Consciousness and Self-Regulation Advances in Research
VOLUME 1
A Continuation Order Plan is available for this series. A continuation order will bring delivery of each new volume immediately upon publication. Volumes are billed only upon actual shipment. For further information please contact the publisher.
Consciousness and Self-Regulation Advances in Research VOLUME 1 Edited by GARY E. SCHWARTZ Yale University
and DAVID SHAPIRO University of California, Los Angeles
PLENUM PRESS· NEW YORK AND LONDON
Library of Congress Cataloging in Publication Data Main entry under title: Consciousness and self-regulation. Includes bibliographical references and index. 1. Consciousness. 2. Self-control. I. Schwartz, Gary E., 1944II. Shapiro, David, 1924BF311.C64 153.8 76-8907 ISBN-13: 978-1-4684-2570-3 e- ISBN -13: 978-1-4684-2568-0 DOl: 10.1007/978-1-4684-2568-0
© 1976 Plenum Press, New York Softcover reprint of the hardcover 1st edition 1976 A Division of Plenum Publishing Corporation 227 West 17th Street, New York, N.Y. 10011
All rights reserved No part of this book may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, electronic, mechanical, photocopying, microfJlming, recording, or otherwise, without written permission from the Publisher
Articles Planned for Future Volumes ,
,
Gyorgy Adam Interoception, Awareness, and Behavior Bernard Glueck and Charles Stroebel Transcendental Meditation: Comparison to EEG Biofeedback Jerre Levy Brain and Consciousness: Cerebral Asymmetry A. R. Luria Brain and Consciousness: Functional Systems Approach Wesley Lynch Biofeedback: Temperature Regulation F. J. McGuigan Imagery and Thinking: The Motor System Martin T. Orne EEG Biofeedback: Relationship to Anxiety Robert Ornstein Dual Modes of Consciousness Kenneth S. Pope and Jerome Singer Regulation of the Stream of Thought Larry Roberts Biofeedback: Use of Curare
Judith Rodin Perception and Externality: Obesity Harold Sackeim and Rubin Gur Sel/Confrontation, Sel/Deception, and Consciousness Bernard Tursky and Milton Lodge Subjective Experience: Psychophysics, Applications to Assessment of Pain, and Political Opinion Takami Wananabe Meditation: Japanese Research Matisyohu Weisenberg Sel/Regulation Therapies: Pain Norman Zinberg Drugs: Interaction of Set and Setting
Contributors
THOMAS D. BORKOVEc, Department of Psychology, University of Iowa, Iowa City, Iowa MONTE BUCHSBAUM, Unit on Perceptual and Cognitive Studies, Adult Psychiatry Branch, Division of Clinical and Behavioral Research, National Institute of Mental Health, Bethesda, Maryland THOMAS H. BUDZYNSKI, Department of Psychiatry, University of Colorado Medical School and Biofeedback Institute of Denver, Denver, Colorado DAVID B. COHEN, Department of Psychology, University of Texas, Austin, Texas DAVID R. ENGSTROM, Department of Psychiatry & Human Behavior and Student Health Service, University of California, Irvine, California ERNEST R. HILGARD, Department of Psychology, Stanford University, Stanford, California E. Roy JOHN, Departments of Psychiatry and Physiology, New York Medical College, New York, New York DONALD MEICHENBAUM, Department of Psychology, University of Waterloo, Waterloo, Ontario, Canada KARL H. PRmRAM, Department of Psychology, Stanford University, Stanford, California
vii
Preface
The first and foremost concrete fact which every one will affirm to belong to his inner experience is the fact that consciousness of some sort goes on. I -William James, 1893
We are witnessing today a mounting interest among behavioral and biological scientists in problems long recognized as central to our understanding of human nature, yet until recently considered out of the bounds of scientific psychology and physiology. Sometimes thrown into the heading of "altered states of consciousness," this growing research bears directly upon such time-honored questions as the nature of conscious experience, the mind-body relationship, and volition. If one broadly views this research as encompassing the two interrelated areas of consciousness and self-regulation, one can find many relevant contemporary examples of creative and experimentally sophisticated approaches, including research on the regulation of perception and sensory experience, attention, imagery and thinking, emotion and pain; hypnosis and meditation; biofeedback and voluntary control; hemispheric asymmetry and specialization of brain function; drug-induced subjective states; and biological rhythms. Because the material is spread over many different kinds of publications and disciplines, it is difficult for anyone person to keep fully abreast of the significant advances. The overall aim of the new Plenum Series in Consciousness and Self-Regulation: Advances in Research is to provide a scholarly forum for discussing integration of these diverse areas by presenting some of the best current research and theory. It is our hope that these volumes will enable investigators to I
William James, Psychology: Briefer Course (New York: Henry Holt and Company, 1893), p.152. ix
x
PREFACE
become more well-rounded in related areas of research, as well as provide advanced students with a ready means of obtaining up-todate, state-of-the-art information about relevant problems, theories, methods, and findings. By selecting significant developments in theory and research, we also hope that over the years the series can help legitimate the field as a scientific venture as well as delineate critical issues for further investigation. Psychology and biology are going through a reawakening, and research on the issues to which this series is devoted is helping to bring these fields closer together. History tells us that Wundt founded psychology as the science of consciousness, and James expanded it to encompass "such things as sensations, desires, emotions, cognitions, reasonings, decisions, volitions and the like."2 But these ideals could not be achieved, or so it seemed, and psychology turned away from questions of experience and volition, as well as from biology, and was replaced with behaviorism. The transformation was arduous, and it required a certain allowance for inconsistency. For example, Edmund Jacobson, one of the pioneers in the psychophysiology of higher mental processes, recalled, "Lashley told me with a chuckle that when he and Watson would spend an evening together, working out principles of behaviorism, much of the time would be devoted to introspection."3 In William James: Unfinished Business (1969), Mandler summarized the good points, and the bad points, of this era of psychology in his "Acceptance of Things Past and Present: A Look at the Mind and the Brain." He aptly noted: I think the Watsonian behaviorist development was inevitable-I think it was even healthy-if we learn not to do it again. Watson and the behaviorists did, once and for all, clean up the problem of the proper data language for psychology. In that sense, we are all behaviorists. The behaviorists inveighed against an establishment which imported theoretical notions and hypotheses into purely descriptive realms of psychology. They successfully excluded vague notions about the causes of behaviorthe introspective statements-from the facts of psychology. But in the process the Watsonians felt called upon to do the reverse and to remove complex and imaginative models from psychology . . . . Behaviorism has been one of the most antitheoretical movements in science .... 2
3
Ibid., p. 1. Jacobson, "Electrophysiology of Mental Activities and Introduction to the Psychological Process of Thinking." In F. J. McGuigan and R. A. Schoonover (Eds.), The Psychophysiology of Thinking (New York: Academic Press, 1973), p. 14.
xi
PREFACE
... I submit that it was this anti theoretical stance that prevented any close attention to physiology . . . . If the mechanisms we postulate are "like" physiological mechanisms, then we will have heeded James in modem terms. But if we are, as we were, afraid to postulate complex mental mechanisms, we will never find the corresponding complex physiological mechanisms.'
This series is dedicated to William James, emphasizing the integration and patterning of multiple processes, coupled with the most significant advances in methodology and knowledge. Some of the chapters will be broad-based and theoretical; others will focus on specific research problems or applications. Inclusion of material in all cases is determined by the investigator's focus on or concern with consciousness and related processes, whether in normal or in abnormal populations. While the editors have a decided bias toward biologically oriented approaches to consciousness and self-regulation, papers that deal primarily with cognition or self-report are included when of particular significance to these topics. Since important findings in this area are often derived from the study of clinical populations and are of direct relevance to the assessment and treatment of psychological and psychophysiological disorders, chapters dealing with basic research are interwoven with chapters of more clinical concern. In this way it is hoped that the series can provide a fertile interchange between the basic and applied sides of this area. To help the reader understand the perspective and rationale for the diverse selections comprising a given volume, a brief overview of each volume is presented by the editors. The impetus for and organization of the series grows out of student response to our interdiSCiplinary seminars at Harvard on the psychophysiology of consciousness, emotion, and self-regulation, coupled with the enthusiasm and support of Seymour Weingarten, Senior Editor of Plenum. Their input, and prodding, is gratefully acknowledged. GARY E. SCHWARTZ DAVID SHAPIRO
4
G. Mandler, "Acceptance of Things Past and Present: A Look at the Mind and the Brain." In R. B. MacLeod (Ed.), William James: Unfinished Business (Washington, D.C. American Psychological Association, 1969), pp. 13, 14.
Overview of Volume 1
In "A Model of Consciousness," E. R. John presents the thesis that "'mind,' under which rubric are subsumed such phenomena as consciousness, subjective experience, the concept of self, and selfawareness, is an emergent property of sufficiently complex and appropriately organized matter." John outlines seven levels of information processing in the brain that correspond to sensations, perceptions, consciousness, content of consciousness, subjective experience, self, and self-awareness. He presents electrophysiological data on both lower animals and man in support of this classification. Based on these findings, he postulates the existence of unique "hyperneurons" in the brain reflecting "complex, three-dimensional volumes of isopotential contours, with a topology encompassing portions of neural membranes, glial membranes, and extracellular binding sites." Karl H. Pribram, in his chapter on "Self-Consciousness and Intentionality ," develops a neuropsychological control-theory model of self-regulation and self-consciousness. Pribram distinguishes among attention, emotion, and motivation and specifically reevaluates Jamesian theory in light of current findings. He argues, drawing on clinical examples as well as research in biofeedback, that "the concepts of feedback and feedforward as they describe closed and open (helical) loop systems are useful in the formulation of a testable model of this domain of inquiry in precise, scientifically useful terms." Beginning with an interest in clinical pain, Monte Buchsbaum reviews the extensive research on augmentation and reduction of sensory input in his chapter "Self-Regulation of Stimulus Intensity." He is particularly interested electrophysiological measures of individual differences in central nervous system control of sensory experience. The relationship of the EEG to psychophysical scaling procexiii
xiv
OVERVIEW OF VOLUME
1
dures is documented, and Buchsbaum illustrates the similarity of these findings to the Russian work on "strength of the nervous system." He introduces the notion of "sensory homeostasis," emphasizing that there exists "an optimal level of continuous sensory stimulation to maintain optimal intellectual functioning or the feeling of well-being." In "Neodissociation Theory of Multiple Cognitive Control Systems," Ernest R. Hilgard notes that "man does more than one thing at a time--all of the time--but the representation of these actions in consciousness is never complete." Drawing initially on the early work in clinical hypnosis and multiple personality, Hilgard presents a "modern comprehensive theory to account for the multiplicity of processes that control overt behavior and conscious processes, with full recognition that something like parallel processing may occur and that all processed information is not available at anyone time to consciousness." Research is reviewed on divided attention, recoverable amnesia, state-dependent learning, hemispheric asymmetry, dissociation within sleep, and multiple personalities, with special attention devoted to new findings on hypnotic analgesia and the recovery of dissociated experiences. Hypnosis and individual differences are discussed further in the chapter by David R. Engstrom on "Hypnotic Susceptibility, EEG Alpha, and Self-Regulation." Engstrom reviews research on the assessment, stability, and modification of individual responses to hypnotic suggestions. Special attention is given to the EEG parameters associated with hypnosis and such related phenomena as perceptual or sensory deprivation. Engstrom explores the relationship between EEG alpha and hypnotic susceptibility in research applying biofeedback procedures in regulating not only the EEG but also the skin temperature. He contends that "biofeedback, hypnosis, meditation, and other training operations which enhance these abilities (muscle relaxation, concentration of attention, and reduction of distraction) should have a similar effect on highly susceptible subjects, reflected in the EEG." In "Toward a Cognitive Theory of Self-Contro!," Donald Meichenbaum sets himself the task cif explaining why "modifying a client's internal dialogue (i.e., self-statements and images) results in behavior change." Drawing on the neurological concept of the final common
OVERVIEW OF VOLUME
1
xv
pathway, Meichenbaum reviews clinical studies and suggests that "the final common pathway to behavior is the internal dialogues in which our clients engage." A three-stage process of cognitive selfcontrol is outlined. He suggests that self-instructions and images affect behavior through influencing attentional direction, as well as influencing a person's interpretation and experience of his physiological state. The interaction of cognitive and physiological processes in a clinical context is discussed by Thomas D. Borkovec in his chapter on "Physiological and Cognitive Processes in the Regulation of Anxiety." Borkovec reviews findings from his research program and outlines a descriptive multiprocess model of anxiety and its regulation. According to Borkovec, the experience of anxiety is elicited both by external fear cues and by internal fear cues, the latter consisting of autonomic arousal, verbal and nonverbal images, and proprioception from overt behavior. Borkovec comes from a decidedly behavior-therapy orientation with strong interests in self-control procedures, and his observations have relevance to an understanding of how different processes interact and combine to elicit the subjective experience of anxiety. In "Dreaming: Experimental Investigation of Representational and Adaptive Properties," David B. Cohen reviews the diverse strategies employed for the study of mental processes during sleep. Dreaming, according to Cohen, is a "psychological process (analogous to thinking) presumably inherent in the neurophysiological activity of the sleeping nervous system." The chapter considers problems of dream recall, the validity of dream reports, and the special adaptive role that dreaming may play in optimal functioning in the waking state. Using available data, Cohen speculates that if "problem-oriented dreaming is an effective vehicle for promoting desirable change in the individual, would it be possible to encourage such changes by experimental manipulation of dream content through pre sleep or sleep suggestion?" In the final chapter, Thomas H. Budzynski considers this question in "Biofeedback and the Twilight State of Consciousness." Interested in altered states of consciousness, Budzynski notes that "when patterning of input to the brain from internal and external stimuli is unusual, out of the ordinary, then the experience may be labeled an altered state." The particular state of consciousness emphasized in the chapter is the "transitory condition wherein one is neither fully awake nor deep asleep," a state once defined by William James as the "fringe
xvi
OVERVIEW OF VOLUME
1
of consciousness." Budzynski reviews the available evidence of the use of biofeedback procedures to help induce and sustain this state and considers possible changes in novel mentation and learning, including sleep learning. Budzynski argues that this learning may involve the minor hemisphere, with associated problems of memory retrieval. Case studies are offered as examples of how biofeedback training of this low-arousal state of awareness, coupled with verbal input, may be used as a behavior-change procedure in psychotherapy.
Contents
1
A Model of Consciousness
1
E.ROY]OHN I. Levels of Information 3 II. A Personal Research Strategy 8 III. EEG Studies 10 A. Changes in Synchrony 10 B. Tracer Technique 11 IV. Average Evoked Potentials 14 A. Appearance of New Components and Increased Similarity of AERs from Different Brain Regions during Learning 15 B. Readout to Absent but Expected Events 15 C. Propagation of Readout from Central Structures 17 D. Differential Readout in Differential Generalization 17 E. AER Correlates of "Meaning" in Human Perception 21 F. Anatomical Distribution of the "Engram" 24 V. Unit Studies 26 VI. Brain Stimulation Studies 31 A. Rapid Transfer to Direct Electrical Stimulation of the Brain 32 B. Peripheral-Central Conflict 32 C. Perceptual Integration 33 D. Loci Responsible for Perceptual Integration 33 E. Role of Cortex and Thalamic Reticular Nuclei 35 VII. Theoretical Discussion of Electrophysiological Evidence 38 References 46 xvii
xviii
2
CONTENTS
Self-Consciousness and Intentionality: A Model Based on an Experimental Analysis of the Brain Mechanisms Involved in the Jamesian Theory of Motivation and Emotion 51 KARL
H.
PRIBRAM
1. A Neurobehavioral Analysis of Brain Mechanisms in Motivation and Emotion 51 A. Introduction 51 B. Case History 53 C. A Mediobasal Motor System 54 D. The Limbic Systems and Behavior 59 II. The Role of Attention in Motivational and Emotional Reactions 66 A. Transfer of Training 66 B. Psychophysiological Experiments 68 C. Habituation 69 D. James Reconsidered 73 III. Effort and the Expression of Motivation and Emotion A. Part Behaviors and Their Integration 74 B. The Precentral Motor Cortex and Action 76 C. Effort and Volition 80 D. The Jamesian Theory of Will 81 IV. A Control-Theory Model of Self-Regulation and SelfConsciousness 83 A. The Model 83 B. Attention Span and Self-Consciousness 88 C. Central Competency 89 D. External Versus Internal Constraint 91 References 95
3
74
Self-Regulation of Stimulus Intensity: Augmenting! Reducing and the Average Evoked Response 101 MONTE BUCHSBAUM
I. Introduction 101 II. Sensory Experience and Augmenting/Reducing
101
CONTENTS
III.
IV. V.
VI.
VII.
VIII.
IX.
X.
xix
A. Petrie and Kinesthetic Figural Aftereffects 101 B. Evoked Responses and Augmenting/Reducing 103 Amplitude/Intensity Relationships in Man 105 A. Visual AERs 105 B. Auditory AERs 107 109 C. Somatosensory AERs D. Summary of Amplitude/Intensity Relationships 110 Augmenting/Reducing Reliability and the Measurement of the AER 111 Genetic Factors in Augmenting/Reducing 115 A. Twin Studies 115 B. Sex and Chromosome Differences 117 117 Tolerance for High-Intensity Stimulation 117 A. Pain Tolerance B. Noise Tolerance 119 Effects of Arousal, Attention, and Sensory Overload 120 A. AER Decrement over Sessions 120 B. AER Decrement with Mental Arithmetic 120 C. AER Decrement with Loud Noise 122 D. Differential Types of AER Decrement 122 Individual Differences and Intensity Judgments 123 A. Psychological Magnitude and Power Functions 123 B. Power Function Exponents and Augmenting/ Reducing 123 C. AER and Psychophysical Scaling 124 Sensory Sensitivity and "Strength of the Nervous System" 125 A. Response to Low-Intensity Stimuli 125 B. "Strength of the Nervous System" and Reducing 125 C. Determination of Strength 126 Self-Regulation and Sensory Homeostasis 127 127 A. Optimum Levels of Stimulation B. Relationships between Pain Tolerance, Sensory Homeostasis, and Distraction 127 128 C. Conclusion References 128
XX
CONTENTS
4
Neodissociation Theory of Multiple Cognitive Control Systems 137 ERNEST
R.
HILGARD
I. II. III. IV.
Pierre Janet's Theory of Dissociation 138 Why a Neodissociation Theory? 141 The Hypnotic Model 142 Neodissociation Model of Multiple Cognitive Control Structures 145 V. Empirical Approaches to Multiple Control Structures and Divisions of Consciousness 152 VI. The Duality of Responsiveness to Pain as Related to Neodissociation Theory 157 VII. Conclusion 168 References 169
5
Hypnotic Susceptibility, EEG-Alpha, and Self-Regulation 173 DAVID
R.
ENGSTROM
I. Introduction 173 II. The Assessment of Hypnotic Susceptibility 175 A. Early Objectification 175 B. Modem Hypnotic Susceptibility Scales 176 III. Stability of Hypnotic Susceptibility 180 IV. Modification of Hypnotic Susceptibility 181 V. Hypnotic Susceptibility and Personality 182 A. Age and Development 183 B. Motivation 184 VI. Hypnosis and the EEG 185 VII. EEG and Hypnotic Susceptibility: Indirect Relationships 187 A. Age 187 B. Perceptual or Sensory Deprivation 188 VIII. EEG and Hypnotic Susceptibility: Direct Evidence A. Base-Rate Alpha Density 189 B. Base-Rate Alpha Amplitude 191 C. EEG Asymmetry 192 D. Evoked Potentials 192 E. Conclusion 192
189
xxi
CONTENTS
IX. X. XI. XII. XIII.
6
The Stability of EEG Base Rates 193 Increasing Susceptibility by EEG Feedback Changes in EEG during Hypnosis 203 Task-Specific EEG Changes 207 Conclusions 215 References 217
195
Toward a Cognitive Theory of Self-Control 223 DONALD MEICHENBAUM
I. Introduction 223 224 II. Conclusions from Treatment A. How Shall We Treat Our Clients' Cognitions? 225 B. Cognitions as Final Common Pathways 238 C. Initial, Conceptualization Phase of Therapy 239 243 III. A Cognitive Theory of Self-Control A. A Three-Stage Process 243 B. How Does Behavior Change through Internal Dialogue? 248 IV. Summary 253 References 255
7
Physiological and Cognitive Processes in the Regulation of Anxiety 261 THOMAS D. BORKOVEC
I. A Descriptive Model of Anxiety Process 264 A. Current Stimulus Conditions 264 B. The Immediate Anxiety Reaction 266 C. Subsequent Maintaining and Reducing Reactions 268 272 D. Intervention Strategies II. Research Studies on the Maintenance and Reduction of Anxiety 276 A. The Role of Physiological Arousal and Cognition 279 B. The Role of Individual Differences in Physiological Arousal and Autonomic Perception 289 III. Summary and Conclusions 305 References 308
xxii
8
CONTENTS
Dreaming: Experimental Investigation of Representational and Adaptive Properties DAVID
B.
313
COHEN
I. Dream Recall 313 A. The Role of Repression 313 B. Alternative Factors: Salience and Interference 317 C. Implication for Theory 324 327 II. Representational Properties of Dreaming 327 A. Validity of Dream Reports B. Two Strategies for Investigating Dreaming 328 III. Functional Properties of Dreaming 345 A. Functions of REM versus NREM Sleep 346 B. REM Psychology versus REM Physiology 349 C. Dream Content and Psychological Change 351 References 355
9
Biofeedback and the Twilight States of 361 Consciousness THOMAS
H. BUDZYNSKI
I. The A. B. C. D.
Twilight State 362 Is a Twilight State the Source of Creative Ideas? Biofeedback and Creativity 364 Learning in the Twilight State? 367 The Production of Low Arousal through Biofeedback 373 E. A Twilight-State Biofeedback System 374 II. Future Considerations 379 A. Is Twilight Learning Minor-Hemisphere Learning? 380 B. A Language for the Minor Hemisphere 381 C. Retrieval Difficulties 381 D. Cognitive Balance 382 References 382 Author Index Subject Index
387 395
363
1
A Model of Consciousness E.RoYJOHN
In the first textbook of physiological psychology, written by Wilhelm Wundt (1910) at the end of the 19th century, Wundt defined the task of physiological psychology as the analysis of the physiological bases of consciousness and subjective experience. In the textbook of physiological psychology which I used when a student, written by Morgan and Stellar (1950) in the middle of the 20th century, physiological psychology was defined as the study of the physiological bases of behavior. The word consciousness does not even appear in the index of the latter volume, nor have I encountered it anywhere in the text. Behaviorism, and "operationism," virtually legislated the problems of consciousness and subjective experience out of the domain of the legitimate concerns of "scientific" and especially physiological psychology, whence they remain essentially excluded until this day. Contemporary experimental and physiological psychology, in its zeal to sanitize itself from any taint of its philosophical heritage and to be even more scientific than the "real" sciences, has preoccupied itself with the analysis of behaviors as if they were performed by unconscious or mindless automata. Attention has largely been focused on clarification of the effects of various schedules of reinforcement on operant responses or the brain mechanisms mediating conditioned responses, rather than on the neural bases of cognition. I have a confession to make. I am not now, nor have I ever been, interested in behavior as such. The main reason I work in physiological psychology is because I am interested in the physiological bases of consciousness and subjective experience. I believe that "mind," under which rubric are subsumed such phenomena as consciousness, subjective experience, the concept of self, and self-awareness, is an emergent property of sufficiently complex and appropriately organized matter. E. Roy JOHN . Departments of Psychiatry and Physiology, New York Medical College, New York, New York.
1
2
E.RoYJOHN
In some fashion, cooperative processes between elements of living matter which individually possess only rudimentary properties generate this emergent property for the system, which qualitatively transcends a simple summation of the elementary properties of the constituent parts. One system which possesses this emergent property is the brain, and the relevant constituent elements are the neurons and the glial cells. We do not understand the nature of this cooperative process, the physical and chemical interactions between the elements of matter which produce mental experience. We do not know how big a neuronal system must be before it can sustain the critical reactions, nor whether the critical reactions depend exclusively upon the properties of neurons or only require a particular organization of energy in matter. These fascinating and enormously important problems, in my opinion, should be among the central topics of investigation in physiological psychology and neurophysiology. They have been neglected far too long, while seemingly inexhaustible energy has been lavished on problems of lesser import. I welcome the signs of a resurgent interest in consciousness and subjective experience, as evinced by the appearance of this series of volumes on consciousness and self-regulation. I am convinced that sufficiently powerful experimental and analytical tools are now available to permit significant progress to be made in the understanding of these issues. This chapter provides an opportunity to examine these problems and to discuss how current research findings might be relevant. One becomes painfully aware of the paucity of contemporary thinking about these issues at the very outset of any attempt to formulate meaningful experimental or analytic approaches to the physiological processes responsible for consciousness and subjective experience. A prerequisite for experimental analysis of these problems must be an adequate definition of what is to be analyzed. Especially because most experiments requiring manipulation of the brain must be carried out in animals, except for the small although invaluable body of data slowly accumulating from the study of "nature's experiments" in cases of human brain injury or disease, an operational definition of consciousness is absolutely essential. Without an unequivocal definition of consciousness, it is hopeless to attempt to identify the responsible processes in the brain. How shall we decide when consciousness is present in an experimental preparation? What constitutes the content of consciousness? Is
A
MODEL OF CONSCIOUSNESS
3
there a difference between consciousness and self-awareness? What do we mean by subjective experience? Is subjective experience sensation, or the perception of sensations, or the apperception of sensations, or something more than any of these?
I.
LEVELS OF INFORMATION
It is inordinately difficult to formulate answers to these questions which seem at all adequate, evading the pitfalls of triviality on the one hand and of useless vague generality on the other. I propose the following definitions as first approximations which provide a basis for an experimental approach. In a later portion of this article, I will relate some current experimental results to these definitions. 1. Sensations are the spatiotemporal patterns of information arriving in the central nervous system because of the excitation of exteroceptive and interoceptive organs. They are a product of the irritability of living matter and constitute first-order information. Such irritability is manifested throughout the phylogenetic scale and is already present in protozoans. Sensations can elicit reflex responses, adjusting the organism to its environment. 2. Perceptions are the interpretation of the meaning of sensations in the context of stored information about previous experiences. Perceptions constitute second-order information resulting from an interaction between sensations and memories. Wundt and his contemporaries argued that the presence of consciousness was revealed when behavioral responses to stimuli ceased to be reflexive and displayed "purposiveness," by which they meant actions which were adaptive and resulted in the adjustment of the organism to its environment as a function of the experiential context of a stimulus rather than to the action of the stimulus alone. For this reason, they considered identification of the lowest phylogenetic level showing learning as crucial for the decision as to the lowest level of organization capable of sustaining consciousness. In this regard, it is noteworthy that Coming, Dyal, and Willows, in their authoritative review of invertebrate learning (1973), reached the conclusion that although the evidence for simple learning or associative conditioning remains highly controversial, there exists compelling evidence that protozoans display the ability to learn not to respond, i.e., habituation,
4
E.RoYJOHN
and some evidence for associative learning has been forthcoming. The capacity for complex learning clearly appears in the phylum Platyhelminthes, with the advent of a brain, defined sensory systems, and complex nerve bundles. We choose to define perception, as well as sensation, provisionally as preconscious or unfelt categories of information processing. Sensations and perceptions are unimodal, referring to the detection and interpretation of stimuli within individual sensory modalities. These functions can be performed by machines, which do not possess consciousness. We contend that under ordinary circumstances, fundamental sensations and much of perception, as defined, do not enter consciousness, although we can make ourselves aware of them by an analytic process. 3. Consciousness is a process in which information about multiple individual modalities of sensation and perception is combined into a unified, multidimensional representation of the state of the system and its environment and is integrated with information about memories and the needs of the organism, generating emotional reactions and programs of behavior to adjust the organism to its environment. Consciousness is third-order information. Many levels of consciousness can exist, in which these dimensions are present in variable amounts. The content of consciousness is the momentary constellation of these different types of information. At the same time that consciousness is the product of an integration of preconscious sensations and perceptions structured in the light of previous experience and reflecting emotional state, drive level, and behavioral plans, feedback from consciousness to these more fundamental levels must take place. Memories are activated, attention is focused, perceptions influenced, emotions aroused, drive priorities altered, and plans of behavior revised as a result of this feedback, producing a continuous reorganization of basic processes because of the influence of higher-level integrative and analytical functions. 4. Subjective experience derives from information about the content of consciousness. It is a process which reorganizes the sequential series of events into a single experiential episode, which merges sequential constellations of multisensory perceptions, memories, emotions, and actions into a unified and apparently continuous event, or "experience," which has a beginning and end. Two critical transformations occur as a result of the process which generates this fourth-order infor-
A
MODEL OF CONSCIOUSNESS
5
mation. First, although the information impinging upon the neuronal populations mediating each of the different dimensions of consciousness is represented by the same mechanism (spatiotemporal patterns of neural discharge) in every such population, the fourth-order information about each different dimension of consciousness is qualitatively distinct. Subjective experience consists of diverse colors, shapes, sounds, textures, smells, tastes, emotions, plans, movements, and thoughts, rather than a uniformly encoded description of these disparate facets of experience. Somehow, qualitative diversity at this higher level of information is constructed out of representational uniformity at lower levels. At the same time, in spite of these qualitative distinctions between the different facets of consciousness and the capability to decompose experience into its constituent components, subjective experience merges these facets into an apparently simultaneous and continuous multidimensional unity. As this unified subjective experience begins to take shape from a related series of episodes, memories relevant to this holistic event are activated, many of them in modalities not involved in the episodes taking place. Some of these memories are of rudimentary or fragmentary sensations, while some are of prior subjective experiences (see below). 5. The self: Second, as subjective experience extends through time and an individual history is accumulated, memory of the sequence of episodes is constructed. This personal history, the accumulated memories of sets of fourth-order information, constitutes the basis for what we call the self. The concept of the self arises as a result of longterm memories constituting the record of an individual's subjective experiences. This individual historical record constitutes fifth-order information. 6. Self-awareness: If we consider subjective experiences as "higherorder sensations," then "self-awareness" is analogous to the perception of those sensations. By this is meant the interpretation of subjective experience in terms of the previous history of life experiences of the individual. Self-awareness is the interpretation of present subjective experience in the context of the salient features, especially the more invariant features, of the pattern of previous subjective experiences. Self-awareness constitutes sixth-order information. As the momentary content of consciousness is interpreted in the light of past experience, feedback to lower levels occurs which is
6
E. Roy
JOHN
probably more powerful than any described thus far. This feedback can be expected to activate trains· of memories of other relevant life experiences, with a high probability that important occurrences (high drive level, high-emotion events) will be followed by systematic or "rational" memory searches. The relatively global feedback resulting from the integration of lower-level information as it enters consciousness is modulated and made far more selective and better-focused. Among the consequences envisaged as resulting from this highest level of information are systematic evaluation of a flood of memories, identification of appropriate perceptions and rejection of more inappropriate perceptions which arose earlier in the experience, selection of the most appropriate emotional response, adjustment of drive levels to correspond to the exigencies and possibilities of the moment, and rational construction of the optimal program of behavior. These processes are far more deliberate and analytical than those previously described. A characteristic of self-awareness is the capacity for cognitive processes. By cognition or thought we mean the ability to have subjective experience vicariously, by activating stored memories about perceptions and prior experiential episodes in a fashion which may be arbitrarily organized rather than occurring according to a previously established sequence. Because of this ability to manipulate, recombine, and reorganize the accumulated store of memories, the self is continuously in the process of modification and of analysis of its own experience. A cognitive process is the representation of an experience in an abstract symbolic fashion, whether or not that experience actually occurred in that form in the personal history of the individual. The distinction between the memory of a rudimentary sensation postulated as essential for perception and the memory of a subjective experience is the amount and diversity of the stored information. The basic neurophysiological mechanisms may be quite the same, and even the anatomicalloci may be shared. I see no compelling need to separate those mechanisms conceptually. Under some circumstances, particularly when cross-modal stimulation is utilized, generalization affords evidence for the presence of cognitive processes. In generalization, an organism interprets the meaning of a sensory stimulus as equivalent to some other sensation because of similarities in the abstract properties common to both stimuli. If the two stimuli are in different sensory modalities, it is clear that some nonsensory specific abstraction has been performed. If the stimuli are in the same senso'ry modality,
A
7
MODEL OF CONSCIOUSNESS
interpretation becomes more equivocal because of the possibility that similar receptors were activated. Observational learning seems to constitute a more unequivocal type of evidence for the presence of cognitive processes. We and others (Chesler, 1969; John, Chesler, Bartlett, and Victor, 1968; GrinbergZylberbaum, Carranza, Cepeda, Vale, and Steinberg, 1974) have shown that naive animals can learn complex discriminative behaviors simply by observing the performance of trained animals. Since the observing animals do not directly experience the reinforcing stimuli, their acquisition of the discrimination must be attributed to their interpretation of
,I
SENSATIONS - - -
I
ENVIRONMEHT
SEQUENCE OF VISUAL STIMULI
I
SEQUENCE OF AUDITORY STIMW
1 1
I
MEMORIES OF PAST SENSATIONS
1 SEQUENCE OF VISUAL PERCEPTS
PERCE PTIONS--
I
I
I.
I
-
.-----l
f----
-
I
J
'~
1
I
SEQUENCE OF AUDITORY PERCEPTS j
.~
EMOTIONS l'
I
DRIVE LEVELS
-1
PR~m~I~R OF
,(.
t ,(.
~
t: .
~
INTEGRATION OF UNIMODAL PERCEPTS, MEMORIES, EMOTIONS, DRIVE LEVELS,AND PROGRAMS OF BEHAVIOR INTO A SEQUENCE OF MULTIVARIATE "FRAMES'
CONSCIOUSNESSCONTE NT OF CONSC IOUSNESS
---
- -
MERGING OF SEQUENTIAL EPISODES INTO UNIFIED AND CONTINUOUS SUBJECTIVE EXPERIENCE
SELF
MEMORIES OF PREVIOUS SUBJECTIVE EXPERIENCES
!
_l
-I MOMENTARY CONTENT OF CONSCIOUSNESS
I
SUBJE CTIVE EXPER IENCE - -
SELF-AWARENESS-
1
---
I
I INTERPRETATION OF MOMENTARY ~ CONTENT OF CONSCIOUSNESS
FIGURE 1. Flow scheme for the levels of information involved in consciousness and selfawareness.
8
E. Roy JOHN
what they observe by referring it to memories of previous experiential episodes which they did experience directly. Observational learning already requires consciousness and probably requires self-awareness. I have found it useful to postulate a series of different levels of information processing, each dependent upon all the levels below (feedforward) and each influenced by the levels above (feedback), in order to define sensation, perception, consciousness, the content of consciousness, subjective experience, the emergence of a self-concept, and self-awareness. The proposed definitions treat each of these processes as fundamentally similar to all the others in that they are all representations of information, presumably in a common neuronal code. They are all different in that they constitute successively higher derivations extracted from the information representing the lower derivations. These ideas are illustrated in Figure 1, which has been limited to two sensory modalities to simplify the diagram. If we accept these formulations as working definitions, the task of experimental analysis of these processes may become easier. The processes representing information at the lower levels must be analyzed first. As we gain insights into the representation of lower-level information, it becomes possible to seek invariances across the representation of multiple items on the same level, which share a common informational feature. Such invariances constitute the representation of information on a higher level. In this "bootstrap" fashion, it would appear possible to progress in a systematic development from initial studies of sensory mechanisms to eventual investigations into the neurophysiological basis of self-awareness.
II. A
PERSONAL RESEARCH STRATEGY
When I began to do research, the physiological process by which neuronal activity became transmuted into subjective experience was the problem of greatest interest to me. The difficulty of objective definition of the momentary content of a spontaneous stream of consciousness seemed insuperable, especially in animal experiments. The study of memory offered what appeared to be a unique solution to this dilemma. If one could succeed in understanding how the information about a specific experience was encoded, stored, and retrieved, then one could identify the physiological processes corresponding to a spe-
A
MODEL OF CONSCIOUSNESS
9
cific memory. When that memory was remembered, the corresponding physiological process would appear. Appearance of that process would constitute an objective indication that a specific past experience was the content of consciousness at that moment. Examination and analysis of the features of such a representational process would provide a description of the physiological mechanism which generated or corresponded to a specific thought. While this description might not explain how the physiological processes engendered the subjective experience, it would tell us what the relevant processes might be. Ideally the experimental situation would be so devised that performance of some particular behavior became much more likely when a specific past experience was remembered, providing objective reassurance that the subjective experience did in fact take place when predicted. For more than 20 years, I have been pursuing this strategy, constantly trying to improve the resolution of my measurements of neurophysiological processes and the design of the experimental procedure. These experiments have primarily been aimed at obtaining a detailed description of the electrical activity of different brain regions in unrestrained animals as they acquired and performed differentiated conditioned responses to discriminative stimuli in each of several sensory modalities. At first, these studies evaluated changes in ongoing electroencephalographic (EEG) activity during conditioning. With the advent of average response computers, our attention shifted to the details of the evoked potentials elicited by the discriminanda. As the results of evoked potential studies provided a relatively clear and consistent picture of the slow-wave phenomena related to information coding and memory retrievel, we used the higher resolution afforded by microelectrode techniques to investigate the behavior of single neurons and small neuronal ensembles under comparable experimental conditions. Finally, when experimental observations permitted formulation of a tentative theory about the salient features of the process by which a past experience was represented, we made an attempt to test the theory directly by using electrical stimulation of the brain to reproduce those hypothetically crucial features, observing behavior to infer what subjective experience had ensued. During this period, other workers carried out a large number of related experiments. In the next section of this article I will briefly
E.RoYJoHN
10
survey this body of research. I will rely mostly upon the work of my own laboratory, because phenomena observed by me personally have had the most impact upon my thinking. Electrophysiological methods have provided unique insights into the details of physiological processes within various anatomical regions and the dynamic transactions between as well as within those regions which take place during learning and which occur when memories are activated. These insights allow us to construct a description of how experiences build an anatomically distributed mediational system in which different parts of the brain cooperate in the representation and procession of information. Detailed reviews of the voluminous evidence on which this description is based are available elsewhere (John, 1961, 1967b, 1971, 1972, 1974; John and Thatcher, 1976; Morrell, 1961b; Thompson, Patterson, and Teyler, 1972). In this article, we have ignored the problem of the chemical processes involved in information storage, which we have discussed elsewhere in detail (John, 1967b).
III. EEG
STUDIES
A. Changes in Synchrony Since the discovery that tiny electrical voltage fluctuations could be recorded from the scalp, the EEG correlates of conditioning have been studied by numerous workers. The general features of the EEG changes observed in such studies are that when a conditioned stimulus (CS), to which the subject has previously been habituated, is initially paired with an unconditioned stimulus (US), widespread changes from relatively low-frequency high-voltage activity (synchronization) to higherfrequency low-voltage activity (activation) occurs in the scalp EEG. As training proceeds, this activation or desynchronization pattern becomes limited to only a few "relevant" regions, for example, over the motor cortex if the conditioned response (CR) requires a movement, over the visual cortex if the CS is a visual signal. Usually changes in the EEG occur prior to the appearance of the first behavioral CRs. During extinction, learning-induced EEG changes persist beyond the disappearance of CRs, with a gradual reversal of the changes seen during acquisition.
A
MODEL OF CONSCIOUSNESS
11
Such findings led to the conclusion that during conditioning there was initial widespread "irradiation" of information over the cortex (adduced as evidence of involvement of the mesencephalic reticular formation early in learning), followed by "consolidation" or more differentiated and localized mediation of performance of well-learned responses (interpreted to indicate a later shift to a dominant role for the thalamic reticular formation, the intralaminar nuclei of the diffuse projection system). Studies of the habituation of the "arousal" response, i.e., gradual disappearance of the activation pattern caused by repeated presentation of a novel stimulus (often considered as a primitive type of perceptual learning), led to analogous concepts of adjustment of iterated inconsequential events, initially involving a phasic diminution of response in the thalamic reticular formation followed by tonic adaptation in the mesencephalic reticular formation, eventually leading to complete suppression of the desynchronization response.
B. Tracer Technique Since its introduction in the Soviet Union by Livanov and Poliakov (1945) and in the United States by John and Killam (1959), tracer technique has been the most useful method for distinguishing between the electrical activity of the brain related to information processing about the learned experience (which we will define as "signal") and the other ongoing business of the brain (which we arbitrarily refer to as "noise" because of our primary concern with brain mechanisms involved in learning and memory). Like many crucial methodological innovations, the idea underlying tracer technique is very simple. The signal of "tracer-conditioned stimulus" (TCS) for the learned behavior under study is presented intermittently at a characteristic rate of repetition. Electrical rhythms which appear in different brain regions at the frequency of the TCS are considered to be "labeled responses" reflecting processing of information about the stimulus. 1. Participation of Many Brain Regions in Learned Behavior
The first findings provided by tracer technique showed that during learning, widespread changes take place in the anatomical distribution
12
E. Roy JOHN
of the brain's responses to the CS. Although the same phenomena of irradiation and consolidation described in prior studies of the desynchronization of the EEG were also observed with labeled rhythms, the decrease in the anatomical extensiveness of the responsive system with well-learned behaviors was only observed in simple CRs, where the animal merely needed to detect the CS. When differential responses required discrimination between different signals, the labeled responses stabilized throughout a widespread anatomical system. 2. Display of Similar Electrical Activity by Many Brain Regions
In the original studies of Livanov and Poliakov and of John and Killam, it was noted with surprise that a number of brain regions which showed markedly different electrical responses to the TCS before conditioning acquired striking similarities in electrical activity during acquisition and subsequent performance of a new behavioral response to that stimulus. Many other workers have commented upon the same phenomenon (Yoshii, Pruvot, Gastaut, 1957; John and Killam, 1959, 1960; Liberson and Ellen, 1960; Galambos and Sheatz, 1962; Glivenko, Korol'kova, and Kuznetsova, 1962; Livanov, 1962, 1965; John, Ruchkin, and Villegas, 1963, 1964; Dumenko, 1967; Knipst, 1967; Korol'kova and Shvets, 1967). These findings indicate that during learning a representational system is established which involves many different anatomical regions in a cooperative, similar mode of activity.
3. "Assimilation" of the Rhythm of the TCS The most intriguing phenomenon observed in the early studies with tracer technique, and since confirmed in many different species and experimental situations, was named assimilation by Livanov and Poliakov. This term referred to the fact that while a CR was being established to a TCS, the spontaneous EEG during the intertrial intervals became dominated by electrical rhythms at the frequency of the absent stimulus. Such rhythms were absent in the home cage but appeared as soon as the animal entered the familiar training environment (Yoshii and Ogura, 1960). It was as if the animal were rehearsing the experience of the meaningful signal which had previously appeared
A
MODEL OF CONSCIOUSNESS
13
in that situation and was again anticipated. Recently, assimilated rhythms with remarkable precision of frequency have been found in the firing patterns of small groups of cortical cells (Ramos and Schwartz, 1976b). The functional significance of these assimilated rhythms is clearly established by the fact that they appear only on the trained side of a split-brain cat (Majkowski, 1967). These phenomena show that the representational system built during an experience with a rhythmic stimulus can produce an electrical facsimile of that rhythm in the absence of the TCS.
4. Exogenous and Endogenous Components of EEG Rhythms Appearance of assimilated rhythms often precedes spontaneous performance of the CR (Yoshii, 1962). Another line of evidence further suggests the functional significance of endogenous electrical patterns, i.e., temporal patterns of activity originating within the brain. As mentioned previously, many different brain regions come to display similar electrical rhythms during conditioning with a TCS. However, when animals who have been highly overtrained in the performance of differential CRs to discriminated stimuli at two different frequencies commit errors, certain brain regions often display electrical rhythms inappropriate to the actual stimulus. Instead, these rhythms correspond to the frequency of the absent stimulus which would be the appropriate cue for the behavior which was performed (John and Killam, 1960; John, 1963, 1967a, 1967b, 1972; John, Leiman, and Sachs, 1961; Lindsley, Carpenter, Killam, and Killam, 1968; Majkowski, 1966). This phenomenon is illustrated in Figure 2. Thus the similar electrical patterns observed in different brain regions of trained animals come from two origins. One, which we term exogenous, reflects afferent input due to external reality. The other, which we term endogenous, reflects the release of previously stored electrical patterns from some internal representational system. The appearance of two different electrical patterns in the brain of an animal as it commits a behavioral error, indicative of misinterpretation of a signal, suggests that the exogenous activity caused by the actual environment somehow activated inappropriate endogenous activity reflecting the significance normally attributed to a different signal. A mismatch has taken place between stimulus input reflecting reality and the retrieval from memory necessary to interpret that reality.
14
E. Roy
JOHN
CORRECT INCORRECT L MOT ,.-..}o-_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ ...}--I-._..,.,..._ _......._ _ _ _...-_......_ _..-_ _ R MOT""",,_
.'PII~.,... - - - - - - - - - - - - - - -
LAUDJ~~~...~~
RAUD ..~.;o.vJ~,~At-..J.,.~
LVIS,.Jw-WV~rtw~~..~"'W\~ RVIS.-J'v..~IM~-I'Mf~...\\·W......,..~~~ LVPL~"~ ...~~,....,..W\"......~~~~ L GL ,~·"'.:;M,"trWtr.,.i'i"",-y.y,...,~.- ~·~'A.'t'..I."PfWrMrl'~l~ R GL"J'I\N""Y...."·,......v..~oJ/V' ~"'."v."'N~"~Wwr~ L GM~ _ _--............""""-""""""""""'.......""",..,.....,.,..,.,.,.."""__.... RGM""W"'-~'"".y..".,...·.••·_ _~ L MRF',.,..Jtv~""-'\~~,...,.J-r...,.'HV.....JI"'.f'.......,..~~ FLICKER 0
FIGURE 2. Comparison of electrophysiological activity elicited by the same visual signal interpreted in two different ways. Records on the left were obtained as the differentially trained cat responded correctly to a negative CS (707-Hz flicker) by performing a conditioned avoidance response. Records on the right were obtained on the next trial, when the cat erroneously responded to the same flicker frequency by performing the conditioned approach response appropriate to the positive stimulus (301-Hz flicker). (MOT, motor cortex; AUD, auditory cortex; VIS, visual cortex; VPL, ventroposterolateral nucleus; GL, lateral geniculate nucleus; GM, medial geniculate nucleus; MRF, mesencephalic reticular formation.) All records bipolar. (Data from John, 1972.)
IV.
AVERAGE EVOKED POTENTIALS
With the advent of special and general-purpose minicomputers capable of averaging many evoked potentials while an experiment was in progress ("on-line" average-response computation), it became possible to analyze the data obtained in conditioning studies using tracer technique by computation of the average evoked response (AER) to the TCS. This enabled examination of the actual waveshape of the response to the stimulus in each brain region and largely replaced the previous preoccupation with the frequency and amplitude of labeled rhythms. AER studies confirmed and extended the conclusions reached in the earlier EEG studies of conditioning. During learning, the anatomical distribution of evoked responses to the TCS becomes more widespread. While responses continue to be displayed by sensory-specific structures of the modality of the signal, new responses appear in regions which showed little or no response to the CS before training.
A
MODEL OF CONSCIOUSNESS
15
These new responses are particularly striking in the mesencephalic reticular formation, the intralaminar nuclei of the thalamus, and in various portions of the limbic system, especially the hippocampus.
A. Appearance of New Components and Increased Similarity of
AERs from Different Brain Regions during Learning As conditioning proceeds, a new late process with an onset latency about 60 milliseconds after the CS appears in the AER recorded from many brain regions (Asratyan, 1965; Begleiter and Platz, 1969; Galambos and Sheatz, 1962; John, 1963, 1967a, 1967b; John and Killam, 1959; John and Morgades, 1969a; Killam and Hance, 1965; Leiman, 1962; Lindsley, Carpenter, Killam, and Killam, 1968; Sakhuilina and Merzhanova, 1966). Different brain regions display markedly disparate AERs at the onset of training and acquire similarities in AER waveshape as conditioning proceeds. Further, when the signal fails to elicit CR performance, the new late process often is absent from the AER in many regions. These observations are illustrated in Figure 3.
B. Readout to Absent but Expected Events A body of evidence has accumulated which shows that certain aspects of the evoked potential (EP) may reflect previous experience rather than responses to afferent input and are thus of endogenous rather than exogenous origin. One important line of such evidence comes from studies primarily carried out on human subjects and is particularly important in the assessing of the likelihood that these released electrical patterns actually correspond to the activation of specific memories, because it has been possible to establish unequivocally that there is a subjective correlate to the appearance of these released potentials. These studies show that when an expected event does not occur, a cerebral potential appears at a latency similar to that of potentials usually evoked by the expected stimulus. EPs elicited in man by absent events have been reported (Barlow, Morrell, and Morrell, 1967; Klinke, Fruhstorfer, and Finkenzeller, 1968; Picton, Hillyard, and Galambos, 1973; Riggs and Whittle, 1967; Rusinov, 1959; Sutton, Tueting,
E. Roy JOHN
16 CONTROL VISUAL CORTEX -" LATERAL GENICULATE AUDITORY
JL -f'v---- .J'\;- ~
-vv ~ V'- y--
r v---- -v-:--,--
J'v. ~
MES. RETICULAR FORMATION
~
NUe. RETICULARIS
,~
NUC. CENTRAlIS
OIFFEREHTlAI. OVERTRAINED CAR CAR
,r---. ~ V'-- ~
CORTEX ........--~
MEDIAL GENICULATE
EARLY CAR
~
\/./'------ ""\ v
r---
~
LATERALlS ~' --yv---" ~. r----- \f\t/
DORSAL HIPPOCAMPUS ........./'\...rJ' MEDIAl. FOREBRAIN BUNDLE PREPYRIFORM
~"'~
r
~- .~ ~ ~ v
' /---"-
v.~
CORTEX -' _~ /"- ' _~....... YI
NUC. SUBTHALAMUS •. ,J~" SUBSTANTIA NIGRA
"-
_ "\.
t
'""'~
'\J\.,
~
.J'" ... \ :. \.........
Ay.- "'-
V" t \;' .-,, - . ., " ~ -~
'v-~.. "
v
12' ..
I \ /
t
FIGURE 3. Evolution of visual evoked responses. Control, average responses evoked in different brain regions of a naive cat by presentation of a novel flicker stimulus. Several regions show little or no response. Early CAR, responses to the same stimulus shortly after elaboration of a simple conditioned avoidance response (CAR). A definite response with similar features can now be discerned in most regions. Differential CAR, changes in the response evoked by the flicker CS shortly after establishment of differential approachavoidance responses to flicker at two different frequencies. As usual, discrimination training has greatly enhanced the response amplitude, and the similarity between responses in different structures has become more marked. Overtrained CAR, after many months of overtraining on the differentiation task, the waveshapes undergo further changes. The arrows point to a component usually absent or markedly smaller in behavioral trials on which this animal failed to perform (nuc. reticularis, nucleus reticularis; nuc. subthalamus, nucleus subthalamus) . (Data from John, 1972.)
Zubin, and John, 1967; Weinberg, Walter, and Crow, 1970; Weinberg, Walter, Cooper, and Aldridge, 1974). Similar findings in the cat were reported by John (1963). These cerebral events, termed readout or emitted potentials, have been interpreted by Weinberg et al. to reflect the generation of processes corresponding to the memory of past or imaginary stimuli.
A
MODEL OF CONSCIOUSNESS
17
C. Propagation of Readout from Central Structures When generalization occurs upon presentation of a novel test stimulus, the AER in the lateral geniculate body and in many other structures closely resembles the waveshape usually evoked by the visual CS. However, if generalization fails to occur, the response to the test stimulus differs radically from the typical AER to the CS, lacking the late components (Ruchkin and John, 1966). The same phenomenon has been found in the firing patterns of neuronal ensembles in the lateral geniculate during generalization (John and Morgades, 1969b). This phenomenon is illustrated in Figure 4. By subtraction of AERs from trials in which no behavioral response was elicited by presentation of the test stimulus from AERs computed during trials resulting in generalization, it was possible to construct the difference waveshapes, showing the forms and latency of the readout process released in different brain regions during generalization (John, Ruchkin, Leiman, Sachs, and Ahn, 1965). Readout processes were found in most brain regions studied and displayed a general similarity of waveshape with marked latency differences from region to region. The readout process seems to arise in a central corticoreticular system, from which it propagates to involve other brain regions in a systematic sequence, appearing last in the lateral geniculate body (when a visual cue is used). The fact that the thalamic "relay" nucleus for visual information is so dramatically influenced by this centrifugal process gives some insight into the compelling influence of experience upon perception. This finding is illustrated in Figure 5.
D. Differential Readout in Differential Generalization The readout process is not merely a nonspecific indicator that memory retrieval is in progress. The shape of the readout process depends upon which memory is activated. This has been established by a technique called differential generalization. Cats are trained to perform two different CRs (CRI and C~) to two discriminated visual stimuli, consisting of flicker at two different repetition rates (VI and V2 ). After thorough overtraining, a test stimulus (Va) is occasionally introduced into random sequences of VI and V2 • The frequency of Va is midway between VI and V2 •
18
E. Roy
JOHN
CONDITIONED RESPONSE To IO/SEC
II
GENERALIZATION RESPONSE TO 7.7/SEC
NO RESPONSE TO 7.7/SEC
I A
I
R. LATERAL GENICULATE
R. NUCLEUS RETICULARIS
GENERALIZATION TO I/S TEST STIMULATION
,..',
,,
B
'
A
MODEL OF CONSCIOUSNESS
19
Sometimes the cat treats V3 as equivalent to VI and CR1 is performed (V3CR1 ). On other trials the cat treats V3 as if it were V2 and C~ is elicited (V3C~), AER waveshapes during V3CR1 trials are significantly different from those found during V3C~ trials. When V3 presentation results in CR1 performance, the AER elicited by V3 closely resembles the usual evoked response to V2 • Conversely, when V3 presentation results in C~ performance, the AER to V3 is like that usually caused by V2 (John, Shimokochi, and Bartlett, 1969). These findings are illustrated in Figure 6. This phenomenon has been analyzed in great detail, by the use of visual, auditory, and electrical stimuli delivered directly to brain structures. A wide variety of different instrumental tasks have been utilized and many controls introduced to rule out possible unspecific causes for this phenomenon. Methods of computer-pattern recognition have been developed to permit classification of single evoked response waveshapes. These further studies have shown that readout processes during differential generalization can be found in most brain regions, are demonstrable under all stimUlus-response contingencies thus far explored, and cannot be attributed to unspecific origins. Repeated presentation of the repetitive test stimulus elicits a variable sequence of EP waveshapes or modes. The CR subsequently performed to the
4. (A) Computations of average responses obtained from the lateral geniculate nucleus and nucleus reticularis of the cat under various conditions during the same experimental session. First row of averages is based upon 100; second and third rows are based upon 42 repetitions of the same stimulus applied during a number of behavioral trials. Analysis epoch was 90 msec. First row: Average responses evoked in structures by the 10-Hz CS (flicker) actually used in training, during repeated correct behavioral performances. Second row: Average responses evoked by a novel 7.7-Hz CS, during repeated generalization behavior. Test trials with the 7.7-Hz stimulus were interspersed among trials with the actual 10-Hz CS and were never reinforced. Third row: Average responses evoked by the 7.7-Hz flicker on presentations when no generalization behavior was elicited. The waveshape elicited by the actual CS is similar to the response evoked by the novel stimulus during generalization behavior. Notice the absence of the second positive component in the EP when generalization behavior failed to occur. (Data from Ruchkin and John, 1966.) (B) (Top) Records of AER's and PSH's obtained during 18 trials that resulted in CR to the 2-Hz CS (dotted curves) and during 32 trials that resulted in behavioral generalization in response to a I-Hz flicker used as a test stimulus (solid curves). The test stimuli were randomly interspersed between presentations of 2-Hz (dotted curves) and 8-Hz flickers in a long experimental session. (Bottom) Records of AER's and PSH's obtained during 17 trials that resulted in failure to elicit generalization behavior in response to the test stimulus. Note change in late components. Analysis epoch, 100 msec. (Data from John and Morgades, 1969a.)
FIGURE
20
E. Roy
JOHN
test stimulus is consistently related to the predominant EP mode identified by the pattern recognition procedure (John, Bartlett, Shimokochi, and Kleinman, 1973). Thus an ambiguous stimulus activates a variety of readout processes identifiable by the different features of late portions of the EP. The behavior eventually displayed seems to depend upon the particular readout mode which becomes dominant in the representational system. These findings indicate that in many brain regions the waveshape of AER elicited by a stimulus is not determined solely by its phYSical
RPOST
"A::t~
L.POSTSS~
R.MRF
R.N.VENT
7\r
R.MARG
R'OORS""~ R.N.LAT p~
5. Difference waveshapes obtained for a number of regions by subtraction of averaged responses, evoked by a 7.7-Hz test stimulus during nonperformance from averaged responses elicited when generalization occurs. All averages in these computations were based upon 200 EPs distributed among a number of behavioral trials in each category, with a 62.5 msec analysis epoch. (Data from John, Ruchkin, Leiman, Sachs, and Ahn, 1965.) FIGURE
R.N.RETIC~ R.ANTLG~
A
MODEL OF CONSCIOUSNESS
21
FIGURE 6. Waveshapes of average responses recorded from the LG (bipolar) under various stimulus-response contingencies: V,CR, during trials resulting in. correct performance of an approach response (CR) to a 3.1-Hz flicker CS; V,CAR, during trials in which a conditioned P (.01 avoidance response (CAR) was correctly performed in response to a 7.7-Hz CS; V3 CR, during V3 CAR generalization trials in which a neutral 5-Hz test stimulus elicited CR behavior; V3 CAR, during generalization trials in which the same 5-Hz test stimulus elicited CAR behavior. The interrupted line between V3 CR and V3 CAR indicates time 100 MSEC intervals during which V3 CR and V3 CAR were significantly different at better than the P = .01 level. The numbers at the right indicate the correlation coefficients between the corresponding bracketed waveshapes. (Data from John, 1972.)
parameters but is strongly influenced by the meaning attributed to it in the context of memories about previous similar experiences.
E. AER Correlates of "Meaning" in Human Perception Jacobo Grinberg-Zylberbaum and I recently carried out an electrophysiological experiment on human subjects which showed that the shape of the AER in some cortical regions depends upon the meaning attributed to a visual stimulus, rather than upon its form. The experiment consisted of two portions. In the first part, subjects seated before a tachistoscope viewed brief presentations of a vertical line followed by presentation of the number 2. This stimulus sequence was repeated 100 times at intervals of 400 msec, while evoked responses to the vertical line were recorded from occipital (01 and O 2 ), parietal (P3 , P4 , and Pz ), and temporal (Ts and Ts) derivations by use of a linked earlobe reference. 1 The subject then viewed 100 presentations of the same vertical line, but now followed by the letter K. Evoked responses to the vertical line were again recorded during this second stimulus sequence. During the first sequence, in which the vertical line was followed by the number 2, it was perceived as the number 1. During the second sequence, when the vertical line was followed by the letter K, it was perceived as the letter I. Thus the same vertical line (sensation) activated two different perceptions. Using a PDP 12 computer, we computed the AERs and the stand1
Letters refer to electrode position in the International 10/20 System.
22
E. Roy JOHN
I AS A NUMBER
I AS A LETTER DIFFERENCE WAVE
t. TEST
0.01 LEVEL - I1iW. jIl. i.L.~_
[J)N\. T5
P4
I AS A NUMBER
I AS A LETTER
DIFFERENCE WAVE
t. TEST
JV\N
rJ\; ~~~
0.01 LEVEL -+ ~f0 L.T.
S.J.
S.J.
S.J.
_1£'.1.1:
I'-,,~
~~
~ ..
~ J.C.
I 396 msee. I
FIGURE 7. (Top) Examples of averaged EPs to a vertical line presented in a context of numbers (Line 1) and in a context of letters (Line 2). The difference wave obtained by the subtraction of Line 2 from Line 1 is shown in Line 3. Line 4 shows the value of the t test at each point along this analysis epoch. Statistical significant differences were obtained in parietal and temporal derivations in the EP components located between 150 and 200 msec of latency. Each average EP was computed from 100 samples. Average responses, variances, difference wa'les, and the t test were computed with a PDP-12 computer. (Bottom) Same data from a second subject.
ard deviations from each deviation for the vertical line in the two different sequences. The AER to the vertical line perceived as a number was then subtracted from the AER to the vertical line perceived as a letter. The significance of the resulting difference wave was assessed at many points along the wave, each representing successive latency
A
23
MODEL OF CONSCIOUSNESS
increments of 2 msec, by use of the t test. The results from typical subjects are illustrated in Figure 7. Figure 7 shows that no significant differences were found between the AERs to the vertical line under the two different perceptual sets in the primary visual receiving areas (01 and ~). That is, the sensation caused by the vertical line was essentially the same in both stimulus sequences. However, significant differences did occur in the parietal and temporal derivation. Essentially, the same procedure was used by Johnston and Chesney (1974), who obtained results comparable to ours. Differences between the two perceptual sets were found in frontal but
DiU. Wavi
!\ ................ 0',&
.. .0
'.rn=p . .
0.01
IIvel
't Tlst
01
P3
Diff. Wave
• V"'""
0", •
"'V
Q ....,.,...
6,
·V., ..."'-"
...
eT~M-l. ~ &~ r%"F\>. B.K.
1 396111S1c
I
FIGURE 8. (Top) The difference wave (Line 1) and the t test (Line 2) obtained by comparison of average EPs (100 samples) elicited by a big A and a little Q. Only the occipital location shows a statistically significant difference. (Bottom) The same calculations, but now of EPs elicited by a capital A and capital E. All the locations show highly significant differences between these EPs.
24
E. Roy JOHN
not occipital regions. No data were obtained from parietal or temporal derivations in that study. In the second experiment, a four-stimulus sequence was tachistoscopically presented, consisting of a large A, a small a, a large E, and a small e. This sequence was repeated 100 times. AERs and standard deviations were again computed for the response to each stimulus from every derivation. The results are shown in Figure 8. When the AERs elicited by small and large versions of the same letter were compared, significant differences were found in the occipital derivations. That is, large and small letters produce different sensations. However, no significant differences were found in temporal or parietal derivations. Large and small versions of the same letter activate the same perception, denoting a particular symbol in the alphabet. Finally, when AERs elicited by As and Es of the same size were compared, significant differences were found in all derivations. Both the sensations and the perceptions elicited by two different letters are different.
F. Anatomical Distribution of the "Engram" If these endogenous or readout processes represent the activation of specific different memories, the anatomical distribution of these endogenous electrical patterns provides information about the locus of the neural representational system which mediates the storage and retrieval of a memory. Traditionally this representational system has been referred to as the engram. By appropriate computer manipulations of AERs from different stimulus-response combinations, the waveshapes of residuals reflecting only exogenous or only endogenous processes can be constructed. By these methods, it is possible to obtain a quantitative estimate of the relative contributions of exogenous and endogenous processes to the AERs obtained from any brain region. Such quantitative estimates have been computed for a wide sample of brain regions in many cats, by the use of data from thousands of behavioral trials of differential generalization to auditory or visual stimuli (Bartlett and John, 1973). For each brain region, the contribution of exogenous processes to the AER was plotted versus the contribution of endogenous processes. The results are illustrated in Figure 9.
A
25
MODEL OF CONSCIOUSNESS
.40 ~
C
;:)
•
FLICKER 0 CLICK
. ~, L VISUAL / SYSTEM •
.30
0
en.... a:
fI)
;:)
0
z ....
NON SPECIFIC, • MOTOR, SYSTEM
.20
'" 0 0
z.... .15 z
........
S~~~~Y
• MESENCEPHALIC NON SPECIFIC
• • AUDITORY SYSTEM
~ I-
.... CD
.10 .09 ~ .......Ja: .08 a: .07 0 Z 0
LIMBIC SYSTEM
u
z .06
'":l
.05 .04 .00
.10
.20
.30
.40
.50
.60
FIGURE 9. Plot of mean correlation coefficients between exogenous residuals versus endogenous residuals for different neural systems and for different cue modalities. Closed circles: Flicker frequencies as stimuli. Auditory system: N = 305; aud. cx. (16 cats), med. genic. (16), brach. info coil. (1). Limbic system: N = 303; hippocampus (16), dentate (5), cingulate (5), septum (5), prepyriform (6), med. forebrain bundle (6), mamm. bodies (5), hypothalamus (7). Mesencephalic nonspecific: N = 158; retic. form. (18), cent. gray (1), cent. teg. tract (1). Motor system: N = 146; motor cs. (4), subs. nigra (10), nuc. ruber (4), nuc. vent. (9), subthal. (5). Other sensory: N = 54; sensorimotor CX. (4), nuc. post. lat. (1), nuc. vent. post. lat. (5), nuc. vent. post. med. (1) Thalamic nonspecific: N = 139; cent. lat. (13), nuc. retic. (6), nuc. reuniens (1), med, dors. (5), pulvinar (1). Visual system: N = 394, visual cX. (18), lat. genic. (18), sup. coil. (2). Open circles: Click frequencies as stimuli. Auditory system: N = 48; aud. cx. (5 cats), med. genic. (5). Limbic system: N = 69; hippocampus (5), dentate (3), cingulate (3), septum (3), prepyriform (2), med. forebrain bundle (3), mamm. bodies (3), hypothalamus (2). Motor system: N = 37; motor cx. (1), subs. nigra (4), nuc. ruber (1), nuc. vent. ant. (5), subthal. (2). Nonspecific system: N = 50, mesen. retic. form. (6), cent. gray (1), cent. teg. tract (1), cent. lat. (3), nuc. retic. (3). Visual system: N = 55; visual cX. (6), lat. genic. (6), sup. coli. (1). N denotes the number of independent measurements within the designated system. Data from monopolar and bipolar derivations were combined. Replications varied across cats and structures. (Data from Bartlett et a/., 1975.)
E.RoYJOHN
26
Figure 9 reveals a systematic relationship between the contribution of exogenous and endogenous processes to the AER: The amount of
endogenous process in any brain region is logarithmically proportional to the amount of the exogenous process in that region. Both exogenous and endogenous processes vary greatly from region to region. All regions are not equivalent with respect to signal to noise ratio nor participation in the representation of experience. In spite of these large quantitative differences, the qualitative effects of a sensory stimulus are distributed throughout the brain, and all brain regions participate in the representation of experience to an extent proportional to that afferent input.
V.
UNIT STUDIES
The EEG and AER observations thus far described suggest that neurons in widespread regions of the brain are involved in the mediation of learning and memory. A large body of evidence supports this conclusion and has been thoroughly reviewed (John, 1967a,b, 1971, 1972; John and Thatcher, 1976). Abundant documentation has established the existence of strong correlations between spontaneous or evoked activity on one hand and neural discharge on the other. In view of the anatomically extensive changes in slow waves during learning, these correlations lead one to expect that changes in single unit activity during learning should be detected with ease. That such is indeed the case can be seen from studies of single neurons during conditioning, which consistently report that a large proportion of cells (10%-60%) alter their response to the CS during conditioning (Jasper, Ricci, and Doane, 1960; Morrell, 1961a; Morrell, 1967; Olds and Olds, 1961; Bure!!, 1965; Bure~ and Bure!!ova, 1965, 1970; Kamikawa, Mcnwain, and Adey, 1964; Buchwald, Halas, and Schramm, 1965; Hori and Yoshii, 1965; Adam, Adey, and Porter, 1966; Yoshii and Ogura, 1960; Woody, Vassilevsky, and Engel, 1970; Travis and Sparks, 1967, 1968; Travis, Hooten, and Sparks, 1968; Travis, Sparks, and Hooten, 1968; O'Brien and Fox, 1969a, 1969b; Olds and Hirano, 1969; Olds, Disterhoft, Segal, Kornblith, and Hirsch, 1972; Leiman and Cristian, 1973). The fact that such a large percentage of neurons change their behavior during even the simplest learning task constitutes a strong argument against connectionistic formulations which localize learning as synaptic alterations causing facilitation of the firing of cells in selected pathways, with such firing constituting the activation of memory
A
MODEL OF CONSCIOUSNESS
27
about the learned experience. If mere neuronal firing per se constituted information about present or prior experience, such representations would be hopelessly ambiguous since so many cells alter their reactivity in a single learning session and since neurons also fire spontaneously, display variable responses to identical stimuli, and discharge in response to a wide variety of events. Such reasoning, discussed elsewhere in great detail (John, 1967b, 1971, 1972; John and Thatcher 1976), has led to the formulation of a statistical theory of neuronal representation of memory, which proposes that information is repre-
sented by the time sequence of deviations from random or baseline activity in extensive ensembles of neurons. In this formulation, learning is a cooperative process and the activity of any single neuron is informationally relevant only insofar as it contributes to the statistical features of the behavior of the ensemble. With the use of multiple chronically implanted movable microelectrodes, neuronal responses to discriminated stimuli have been studied (John and Morgades, 1969a, 1969b, 1969c). As the movable microelectrodes traversed across extensive anatomical domains, it was found that both the waveshape of the AER and the firing pattern of neuronal ensembles (average poststimulus histograms) were basically similar at different electrode positions. Analysis of variance was carried out comparing the variance within the responses to the two different signals at each region to the variance between responses to the same signal in many electrode positions. The difference in response to the two discriminated signals at any location was greater than the variability in response to a particular signal at different locations. This finding is illustrated in Figure 10. In other words, not only AER waveshapes but the firing patterns of neuronal groups in different brain regions displayed a characteristic temporal pattern whenever a particular learned signal was presented. Different anatomical regions were qualitatively homogeneous in their representation of a particular learned stimulus. However, signals with different significance elicited different patterns of ensemble responses. Finally, for three years my colleagues and I have been studying unit behavior in discrimination learning (Ramos, Schwartz, and John, 1974, 1976a-d; John, 1974; Ramos and Schwartz, 1976a, 1976b; Schwartz, Ramos, and John, 1976). In those studies we have utilized chronically implanted, movable microelectrodes to examine the responses of small groups of neurons and, more recently, well-isolated single neurons during correct responses and errors in tasks requiring
E.RoYJOHN
28 (Grand average)
2 Hz minus 8 tlz
8 tlz
2 liz
Average evoked
e,
AER P value (t·test) .00001 .0001 .001
.01
S.D.I:::!!::!.:::::::~::::::::
o
125 msec
0
0
125 msec
0
125 msec
(Grand average) 6
Post stimulus histograms
PSH P value (t-test) .0001
.001
.01
S.D.
o
125 msec
o
125 msec
10. (Top left) The top curve shows the grand average of the AERs elicited by the 2-Hz es across all electrode positions in the mapped region, while the lower curve shows the standard deviation (S.D.) of the group of AERs. (Bottom left) The top curve shows the grand average of the poststimulus histograms (PSHs) elicited by the 2-Hz es across the same electrode positions and the lower curve shows the standard deviation. (Top center) The grand average of the AERs elicited by the 8-Hz es and the corresponding S.D. (Bottom center) The grand average PSH elicited by the 8-Hz es and its S.D. (Top right) The top curve shows the difference waveshape resulting from the subtraction of the grand average AER elicited by the 8-Hz es from the grand average AER elicited by the 2-Hz es. The lower curve shows the p value, as computed by the t test, for each point of the difference wave. (Bottom right) The top curve shows the difference waveshape resulting from the subtraction of the grand average PSH elicited by the 8-Hz es from the grand average PSH elicited by the 2-Hz es. The lower curve shows the p value for each point of the difference. (Data from John and Morgades, 1969b.) FIGURE
differentiated behavioral responses to discriminated visual stimuli and also during differential generalization to ambiguous stimuli delivered to differentially trained cats. EPs and unit responses were simultaneously recorded from both cortical and subcortical microelectrodes in these cats. Cortical electrodes were located both in specific projection and in association areas. Computer-pattem-recognition techniques (John, Bartlett, Shimokochi, and Kleinman, 1973) and multidimensional scaling methods (Ramos, Schwartz, and John, 1976a-d) were used to classify single EPs from trials resulting in correct versus erroneous performance to the same CS
A
MODEL OF CONSCIOUSNESS
29
or from differential generalization trials in which two different behaviors were elicited by the same novel test stimulus. This classification procedure identified the EPs as belonging to one or another of the "readout modes" which reflected the activation of memories about different stimulus-response contingencies. Particular readout modes were found to be differentially correlated, at extremely high significance levels, with the subsequent behavioral performance. Therefore occurrence of a particular readout mode can be interpreted as evidence that the stimulus eliciting that electrophysiological and behavioral response was perceived as a signal with a particular significance. When the pattern-recognition procedure had identified the readout mode activated by each stimulus in the behavioral trial, the firing patterns of the simultaneously recorded unit activity corresponding to each readout mode were separately analyzed. We found two types of neurons in these analysis. One type showed an invariant average pattern of response to a given stimulus, no matter how it was perceived (Le., no matter what behavioral response ensued). These units might be described as "stimulus-bound," responding to the signal in the same way independent of perception. The second type of neuron showed one average temporal pattern of response during one readout mode and a different average temporal pattern of response during another readout mode. These units might be described as "gnostic" units, with a response pattern related to the perception rather than determined by the sensation. Such units showed great variability in their responses during a specified readout mode but displayed a characteristic and specific average response in each mode. Different units of this type in the same anatomical region showed closely similar average responses during the same mode, although their responses to single stimuli were poorly synchronized. These findings suggested an "ergodic" hypothesis, that is, the response to a single stimulus presentation
eliciting a specific readout mode averaged across the set of units of this type corresponded to the response of any unit of this type averaged across a set of stimulus presentations eliciting that specific readout mode. In view of these data, it would seem that the perception of a stimulus is mediated by the averaged temporal firing pattern of an ensemble or set of units of this type. The activity of any single unit is important only insofar as it contributes to the statistical behavior of the ensemble (Ramos, Schwartz, and John, 1976a, 1976b, 1976c, 1976d). Further, firing of any unit per se seems to have no unique informa-
30
E. Roy JOHN
o 0 0
•
AEP
•
8
~PSH
left response
~
•
2Qijms
:.
AEP
right • response
AvJVv
30jlV
20I~60jlV
~30)lV 20I~60jlV 200ms
FIGURE 11. On the left of this figure is shown the way in which single evoked responses obtained from the ends of several differential generalization trials were classified by use of a multidimensional scaling technique. Open circles represent EPs from trials resulting in eN.. The open and closed circles fall into domains which are clearly different, showing that a consistent waveshape or mode difference existed between the two sets of EPs referred as Mode 1 and Mode 2. In the middle of the figure are shown the average of the EPs in the two portions of the signal space. On top is the average of the EPs classified as Mode 1, while the average of the EPs classified as Mode 2 is on the bottom. These two averages display clear differences in a late component. On the right of the figure are shown two PSHs of the firing pattern of two different cells separated by computer spike-height discriminators during each mode. The two PSHs on the top show the firing pattern of these two cells during Mode 1 corresponding to the readout of memory related to CR,. The two bottom PSHs show that one of these cells displayed a completely different firing pattern during Mode 2 corresponding to the readout of memory relating to eN..
tional value. All units thus far observed not only fire "spontaneously" and show great variability in response to any stimulus but show a
differential response consisting of differences in graded temporal patterns, rather than an all-or-none behavior. As yet, we have not observed a single neuron which fired only during one memory readout mode and not during the other. An example of these findings, showing the different AER readout modes and the corresponding firing patterns of well-resolved units isolated by electronic discriminators, is presented in Figure 11.
A
MODEL OF CONSCIOUSNESS
VI.
31
BRAIN STIMULATION STUDIES
The electrophysiological observations summarized above provided the basis for a tentative picture of how memories were stored in the brain. The data suggested that many regions of the brain were directly or indirectly affected by presentation of the CS. Rapid informational transactions between these regions resulted in the establishment of a common mode of neural activity, reflected by the emergence of similar AER waveshapes in different brain regions as learning progressed. Presumably these similar electrophysiological processes represented not only the influence of the stimulus but acquired information about the stimulus-response contingencies as well as the emotional reactions and motivational state of the animal. Stimulus-bound features of this common mode of activity were represented by relatively short-latency, exogenous processes in the AER, while reactive features related to the meaning attributed to the stimulus were represented by longer-latency, endogenous processes. As the learned behavior became well consolidated, this anatomically extensive representational system acquired the capacity to reproduce an accurate facsimile of the common mode of activity in the absence of the conditioned stimulus, when situational factors or generalization stimuli activated the corresponding memory. These facsimiles of the effects of previous meaningful experiences were apparently generated in a cortico-thalamic-mesencephalic system, whence they propagated centrifugally to other regions of the brain. Investigation of the neuronal activity correlated with these slowwave and EP phenomena revealed that extensive neuronal populations displayed an average temporal firing pattern corresponding to the different memory readout modes. Single units displayed highly variable firing patterns to individual stimulus presentations but converged to the average ensemble behavior as their long-term behavior was observed. It appeared that the activity of any single neuron was informationally significant only insofar as it contributed to the ensemble statistics. No evidence was obtained for the existence of neurons whose firing uniquely denoted the identification of a specific CS, as would be required by theories attributing learning to the establishment of new connections in specific neuronal pathways. Electrical stimulation of the brain with appropriate temporal patterns of current provided a method of testing of these inferences to ascertain whether this model was sufficiently accurate to permit direct
32
E. Roy
JOHN
control of decision-making behavior by the manipulation of the average firing pattern of neuronal ensembles in differentially trained animals. A series of such experiments have been performed.
A. Rapid Transfer to Direct Electrical Stimulation of the Brain A group of cats was trained to discriminate between two different repetition rates of either visual (Vl and V2) or auditory (Al and Az) stimuli. After prolonged overtraining, these animals were subjected to electrical stimulation of the mesencephalic reticular formation (RF) with high-frequency current pulses at the same two repetition rates (RFl and RF2). Immediate discriminated response at high levels was displayed by most of the animals, and all animals in the study achieved criterion rapidly, indicating a very large transfer of training effect (John and Kleinman, 1975). Subsequent stimulation of the visual cortex (VIS), lateral geniculate (LG), medial geniculate (MG), and intralaminar nuclei of the thalamus (INT) showed less consistent results, but marked transfer was displayed by at least one animal for each of these structures. Electrical stimulation of this sort cannot conceivably activate particular pathways within the stimulated domain in a selective way, corresponding to the hypothetical circuitry postulated by theories which attribute learning to the establishment of new connections or facilitation of unique pathways. It seems necessary to concede that these gross stimuli can only accomplish massive coordinated or coherent firing patterns in the stimulated neuronal population.
B. Peripheral-Central Conflict After transfer of the discriminated behaviors from the peripheral modalities to direct electrical stimulation, "conflict" studies were carried out in which RF stimuli at either frequency were pitted against concurrent discordant visual or auditory stimuli of the other frequency (RFl vs. V2, RF2 vs. Vl t RFl vs. Az, RF2 vs. Al). Parametric tests were performed that explored the effects of ascending and descending series of RF current values. In every case, values of RF stimuli were found such that central stimulation at either frequency completely or substantially controlled the behavioral outcome, preempting the decision and effectively contradicting the concurrent visual or auditory cue. These results are illustrated in Figure 12.
A
MODEL OF CONSCIOUSNESS
33
In contrast, conflict studies using LG stimulation revealed that LG stimuli could control behavior only when contradicting visual cues and then only when the visual cues were at the slower repetition rate (Kleinman and John, 1975). These results suggested that RF stimuli provided uniquely intimate access to the decision-making system, effectively contradicting events in either peripheral sensory modality, while LG stimuli seemed rather to simulate visual sensations.
C. Perceptual Integration Processes of perceptual integration were investigated in the same animals. The cats were trained using a 4-Hz signal and a 2-Hz signal as the discriminated cues. After stimuli at this repetition rate had been established as effective signals whether delivered via vision, audition, or electrical stimulation of RF, VIS, LG, MG, or INT, temporal summation studies were carried out. Modalities of input were studied two at a time, in all possible combinations. In each study, simple stimuli at 4 Hz and 2 Hz were delivered in random order via the two selected modalities in random sequence. This series of simple stimuli served as control measures. Interspersed within this series were two kinds of compound stimuli. "In-phase" compound stimuli consisted of 2 Hz signals delivered simultaneously via the two modalities. These signals served as a control for total current and modality interaction effects. "Out-ofphase" compound stimuli consisted of 2-Hz signals delivered via the same two modalities but occurring 250 milliseconds out of phase. These signals served to test whether the animal would behave as if the two separate signals were perceived as individual 2-Hz cues, or whether they would be perceived as a unified 4-Hz cue. Although some pairs of modalities showed a much higher capability for cross-stimulus integration than did others, highly significant levels of integration were found for most of the combinations studied (John and Kleinman, 1975). Thus it appeared that informational events in many different brain regions could be integrated into a perceptual whole.
D. Loci Responsible for Perceptual Integration In order to identify the brain regions mediating this cross-modal integration, EEG and AER measurements were obtained during such
34
E. Roy Cat No.2 (++)
Cat No.6 (--)
N = 125
ti
= 123
100 c:
o
80
os
60
+:;
::>
E
40
I II
c:
~
.a
,.,
200
150
.a
N
g 100 c:
-'" o
(.)
c:
~
'"
c..
250
300
350
40
Cat No. 1 (--)
50
60
70
= 114
N
100
80
= 112
80
60
60
40
40
20
20
30
50
90
70
110
........•._.......
25
30
-'
...-----
~E :: /~;""""". 40
III
,
40
. 45
~--..
Cat NO.6 (--) ti = 74
N = 273
100
,,
.A.,
V, _ _ V2
Cat NO.3 (++)
, .....' ..
35
Stimulus (u.a)
'';::;
80
Cat No. 5 (--)
, .
100 80 60 40
/
20
c:
100
c:
-
80
c: '"
40
'" c..
20
o
(.)
~
120
140
50
ti
,
60
100
143
.'
.'
ti
100
80
= 164
80
,
60
40
/
---150
70
60
Cat No.4 (++)
Cat No.2 (++)
e
-
100
80
60
20
200
250
150
Stimulus (J.ta)
200
250
300
90
JOHN
A MODEL OF CONSCIOUSNESS
35
summation experiments. During out-of-phase stimulation at 2 Hz via two modalities which resulted in behavior as if a 4-Hz stimulus had been perceived, most brain regions displayed labeled responses and AER waveshapes showing only a 2-Hz component. The only exceptions to this were the VIS and the INT. These structures showed a marked 4Hz component when cross-modal integration occurred. This was particularly marked and consistent in the INT, no matter what pair of modalities was being integrated. These findings are illustrated in Figure 13A-E. These findings suggest that a system including the visual cortex and the INT reticular formation plays a particularly important role in cross-modal perceptual integration.
E. Role of Cortex and Thalamic Reticular Nuclei Other brain stimulation results confirm the apparent importance of the sensory cortex and the INT in perceptual processes. We have observed (John, 1963) that it is possible to interrupt stimulus-controlled behaviors by electrical stimulation of a wide variety of brain regions concurrent with the presentation of a sensory cue. At a sufficiently high current, the brain stimulus blocks performance to the CS (occlusion). In many cases, this occlusion persists for many seconds or even a few minutes after termination of the electrical stimulation. Recording during this "poststimulus absence" reveals high-voltage spindle waves in the INT independent of the locus of the electrical stimulation which produced the absence. Seizurelike afterdischarges can be produced in a variety of regions, especially in the FIGURE 12. Each graph shows the effectiveness with which stimulation of the mesencephalic RF at either of two frequencies (RFl and RF2 ) contradicted simultaneously presented visual stimuli (V2 and Vl' top) or auditory stimuli (Ao and At< bottom), plotted as a function of increasing current intensity. For cats 1, 3, and 6, frequency 1 was 4/sec and frequency 2 was 2/sec. For cats 2, 4, and 5, frequency 1 was 5/sec and frequency 2 was 1.81 sec. Solid lines show the outcomes when peripheral stimulation at the higher frequency (Vl in top graphs, Al in bottom graphs) was pitted against RF stimulation at the lower frequency (RF2 ), while the dotted lines show the outcomes when the higher-frequency stimulus was delivered to the RF. Cats 1, 5, and 6 were trained to perform an avoidanc~ avoidance discrimination (- -), while cats 2, 3, and 4 were trained to perform approachapproach discrimination (+ + ). N refers to the total number of conflict trials carried out in each cat, accumulated in three sessions for cats 2, 5, and 6 and four sessions for cat 1 (visual-RF conflict), and in three sessions for cat 2, four for cat 6, five for cat 4, and seven for cat 3 (auditory-RF conflict). (Data from Kleinman and John, 1975.)
E. Roy JOHN
36
CR as 2 LATERAL GENICULATE
_----\'----".--.....J._____'..... ~I_____'.I___'.~\____.\____.~.~.~.~I_____
STIMULUS VISUAL CORTEX
,.,.: ~~~{1\{4Vl)RM. 2+2
~
I~VRU'
2
- - - - 4 - -........""""'-.................................-......-------
FLICKER
f- 1 SEC -lj VISUAL CORTEX
FLICKER
.,
\
"
,
MRF STIMULUS
\
\
I
I
•
\
I
I
I
\
•
I
\
\
I
\
\
I
, .
I
A
37
MODEL OF CONSCIOUSNESS
o
.v'J.'1~1 ~~"'1~"MV;-." BEGINS TO MAKE CR AS 4
MOVES AWAY FROM LEFT LEVER
PERFORMS CR AS 2 ON RIGHT LEVER
FLICKER
~1 SEC~ MRF
CAR LEFT (2+2 • 4) CAR RIGHT (2+2=2) FIGURE 13. Data from a cat 7 TRIALS N-187 trained to press a, lever on the 6 TRIALS N-I04 left side of a work panel to avoid foot shock when a 4/sec es was AVERAGE A .f\ presented in any of a number of EWlCED modalities, including electrical RESPONSE stimulation of the brain, and to press a lever on the right side of the work panel to avoid foot VARIANCE shock when a 2/sec es was preL2~ IIS.J...2~ IoJSoI sented in any of the same mod110 R. IIRF R.IlAF 110 alities. 9O)lA 6O"A 60.. A 9O"A (A-D) Each of these figures shows the activity recorded from the intralaminar-midline thalaDIFFERENCE mus under two conditions: WAVE (Top) Presentation of a 2/sec es (CAR LEFT -CAR RIGHT) in one modality (MI) plus a 2/sec es in a second modality (M2) delayed by Z50ms with respect to t _TEST p • Ol--l/'r'),.,~ . MI resulting in performance of E the behavior appropriate to a 4/ sec es: (Bottom) Presentation of the same compound stimuli resulting in behavior appropriate to a 21sec es. A-D differ in the modalities of es presentation. (A) MI = LG es; M2 = VIS es. (B) MI = peripheral flicker es; M2 = VIS es. (C) MI = peripheral flicker es; M2 == mesencephalic RF es. (D) MI = peripheral flicker es; M2 = mesencephalic RF es. Note that when the 2/sec ess in the two modalities were effectively merged, a 4/sec rhythm is prominent in the intralaminar record. If the two signals were not effectively combined, a 2/sec rhythm dominates the record. This is particularly striking during behavioral vacillation (D). (E) stimulation as above with MI = medialis dorsalis and M2 = right mesencephalic RF. All data recorded from right VIS, bipolar derivation. (Top left) AER when effective merging of the two 2/sec ess occurs as indicated by behavioral performance of the animal appropriate to a 4/sec es. Note that at the cortex, essentially four afferent volleys per second are occurring. Below is the variance. (Top right) Stimulation identical as top left, however, the two signals were not effectively combined as indicated by the behavioral performance of the animal appropriate to a 21sec es. Note that at the cortex essentially two afferent volleys per second are occurring, the EP to the right mesencephalic (RF) stimulus is barely discernible. Below is the variance. (Bottom) Difference wave obtained by subtraction of the average EP of trials in which summation failed tooccur (top'right) from the average EP of trials in which summation occurred (top left). The I-test wave at the bottom shows that the difference between the two average EPs is statistically significant.
...
~
38
E.RoYJoHN
limbic system, by such a procedure. However, by careful adjustment of current values and recording from the stimulated region as well as the INT, it is possible to confirm that such absences can occur when there is no indication of electrical seizure in the stimulated regions. Presentation of the es during the period of intralaminar spindles may elicit an orientation reflex, but the cat shows no sign of comprehending the signal. As the spindle waves vanish, the animal often gives a startle response and looks about in an agitated fashion, as if awakening. Thus, interference with the activity of the INT serves to disrupt perception. Finally, we have used electrical stimulation to explore the informational significance of early and late components in the AER from visual cortex, by phase-locking high-frequency current pulses to the es so as to coincide with the short-latency exogenous or longerlatency endogenous processes (John, 1967b). Electrical stimulation of sensory cortex during the exogenous components produced no disruption of discriminative responses, while stimulation with identical current parameters timed to coincide with endogenous processes (80110 milliseconds) totally abolished eRs. During such stimulation, the cat would orient toward the es but would behave as though it possessed no cue value. These findings show that the late components of the cortical AER, generally considered to reflect the influence of the nonsensory-specific mesencephalic and thalamic reticular systems, are essential for identification of the meaning of a stimulus. It is not clear whether these results should be interpreted as indicating that the essential functions are performed at the cortical level, or whether the cortical stimulation precludes the corticothalamic outflow necessary to establish an interactive transaction between these sensory-specific and nonsensory-specific domains of the brain.
VII.
THEORETICAL DISCUSSION OF ELECTROPHYSIOLOGICAL EVIDENCE
The evidence which has been summarized suggests that information, past or present, is represented in the brain by a statistical process, the average spatiotemporal pattern of activity in anatomically extensive neuronal populations. The activity of the single neuron is not informationally significant except insofar as it contributes to the activity of the ensemble. The same information can be represented in diverse regions, with a varying signal-to-noise ratio (SIN). In any region, some cells
A MODEL OF CONSCIOUSNESS
39
appear to be stimulus-bound, displaying the same average firing pattern to stimuli independent of how they are perceived, although they may display different response patterns to different stimuli. Such cells would appear to be relatively reliable reporters of sensation, in terms of their ability to construct reproducible average firing patterns characteristic for each different stimulus, in spite of the short-term variability of their responses. Other cells in the same regions display average patterns of response to the same stimulus which are more reactive, depending upon the meaning attributed to the signal. These latter cells seem to be involved in perceptual processes and in the storage of memories about the stimulus-response contingencies. The mixture of these two types of cells varies from region to region, producing variable SIN for both exogenous and endogenous processes, that is, making different relative contributions to sensation and to perception. Consciousness, subjective experience, the concept of self, and selfawareness, while representing successively higher orders of information, must nonetheless also be mediated in the same statistical fashion. There is no compelling evidence or logical argument to suggest that these higher levels of information are represented by qualitatively different neuronal processes. The content of consciousness is the sum of all informational processes in all the various functional systems of the brain. The information in each area comprises a coherent temporal pattern. The outflow of this coherent pattern to other brain regions constitutes afferent input elsewhere. The result of these rapid multidirectional transactions of information between different regions establishes a common mode of activity shared by many anatomical regions, with the relative contribution (SIN) of each type of activity varying from region to region as a function of its afferent connectivities. These regional messages, each with its characteristic pattern and 51 N, converge upon the cortical association areas and via collateral pathways and corticofugal pathways upon the INT, the mesencephalic RF, and the limbic system. In man and other mammals, consciousness depends upon integrity of the thalamic and mesencephalic reticular systems. Lesion of these systems produces long-lasting or permanent coma (Moruzzi and Magoun, 1949). Yet the fact that recovery from such coma sometimes ensues or that multistage lesions of these regions fail to produce a comatose animal or to interfere with information processing (Adametz, 1959; Chow and Randall, 1964) suggests that the process is distributed and can be effectively mediated by other brain regions under appropriate circumstances.
40
E.RoY]OHN
Similarly, although voluminous data from neurological clinics attests to the catastrophic effects of brain damage in certain regions upon specific perceptions or other higher intellectual functions because of head trauma or cerebrovascular accidents, yet the literature also abounds with evidence of functional compensation for a good part of such damage with time although the damaged tissue was irreversibly destroyed. A substantial body of evidence indicates that retention of preoperatively learned tasks often occurs in mammals when brain areas relevant to the task are removed serially in multiple-stage operations, although identical lesions made in a single-stage operation may abolish performance and prevent reacquisition. Interestingly, if animals subjected to multiple-stage lesions of the VIS are deprived of visual experience between surgeries, the resulting visual deficit is comparable to the effects of a one-stage ablation. Recently it was demonstrated that rats permitted unrestrained movement in a patterned visual environment during the interval between two-stage lesions of the VIS can rapidly relearn a pattern discrimination established prior to surgery, while rats passively transported through the same environment fail to do so (Dru, Walker, and Walker, 1975). These findings suggest that functional reorganization must occur during a multiple-stage procedure, that sensory stimulation in the damaged modality is crucial for such reorganization, and that cross-modal transfer may facilitate this process. Phenomena such as compensation for brain damage and absence of functional deficits after multiple-stage lesions constitute strong evidence that the brain possesses alternative methods for performing many functions which are mediated by some particular structure under ordinary circumstances. Perhaps it is possible to reconcile the evidence that information about particular kinds of experience appears to be localized-because local lesions can cause such discrete dysfunctions as alexic agnosias or loss of previously learned discriminations-with the evidence adduced above that information is distributed throughout many different anatomical regions. The electrophysiological data indicate that the endogenous processes reflecting particular perceptual and cognitive functions have a very widespread distribution. The fact that the values found for these processes span a range of 1,000 indicates great quantitative differences between anatomical regions in the density or intensity of representation of a memory. Perhaps under normal circumstances, the brain of an individual requires some threshold value for the SIN to be exceeded in order for information represented by the correspond-
A
MODEL OF CONSCIOUSNESS
41
ing neural activity to be functionally useful. If that threshold value is surpassed by only one particular brain region, damage to that region will produce impairment of that function, while such symptoms should not result from damage of any other region. Nonetheless, the relevant information is available in many other places. Perhaps if the threshold value for the SIN could be lowered, restoration of the impaired function might be achieved even though irreversible damage had been sustained by the region which previously achieved the highest SIN. This reasoning seems particularly plausible if we consider that the greater informational reliability of a high SIN, as life experiences accumulated, would tend to establish functional dependence upon the region displaying the highest SIN and a learned threshold setting which would reject information from regions with a lower SIN. Thus a learned functional inhibition might even be established which prevented such alternate regions from resuming functional utility in the event that the region usually mediating that function were damaged. These speculations offer a way to reconcile the large body of evidence about specific agnosias and other deficits, which seemingly involve memory and consciousness and result from localized brain damage, with the apparently contradictory electrophysiological findings in our studies. For example, perhaps the SIN for activity related to the perception of letters and numbers is highest in a particular cortical region and the threshold in the normal brain is usually set to reject lower SINs for that activity. Such lower SINs might exist in other regions. Thus although information relevant to the perception of letters and numbers is available in those regions, damage there will not result in alexic agnosia, nor can those regions sustain such perception alone if the salient cortical field is damaged. While I have no evidence at present that these speculations are correct, they provide an attractive working hypothesis-attractive not only because thus no contradiction need exist between two bodies of data (lesion and electrophysiological), both of which reflect real aspects of brain function, but also because were this hypothesis correct, much functional deficit due to brain damage which we now consider irreversible might be reversed by procedures which lower the relevant thresholds or block the learned functional inhibitions. Perhaps the apparent dependence of consciousness upon an intact reticular formation can be similarly explained. The convergence of information from every sensory region and many other functional systems upon the reticular formation create here a uniquely favorable
42
E. Roy JOHN
anatomical design for integration to take place. Anatomy also favors the effective and widespread distribution of activation from the RF throughout a host of other brain regions. These factors may well give the RF of the mesencephalon and thalamus a unique ability to achieve a high SIN both for convergent afferent input and divergent efferent output. The brain may well come to rely upon this high SIN, setting thresholds as a result of experience such that it comes to depend upon these regions both for integrative processes and for the maintenance of consciousness, in the sense of organizing excitability in the system so that sufficiently high levels of coherence may be achieved. Single-stage lesions in this system, as attested by a huge volume of literature, do result in long-lasting or permanent loss of consciousness. Perhaps the ability of the brain to maintain consciousness and integrative activity when this sytem is destroyed in multiple stages is due to systematic increases in the relative SIN of other regions participating in the same functions, increased absolute coherence in those regions as the functional inhibition is extinguished, and lowering of the threshold for acceptable SIN as the series of lesions is inflicted. Thus, as with many other functions, it may well be fruitless to ask whether any brain region is uniquely responsible for consciousness and subjective experience. These functions are probably distributed over a widespread anatomical substrate, every region of which makes a contribution to the overall process and many of which may be capable of sustaining the process if damage occurs elsewhere in the system. Is this an evasion of the issue? Is the whole system conscious? Under normal circumstances, whether or not "backup" systems exist, is there some circumscribable system which mediates consciousness of the fluid patterns of information and the continuity of subjective experience? What is the nature of the process constituting the intimate basis of the emergent property of subjective experience, which transcends the activity of the constituent elements of the system? The electrophysiological data reviewed above indicate that sensations are encoded as organized spatiotemporal patterns of average activity in stimulus-bound neurons whose density varies from region to region of the brain, with a concomitant variation in SIN. Perceptions are similarly encoded as average spatiotemporal patterns, but in ensembles of neurons capable of responding to the arrival of information from the stimulus-bound ensemble with a firing pattern creating a facsimile of responses previously displayed to other events. These intermingled firing patterns, rapidly evolving into a common mode which represents
A
MODEL OF CONSCIOUSNESS
43
sensations and perceptions in many modalities, as well as information about drive levels from the hypothalamus and about affective state from the limbic lobe, are distributed throughout widespread brain regions but converge most intensively upon the "centrencephalic" thalamic and mesencephalic reticular systems. Outflow from these systems feeds back upon cortical and thalamic regions which contributed high SIN to the afferent barrage upon the centrencephalic system, further enhancing the SIN of the reverberating cortico-thalamic-centrencephalic pattern which emerges. As a result of this reverberation, unusually high levels of local coherence are achieved in the participating ensembles. These ensembles do not become fully synchronized but do achieve higher coherence levels and higher SIN than could occur without this feedback process. As these coordinated temporal patterns of firing occur in the densely packed cells of the centrencephalic system, the membranes of the participating cells are depolarized and ionic shifts occur. Potassium concentration increases in the extracellular space, and ionic binding probably occurs to mucopolysaccharide filaments and on the surfaces of glial cells. Complex gradients of charge are thereby established, with distributions which depend upon the spatiotemporal coherence patterns in the neuronal ensemble. One can envisage a complex, threedimensional volume of isopotential contours, with a topology encompassing portions of neuronal membranes, glial membranes, and extracellular binding sites. Let us call this set of isopotential contours or convoluted charge surfaces a hyperneuron. Every representational system has a corresponding particular distribution of energy, a unique hyperneuron. The special features of a particular hyperneuron will be determined by the statistical processes in local ensembles which established the set of coherent spatiotemporal patterns within this volume of neural tissue. The contribution of any individual cell to a hyperneuron will be insignificant. Ensembles of neurons in regenerative circuits will contribute a stable component to all or most hyperneurons, while ensembles with lower positive feedback will make more variable contributions. Thus one can envisage sequences of hyperneurons which would display stable, invariant features as well as modulated, reactive features. We postulate that the property of consciousness emerges from the cooperative interaction of neuronal populations, resulting in the establishment of hyperneurons whose characteristics transcend the features of the cellular constituents of the ensemble, which serve as responsive,
44
E. Roy JOHN
charge-distributing elements. The content of subjective experience is the momentary contour of the hypemeuron. As the contour of the hypemeuron fluctuates in time, the content of consciousness varies, so that the invariant features of the hypemeuron constitute the "1" of selfawareness, while the variable features are the fleeting "here-now" of the momentary subjective experience. It may be that only the centrencephalic system and perhaps the limbic and cortical neuronal masses can sustain the hyperstructure required for a hypemeuron to develop. On the other hand, the brain may contain an extended hypemeuron with many lobules located in different anatomical regions and all interconnected with each other and modulating the centrencephalic-cortical hypemeuron. A certain critical mass and critical density of elements may be a prerequisite for tissue to be capable of sustaining a hypemeuron, or this cooperative process may be a property of any mass of neural tissue. It is conceivable that a "rudimentary hypemeuron" can exist in any form of living matter, with a complexity of experience limited by the number of energy states which the matter can attain. A priori, there seems no compelling reason to insist that this cooperative process is restricted only to certain types of brains or to certain types of tissue. We simply do not know enough about the essential features upon which this emergent property depends to be arbitrary about which organisms can and cannot possess it. Finally, we have postulated that mental experiences are produced by and consist of cooperative electrochemical phenomena which arise within volumes of neural tissue. Yet the crucial features of neuronal masses for the production of subjective experience may not depend upon the neurons themselves. Were the hypemeuron postulate accurate, it would not be clear whether the property of mental or subjective experience arose from the action of the charge contours upon the neurons present within that complex field or whether the subjective experience were an intimate consequence of the energy distribution itself. Neurons may not possess any inherent quality essential for this transformation but may be uniquely well suited for the production of a wide variety of improbable distributions of energy. Were it possible to achieve comparable distributions of energy without neurons-in other words, to simulate' a "neuron-free" hypemeuron-perhaps quite the same subjective experience would arise. Subjective experience may actually be a property of a certain level of organization in matter. This article began with a series of questions about the nature of
A
MODEL OF CONSCIOUSNESS
45
subjective experience. A body of experimental evidence has been reviewed which provides some insight into the brain mechanisms which mediate information representation, memory retrieval, and decision making and suggests possible answers to those questions. In view of the distributed statistical nature of the representational processes revealed by such experimental studies, it seemed necessary to postulate that subjective experience is the product of a cooperative process involving both cellular and extracellular constituents of neural tissue, most probably in the centrencephalic system, to which we assigned the label of hyperneuron. This postulate can be subjected to test. If it is correct, then there must exist physical parameters of the energy distribution which will alter the content of subjective experience when manipulated. A major experimental task for us must be to ascertain what features of organized energy in neural tissue produce subjective experience. As these essential features become more apparent, it will be possible to develop a better-informed basis for evaluating whether this emergent property is necessarily limited to brains with certain architectonic specifications, exists in any neuronal systems, is a general property of living matter, or might arise in a sufficiently organized system of energy. Although these questions will be extremely difficult to answer, I am confident that answers will be provided, and relatively soon. A great deal has been learned about informational processes in the brain, and additional information is steadily accumulating. Something very much like the postulated hyperneuron must exist, and it is only a question of time until we understand it. There is one aspect of this set of issues, however, which I still find totally baffling. Reality is a continuously fluctuating distribution of physical energy in different frequency domains, located in various regions of space. This energy, impinging upon the receptors of the body, causes the firing of neurons in afferent pathways and ultimately, if our postulate is correct, produces a modulation in the contours of a hyperneuron. Subjective reality, produced by this hyperneuron, is a constellation of vivid colors, shapes, sounds, and images synthesized from these neuronal patterns, reflecting the energy spectrum of reality in a reproducible but not literal fashion. Reality is not our experience of reality. How is this transformation accomplished? What aspect of the hyperneuron's dimensionality might produce the rich, diverse qualities of this abstraction from reality? Perhaps this is the fundamental question, and as yet I fail to discern the faintest glimmering of an answer.
46
E. Roy
JOHN
ACKNOWLEDGMENTS
This work has been supported in part by Grant #MH20059 from the National Institutes of Health and grant #BMS 7502819 from the National Science Foundation. Dr. John is a Career Scientist of the Health Research .council of the City of New York under Grant #1-752. REFERENCES ADAM, G., ADEY, W. R., AND PORTER, R. W. Interoceptive conditional response in cortical neurones. Nature, 1966,209,920-921. ADAME1Z, J. H. Rate of recovery of functioning in cats with rostral reticular lesions. Journal of Neurosurgery, 1959,16,85-97. ASRATYAN, E. A. Changes in the functional state and pattern of electrical activity in cortical areas involved in the establishment of conditioned connection. Proceedings of XXIII International Congress of Physiological Sciences (Tokyo), 1965,4, 629-636. BARLOW, J. S., MORRELL, L.,AND MORRELL, F. Some observations on evoked responses in relation to temporal conditioning to paired stimuli in man. Proceedings of International Colloquium on Mechanisms of Orienting Reactions in Man (Bratislava-Smolenice, Czechoslovakia), 1967. BARTLETT, F., AND JOHN, E. R. Equipotentiality quantified: The anatomical distribution of the engram. Science, 1973, 181, 764-767. BARTLETT, F., JOHN, E. R., SHIMOKOCHI AND KLEINMAN, D. Electrophysiological signs of readout from memory. II. Computer classification of single evoked potential waveshapes. Behavioral Biology, 1975, 14, 409-449. BEGLEITER, H., AND PLA1Z, P. Modifications in evoked potentials by classical conditioning. Science, 1969, 166, 769-771. BUCHWALD, J. S., HALAS, E. S., AND SCHRAMM, S. Progressive changes in efferent unit responses to repeated cutaneous stimulation in spinal cats. Journal of Neurophysiology, 1965,28, 200-215. BURES, J. Discussion. In D. P. KIMBLE (Ed.), Anatomy of memory. Palo Alto: Science Behavior Books, 1965, pp. 49-50. BUREs,J., AND BURESOVA, O. Plasticity at the single neuron level. Proceedings of the XXIII International Congress of Physiological Sciences (Tokyo), 1965,4, 359--364. BURE~, J., AND BURESOVA, O. Plastic changes of unit activity based on reinforcing properties of extracellular stimulation of single neurons. Journal of Neurophysiology,
1967,30,98-113. J., AND BURESOVA, O. Plasticity in single neurons and neural populations. In G. Hom and R. A. Hinde (Eds.), Short-term changes in neural activity and behavior. London and New York: Cambridge University Press, 1970, pp. 9-35. CHESLER, P. Maternal influence in learning by observation in kittens. Science, 1969, 166, 901-903. CHOW, K. L., AND RANDALL, W. Learning and retention in cats with lesions in reticular formation. Psychonomic Science, 1964, 1, 259--260. CORNING, W. c., DYAL, J. A., AND WILLOWS, A. O. D. Invertebrate learning (Vol. 1). New York: Plenum Press, 1973. DRU, D., WALKER, J. P., AND WALKER, J. B. Self-produced locomotion restores visual capacity after striate lesions. Science, 1975,187,265-266. DUMENKO, V. N. The electrographic study of relationships between various cortical areas
BURE~,
A
MODEL OF CONSCIOUSNESS
47
in dogs during the elaboration of a conditioned reflex stereotype. In l. N. KNIPST (Ed.), Contemporary problems of electrophysiology of the central nervous system. Moscow: Academy of Science, 1967, pp. 104-112. GALAMBOS, R, AND SHEAtz, G. C. An electroencephalograph study of classical conditioning. American Journal of Physiology, 1962, 203, 173-184. GLlVENKO, E. V., KOROL'KOVA, T. A., AND KUZNETSOVA, G. D. Investigation of the spatial correlation between the cortical potentials of the rabbit during formation of a conditioned defensive reflex. Fizicheskii Zhurnal SSSR Sechenova, 1962, 48, 1026. GRINBERG-ZYLBERBAUM, J., CARRANZA, M. B., CEPEDA, G. V., VALE, T. c., AND STEINBERG, N. N. Caudate nucleus stimulation impairs the processes of perceptual integration. Physiology and Behavior 1974, 12; 913-918. HORI, Y., AND YOSHII, N. Conditioned change in discharge pattern for single neurons of medial thalamic nuclei of cat. Psychological Report, 1965,16, 241. JASPER, H. H., RICCI, G., AND DOANE, B. Microelectrode analysis of cortical cell discharge during avoidance conditioning in the monkey. Electroencephalography and Clinical Neurophysiology Supplement, 1960,13, 137-155. JOHN, E. R Higher nervous functions: Brain functions and learning. Annual Review of Physiology, 1961, 23, 451. JOHN, E. R. Neural mechanisms of decision making. In W. S. FIELDS AND W. ABBOT (Eds.), Information storage and neural contra!. Springfield: Thomas, 1963, pp. 243-282. JOHN, E. R. Electrophysiological studies of conditioning. In G. C. QUARTON, T. MELNECHUK, AND F. O. SCHMITT (Eds.), The neurosciences: A study program. New York: Rockefeller University Press, 1967a, pp. 690-704. JOHN, E. R. Mechanisms of memory. New York: Academic, 1967b. JOHN, E. R. Brain mechanisms of memory. In J. McGAUGH (Ed.), Psychobiology. New York: Academic, 1971, pp. 199-283. JOHN, E. R. Switchboard versus statistical theories of learning and memory. Science, 1972, 177, 850-864. JOHN, E. R. Cellular mechanisms in conditioning. Paper presented at International Union of Physiological Sciences, New Delhi, 1974. JOHN, E. R. BARTLETT, F., SHlMOKOCHI, M., AND KLEINMAN, D. Neural readout from memory. Journal of Neurophysiology, 1973,36, 893-924. JOHN, E. R., CHESLER, P., BARTLETT, F., AND VICTOR, l. Observation learning in cats. Science, 1968, 159, 1489-1491. JOHN, E. R., AND KILLAM, K. F. Electrophysiological correlates of avoidance conditioning in the cat. Journal of Tharmacological and Experimental Therapeutics, 1959,125,252. JOHN, E. R., AND KILLAM, K. F. Electrophysiological correlates of differential approachavoidance conditioning in the cat. Journal of Nervous and Mental Diseases, 1960,131, 183. JOHN, E. R., AND KLEINMAN, D. Stimulus generalization between differentiated visual, auditory and central stimuli. Journal of Neurophysiology, 1975,38, 1015-1034. JOHN, E. R., LEIMAN, A. L., AND SACHS, E. An exploration of the functional relationship between electroencephalographic potentials and differential inhibition. Annals of New York Academy of Sciences, 1961,92, 1160-1182. JOHN, E. R, AND MORGADES, P. P. Neural correlates of conditioned responses studied with multiple chronically implanted moving microelectrodes. Experimental Neurology, 1969a, 23, 412-425. JOHN, E. R., AND MORGADES, P. P. The pattern and anatomical distribution of evoked potentials and multiple unit activity elicited by conditioned stimuli in trained cats. Communications in Behavioral Biology, 1969b, 3, 181-207. JOHN, E. R., & MORGADES, P. P. A technique for the chronic implantation of multiple
48
E. Roy JOHN
movable micro-electrodes. Electroencephalography and Clinical Neurophysiology, 1969c, 27, 205-208. JOHN, E. R., RUCHKIN, D. S., LEIMAN, A., SACHS, E., AND AHN, H. Electrophysiological studies of generalization using both peripheral and central conditioned stimuli. Proceedings of the XXIII International Congress of Physiological Sciences (Tokyo), 1965, 618-627. JOHN, E. R., RUCHKIN, D. S., AND VILLEGAS, J. Signal analysis of evoked potentials recorded from cats during conditioning. Science, 1963,141,429-431. JOHN, E. R., RUCHKIN, D. S., AND VILLEGAS, J. Signal analysis and behavioral correlates of evoked potential configuration in cats. Annals of New York Academy of Sciences, 1964, 112, 362-420. JOHN, E. R., SHIMOKOCHI, M., AND BARTLElT, F. Neural readout from memory during generalization. Science, 1969,164, 1534-1536. JOHN, E, R., AND THATCHER, R. W. Integrative neuroscience. New Jersey: Lawrence Erlbaum Associates, 1976, in press. JOHNSTON, V. 1., AND CHESNEY, G. 1. Electrophysiological correlates of meaning. Science, 1974, 186, 944-946. KAMIKAWA, K., MciLWAIN, J. T., AND ADEY, W. R. Response patterns of thalamic neurons during classical conditioning. Electroencephalography and Clinical Neurophysiology, 1964, 17, 485-496. KILLAM, K. R., AND HANCE, A. J. Analysis of electrographic correlates of conditional responses to positive reinforcement: I. Correlates of acquisition and performance. Proceedings of the XXIII International Congress of Physiological Sciences (Tokyo), 1965, 4, 1125. KLEINMAN, D., AND JOHN, E. R. Contradiction of auditory and visual information by brain stimulation. Science, 1975, 187, 271-272. KLINKE, R., FRUHSTORFER, H., AND FINKENZELLER, P. Evoked responses as a function of external and stored information. Electroencephalography and Clinical Neurophysiology, 1968, 26, 216-219. KNIPST, I. N. (Ed.). Spatial synchronization of bioelectrical activity in the cortex and some subcortical structures in rabbit's brain during conditioning. In Contemporary problems of electrophysiology of the central nervous system. Moscow: Academy of Science, 1967, pp. 127-137. KOROL'KOVA, T. A., AND SHVETS, T. B. Interrelation between distant synchronization and steady potential shifts in the cerebral cortex. In I. N. KNIPST (Ed.), Contemporary problems of electrophysiology of the central nervous system. Moscow: Academy of Science, 1967, pp. 16~167. LEIMAN, A. 1. Electrophysiological studies of conditioned responses established to central electrical stimulation. Doctoral thesis, University of Rochester, Rochester, New York, 1962. LEIMAN, A. 1., AND CRISTIAN, C. N. Electrophysiological analyses of learning and memory. In. J. A. DEUTSCH (Ed.), The physiological basis of memory. New York: Academic Press, 1973, pp. 125-173. LIBERSON, W. T., AND ELLEN, P. Conditioning of the driven brain wave rhythm in the cortex and the hippocampus of the rat. In J. WORTIS (Ed.), Recent advances in biological psychiatry. (Vol. 2). New York: GRUNE & STRATTON, 1960. LINDSLEY, D. F., CARPENTER, R. S., KILLAM, E. K., AND KILLAM, K. F. EEG correlates of behavior in the cat: I. Pattern discrimination and its alteration by atropine and LSD25. Electroencephalography and Clinical Neu~ophysiology, 1968,24, 497-513. LIVANOV, M. N. Spatial analysis of the bioelectric activity of the brain. Proceedings of the XXII International Congress of Physiological Sciences (Leiden), 1962, 899-907.
A
MODEL OF CONSCIOUSNESS
49
LIVANOV, M. N. The significance of the distant brain potential synchronization for realization of temporal connections. Proceedings of the XXIII International Congress of Physiological Sciences (Tokyo), 1965,4, 600-612. LIVANOV, M. N., AND POLIAKOV, K. L. The electrical reactions of the cerebral cortex of a rabbit during the formation of a conditioned defense reflex by means of rhythmic stimulation. Izvestiya Akademiya Nauk. USSR Series Biology, 1945,3,286. MAJKOWSKI, J. Electrophysiological studies of learning in split-brain cats. Presented at Electroencephalography Meeting (California), 1966. MAJKOWSKI, J. Electrophysiological studies of learning in split-brain cats. Electroencephalography and Clinical Neurophysiology, 1967,23, 521-53l. MORGAN, C. T., AND STELLAR, E. Physiological psychology. New York: McGraw-Hill, 1950. MORRELL, F. Effect of anodal polarization on the firing pattern of single cortical cells. Annals of New York Academy of Science, 1961a, 92, 86~76. MORRELL, F. Electrophysiological contributions to the neural basis of learning. Physiological Review, 1961b, 41, 443. MORRELL, F. Electrical signs of sensory coding. In G. C. QUARTON, T. MELNECHUK, AND F. O. SCHMITT (Eds.), The neurosciences: A study program. New York: Rockefeller University Press, 1967, pp. 452-469. MORUZZI, G., AND MAGOUN, H. W. Brain stem reticular formation and activation of the EEG. Electroencephalography and Clinical Neurophysiology, 1949,1, 455-473. O'BRIEN, J. H., AND Fox, S. S. Single-cell activity in cat motor cortex: I. Modifications during classical conditioning procedures. Journal of Neurophysiology, 1969a, 32, 267284. O'BRIEN, J. H., AND Fox, S. S. Single-cell activity in cat motor cortex: II. Functional characteristics of the cell related to conditioning changes. Journal of Neurophysiology, 1969b,32, 285-296. OLDS, J., DISTERHOFf, J. F., SEGAL, M., KORNBLITH, c., AND HIRSCH, R. Learning centers of the rat brain mapped by measuring latencies of conditioned unit response. Journal of Neurophysiology, 1972,35, 202-219. OLDS, J., AND HIRANO, T. Conditioned responses of hippocampal and other neurons. Electroencephalography and Clinical Neurophysiology, 1969,26, 159-166. OLOS, J., AND OLOS, M. E. Interference and learning in paleocortical systems. In J. F. DELAFRESNAYE, A. FESSARD, R. W. GERARD, AND J. KORNORSKI (Eds.), ClOMS symposium on brain mechanisms and learning. Oxford: Blackwell, 1961. PICTON, T. W., AND HILLYARD, S. A. Human auditory evoked potentials: II. Effects of attention. Electroencephalography and Clinical Neurophysiology, 1974,36, 191-199. PICTON, T. W., HILLYARD, S. A., AND GALAMBOS, R. Cortical evoked responses to omitted stimuli. In M. N. Livanov (Ed.), Major problems of brain electrophysiology. Moscow: U.S.S.R. Academy of Sciences, 1973. RAMOS, A., AND SCHWARTZ, E. Observation of assimilation of the rhythm at the unit level in behaving cats. Physiology and Behavior, 1976a, in press. RAMOS, A., AND SCHWARTZ, E. Observation of frequency specific discharges at the unit level in conditioned cats. Physiology and Behavior, 1976b, in press. RAMOS, A., SCHWARTZ, E., AND JOHN, E. R. Unit activity and evoked potentials during readout from memory. Presented at XXVI International Congress of Physiological Sciences (New Delhi), 1974. RAMOS, A., SCHWARTZ, E., AND JOHN, E. R. Evoked potential-unit relationships in behaving cats. Brain Research Bulletin, 1976a, in press. RAMOS, A., SCHWARTZ, E., AND JOHN, E. R. Cluster analysis of evoked potentials from behaving cats. Behavioral Biology, 1976b, in press.
E.RoYJOHN
50
RAMOS, A., SCHWAR1Z, E., AND JOHN, E. R. An examination of the participation of neurons in readout from memory. Brain Research Bulletin, 1976c, in press. RAMOS, A., SCHWAR1Z, E., AND JOHN, E. R. Differential neuronal responses during conditioning in cats. Science, 1976d, in press. RIGGS,1. A., AND WHlTILE, P. Human occipital and retinal potentials evoked by subjectively faded visual stimuli. Vision Research, 1967, 7, 441-451. RUCHKIN, D. S., AND JOHN, E. R. Evoked potential correlates of generalization. Science, 1966, 153, 209-211. RUSlNov, V. S. Electroencephalographic studies in conditioned reflex formation in man. In M. A. B. BRAZIER (Ed.), The central nervous system and behavior, New York: JOSIAH MAClIl, JR. Foundation, 1959. SAKHUILINA, G. T., AND MERZHANOVA, G. K. Stable changes in the pattern of the recruiting response associated with a well-established conditioned reflex. Electroencephalography and Clinical Neurophysiology, 1966,20, 50-58. SCHWAR1Z, E., RAMOS, A., AND JOHN, E. R. Single cell activity in chronic unit recording: A quantitative study of the unit amplitude spectrum, Brain Research Bulletin, 1976, in press. SUTTON, S., TUETING, P., ZUUIN, J., AND JOHN, E. R. Information delivery and the sensory evoked potential. Science, 1967, 155, 1436-1439. THOMPSON, R. F., PATTERSON, M. M., ANDTEYLER, T. J. The neurophysiology of learning. Annual Review of Psychology, 1972,23, 73-104. TRAVIS, R. P., JR., AND SPARKS, D. L. Unitary responses and discrimination learning in the squirrel monkey. Physiological Behavior, 1968,3, 187-196. TRAVIS, R. P., JR., AND SPARKS, D. L. Changes in unit activity during stimuli associated with food and shock reinforcement. Physiology and Behavior, 1967,2,171-177. TRAVIS, R. P., JR., HOOTEN, T. F., AND SPARKS, D. 1. Single unit activity related to behavior motivated by food reward. Physiology and Behavior, 1968,3, 309-318. TRAVIS, R. P., JR., SPARKS, D. 1., AND HOOTEN, T. F. Single unit response related to sequences of food motivated behavior. Brain Research, 1968,7, 455-458. WEINBERG, H., WALTER, W. G., COOPER, R., AND ALDRIDGE, V. J. Emitted cerebral events. Electroencephalography and Clinical Neurophysiology, 1974,36,449-456. WEINBERG, H., WALTER, W. G., AND CROW, H. H. Intracerebral events in humans related to real and imaginary stimuli. Electroencephalography and Clinical Neurophysiology, 1970,29, 1-9. WOODY, C. D., VASSILEVSKY, N. N., AND ENGEL, Conditioned eye blink-unit: Activity at coronal precruciate of cat. Journal of Neurophysiology, 1970,33, 838. WUNDT, W. Principles of physiological psychology. (Originally published in 1901.) New York: MACMILLAN, 1910. YOSHII, N. Electroencephalographic study on experimental neurosis, a conditioned partly awake state. Proceedings of the XXII International Congress of Physiological Sciences (Leiden), 1962, 1088. YOSHII, N., AND OGUBA, H. Studies on the unit discharge of brain stem reticular formation in the cat: I. Changes of reticular unit discharges following conditioning procedure. Medical Journal of Osaka University, 1960,11, l. YOSHII, N., PRUVOT, P., AND GASTAUT, H. Electroencephalographic activity of the mesencephalic reticular formation during conditioning in the cat. Electroencephalography and Clinical Neurophysiology. 1957,9, 595.
r
2
Self-Consciousness and Intentionality
A Model Based on an Experimental Analysis of the Brain Mechanisms Involved in the Jamesian Theory of Motivation and Emotion
KARL
I.
H.
PRIBRAM
A NEUROBEHAVIORAL ANALYSIS OF BRAIN MECHANISMS IN MOTIVATION AND EMOTION
A. Introduction The recent revolution in psychology has readmitted cognition and consciousness as legitimate areas of scientific investigation. The study of cognitive processes has made rapid strides by taking as its model brain mechanisms assumed to be similar to those of the digital computer (Miller, Galanter, and Pribram, 1960) and by utilizing reactiontime data investigations of memory for verbally coded materials. The currently projected volumes on consciousness and self-regulation presuppose that equally effective strides can be made in our research on, and understanding of, consciousness. The title of the series, in fact, suggests that data on self-regulation, utilizing biofeedback procedures, will provide the substance upon which such strides will be based. But if understanding comparable to that attained for cognition is to be achieved, an experimentally based model of the brain processes operative in consciousness must also be made available. The purpose of this paper is to provide the outlines of such a model. KARL H. PRIBRAM California.
.
Department of Psychology, Stanford University, Stanford,
51
52
KARL
H.
PRffiRAM
There are many meanings attributed to the term consciousness. Webster's dictionary covers a dozen. There have been articles written on the consciousness of cells; Eastern mystics speak of the consciousness of stones; Jungians deal with universal human consciousness (see Ornstein, 1972, 1973 for review). I have, in another manuscript (Pribram, 1976b), taken a somewhat more direct and perhaps practical approach to definition. These largely definitional issues need not concern us here since self-consciousness can be clearly distinguished from other forms of consciousness. Self-consciousness is said to occur when an observer is able to describe both the observed and the observing. Philosophers (Husserl, 1928) have called this ability intentionality-thus the subtitle of the present manuscript. The term is derived from intention, an aim of an action which mayor may not be realized. The separateness of intent and outcome of an action was generalized by Brentano (1925, 1960) to the objects of perceptual "acts." This generalization has proved to be prescient. Recent evidence from brain research (see Pribram, 1971, 1974, 1976a) has shown that the same parts of the brain (the basal ganglia) that control motor function also control sensory input. These controls operate by changing the set point of receptors (see below) in muscles or sense organs and are therefore ideally suited to function as in ten tionali ty mechanisms. Thus the outcomes of actions and the objects of perception come to form one universe-the realized universe of existence-while the intentional universe is dispositional and may even be unrealizable (the awkward term intentional inexistence was meant to convey this nonreality). The difficulty with such terminolgy is, of course, that other philosophers can counter with the proposition that the phenomenal experience of dispositions exists just as much as the outcomes of actions and the objects of experience and that, to some, these dispositions are the existential reality. For us here, the distinction, not the argument, is the important concern: In studies, of self-regulation both disposition and outcome are realized. After all, the instrumentation that allows the externalization (objectively demonstrating the separate existence) of dispositions is the innovation that makes the scientific study of selfconsciousness now possible. Behavioral psychologists have ordinarily deSignated dispositions by the terms emotion and motivation. The intentionality of motivation is relatively obvious, although Miller, Galanter, and Pribram (1960) dis-
SELF-CONSCIOUSNESS AND INTENTIONALITY
53
tinguish between motive and intent as follows: Jones hires Smith to kill someone. Smith commits the murder, but he is caught and confesses that he was hired to do it. Question: Is Smith Guilty? If we consider only the motives involved, the employer is guilty because he was motivated to kill, but the gunman is not guilty because his motive was merely to earn money (which is certainly a commendable motive in a capitalistic society). But if we consider their intentions, then both parties are equally guilty, for both of them knowingly undertook to execute a Plan culminating in murder. (p. 61)
Motive in ordinary and legal language apparently refers more to the feelings involved, while intent refers to the aims of actions. A similar distinction can be made in the case of emotions: The feelings of emotional elation or upset can be separated from their "aim" or targete.g., accomplishing rapport with someone whom one is in love with. The fact that motivational and emotional feelings (dispositions) can be distinguished from their referents makes intentionality possible. The purpose of studies of self-consciousness is to enhance intentionality by providing independent controls (self-regulation) on motivational and emotional dispositions. Stated in this fashion, it becomes clear that the terms intentionality and volition have a good deal in common. In popular parlance, of course, to "intend" something is to "will" it. The issue of self-consciousness is therefore also the issue of voluntary control, and any proposed brain model must take this into account. Interestingly, William James (1950) dealt with these related issues in a most sophisticated manner. I want here, therefore, to develop and evaluate by both positive and negative comment the Jamesian model, critically but not polemically. Rather the presentation will review a series of clinical observations and laboratory experiments specifically designed to test aspects of the model with the aim of providing a modification based on currently available data.
B. Case History The observations and experiments were begun within the framework of a James-Lange view of the problem, a view that William James (1950) proposed as follows: "Bodily changes follow directly the perception of the exciting fact and-our feeling of the same changes as they occur is the emotion" (Vol. II, p. 449). As did most investigators at the time, and perhaps even now, I took this to mean that emotional feelings
54
KARL
H.
PRffiRAM
result when visceroautonomic mechanisms become activated. This aspect of the theory is attributed by William James to Carl Lange, who had suggested that emotional feelings were due to changes in vascularity and other visceral processes. True, the work of Walter Cannon (1927) had made it mandatory to replace peripheral with central mechanisms, but these were still thought of in terms of visceroautonomic processes. It was, after all, the "head ganglion" of the autonomic nervous system that concerned Cannon. My entry into the problem was due to a patient, a Greek woman in her early fifties, who suffered from Jacksonian epileptic seizures always initiated and almost always limited to the left part of her face. Characteristically, even before any muscular twitching, there would be a profuse outpouring of perspiration sharply restricted to the left side of the face and neck as if by a line drawn to separate the two sides. To make a long story short, Paul Bucy and I (Bucy and Pribram, 1943) diagnosed a brain tumor and found and removed a circumscribed oliogodendroglioma located in the right precentral motor cortex. The patient recovered completely with no residual paralysis and with elimination of the seizures. The localized sweating shown by this woman was caused by irritation of the precentral motor cortex and thus called into question the idea then held that it was the hypothalamus which was the "head ganglion" of the autonomic nervous system. Obviously, cortical mechanisms played some role in the regulation of visceroautonomic activity.
C. A Mediobasal Motor System After publication of this patient's story, it became clear to me that visceroautonomic auras were not altogether rare in epileptic patients. However, the great majority of such auras could be referred to pathology in and around the Island of Reil and the pole of the temporal lobe. I therefore began a program of research to map the cortical sites in nonhuman primates from which visceroautonomic responses could be obtained by electrical stimulations. The initial experiments were performed at the Yerkes laboratory with a Harvard inductorium and produced equivocal and unreliable results. I heard, however, that at Yale a new method of cortical stimulation had been developed-a thyrotron stimulator which put out square waves instead of sine waves-and that
SELF-CONSCIOUSNESS AND INTENTIONALITY
55
pulse duration as well as voltage and frequency could be controlled. With this stimulator, Arthur Ward had been able to produce visceroautonomic effects from excitation of the anterior part of the cingulate gyrus (Ward, 1948), and Robert Livingston had succeeded in showing similar effects from the posterior orbital surface of the frontal lobe (Livingston, Fulton, Delgado, Sachs, Brendler, and Davis, 1948). Because of the connections of these portions of the frontal cortex via the uncinate fasciculus, which had been demonstrated not only anatomically but with strychnine neuronography by McCulloch, Bailey, and von Bonin (Bailey, von Bonin, and McCulloch, 1950)-experiments I had had an opportunity to observe-I decided to go to Yale to extend the stimulation experiments to the temporal pole. There I found Birger Kaada, who had just begun his thesis with the aim of analyzing not only the visceroautonomic but also the "suppressor" effects of cingulate gyrus stimulation. Working in adjacent laboratories, obtaining identical effects from stimulation of the temporal pole and the cingulate cortex, late one night we joined forces and mapped the entire area from above the corpus callosum to below it, and by turning the monkey on his back and letting the frontal lobe hang away from the base of the skull, we traced the entire region effective in producing visceroautonomic responses: cingulate, subcallosal, medial and posterior orbital, anterior insular, periamygdaloid, and temporal polar cortex. This was made especially easy once the Sylvian fissure was opened by gentle retraction and temporarily packed with cotonoid patties. In short, we mapped (Kaada, Pribram, and Epstein, 1949; Kaada, 1951) a continuous region of cortex lying on the edge, the limbus of the anterior portion of the cerebral hemisphere, which, when stimulated, produced respiratory arrest, a drop in blood pressure, changes in heart rate, eye movements, turning of the head, and under proper circumstances, suppression (or occasionally enhancement) of spinal reflexes. We had mapped a mediobasal motor cortex. What then of the visceroautonomic seizures in the patient with the precentral oliogodendroglioma? In another series of experiments Patrick Wall and I (Wall and Pribram, 1950) mapped the lateral surface of the cortex and, again to make a long story short, found that such responses could be obtained from the classical precentral motor cortex. Despite a whole series of attempts, we were unable to specify the difference between these responses and those obtained from the mediobasal cortex.
56
KARL
H.
PRIBRAM
FIGURE 1. The medial-basal motor cortex. Black dots indicate areas for which electrical stimulation produces changes in blood pressure, heart rate, respiratory rate, eye movements, and gross bodily movements. (A) Lateral surface: (B) Medial-basal surface. One accomplishes this view by making a slit in the lateral part of the hemisphere and bringing the basal surface in line with the medial surface.
It should not have been altogether surprising that visceroautonomic responses are obtained from stimulations that also produce somatomotor responses. Even stimulation of the hypothalamus, the head ganglion of the autonomic nervous system, produces somatomotor as
SELF-CONSCIOUSNESS AND INTENTIONALITY
57
FIGURE 2. Points of stimulation in the somatosensory motor cortex of the lateral extent of the hemisphere giving rise to changes in blood pressure, heart rate, respiratory rate, and discrete movement.
well as visceral responses. In fact, William James (1950) had stated the issue clearly: If the neural process underlying emotional consciousness be what I have now sought to prove it, the physiology of the brain becomes a simpler matter than has been hitherto supposed. Sensational, associational, and motor elements are all that the organ need contain. The physiologists who, during the past few years, have been so industriously exploring the brain's functions, have limited their explanations to its cognitive and volitional performances. Dividing the brain into sensory and motor centres, they have found their division to be exactly paralleled by the analysis made by empirical psychology of the perceptive and volitional parts of the mind into their simplest elements. But the emotions have been so ignored in all these researches that one is tempted to suppose that if these investigators were asked for a theory of them in brain-terms, they would have to reply, either that they had as yet bestowed no thought upon the subject, or that they had found it so difficult to make distinct hypotheses that the matter lay among the problems of the future, only to be taken up after the simpler ones of the present should have been definitely solved.
58
KARL
H.
PRIBRAM
And yet it is even now certain that of two things concerning the emotions, one must be true. Either separate and special centres, affected to them alone, are their brainseat, or else they correspond to processes occurring in the motor and sensory centres already assigned, or in others like them, not yet known. If the former be the case, we must deny the view that is current, and hold the cortex to be something more than the surface of "projection" for every sensitive spot and every muscle in the body. If the latter be the case, we must ask whether the emotional process in the sensory or motor centre be an altogether peculiar one, or whether it resembles the ordinary perceptive processes of which those centres are already recognized to be the seat. Now if the theory I have defended be true, the latter alternative is all that it demands. Supposing the cortex to contain parts, liable to be excited by changes in each special sense-organ, in each portion of the skin, in each muscle, each joint, and each viscus, and to contain absolutely nothing else, we still have a scheme capable of representing the process of the emotions. An object falls on a sense-organ, affects a cortical part, and is perceived; or else the latter, excited inwardly, gives rise to an idea of the same object. Quick as a flash, the reflex currents pass down through their preordained channels, alter the condition of muscle, skin, and viscus; and these alterations, perceived, like the original object, in as many portions of the cortex, combine with it in consciousness and transform it from an object-simply-apprehended into an object-emotionally-felt. No new principles have to be invoked, nothing postulated beyond the ordinary reflex circuits, and the local centres admitted in one shape or another by all to exist. (Vo!. II, pp. 472-474)
Note that James emphasizes the sensory aspects of these "reflex currents." We shall return to this point presently. But at the time of the discovery of the mediobasal motor mechanism I was surprised and, in a way, a little disappointed-we had not been able to confirm our hypothesis that some part of the cerebral mantle dealt exclusively with visceral mechanisms and could thus be thought of as a "visceral brain"-a substrate for a Langian conception of "emotion." I might add that everyone did not share this disappointment-Paul MacLean, my office mate and collaborator in experiments mapping by electrical stimulation and strychnine neuronography the organization of mediobasal cortex (Pribram, Lennox, and Dunsmore, 1950; Pribram and MacLean, 1953; MacLean and Pribram, 1953), was more convinced by our reports of visceroautonomic regulations by mediobasal cortex than by their invariable concomitance with somatomotor effects (MacLean, 1949). But for me the disappointment was real and led to puzzlement as to just what could be the meaning of this juxtaposition of visceroautonomic and somatomotor mechanisms to the brain's role in emotion and motivation. I therefore turned to other techniques to help resolve these issues.
59
SELF-CONSCIOUSNESS AND INTENTIONALITY
D. The Limbic Systems and Behavior First, it was necessary to obtain evidence that the limbic mediobasal motor mechanisms are in fact critically involved in motivational and emotional processes. To rephrase the question in experimentally testable terms, evidence had to be obtained to show that behavior ordinarily considered to he representative of motivational and emotional processes is disrupted by resections or excitations of the limbic mediobasal mechanism. As it turned out, the results of the experiments undertaken took us a long way into reformulating the problem of what constitutes such behavior. A series of studies designed to analyze the syndrome described by Heinrich Kluver and Paul Bucy (KlUver and Buey, 1937) to follow total temporal lobectomy provided the evidence. Kluver and Buey had included taming, increased oral and sexual behavior, and visual agnosia
DAve 1 Ootnlnilnl
Self-Assured. Feared
71 . .. ~ .::::"",,,
RIVA3
A9\lfturve. Ac:1rve
HIERARCHY BEFORE ANY OPERATION HERBY 4 PlilCid, un,aggrf$$ive
LARRy 8 Submiui"'e, Cowering, Frequenlly Atl,ac:ktd
SHORTY 7
ARNIE 6
~~
Submfutve to Others, Aggreuive Toward L.etry
aeNNY S AJert,. Active food Getter
FIGURE 3(A) Dominance hierarchy of a colony of eight preadolescent male rhesus monkeys before any surgical intervention.
60
KARL
H.
PRIBRAM
wi1k ZeKE 1
OoMi~nt
J\wrtlsive
RIVA 2 Oarinv. Competes With Zeke
HIERARCHY AFTER DAVE'S OPERATION
1
LARRY 7 AnKk1. O;we
Oomln.alIH .. nd
R.~ ,
1
?
"'-
~
DAVE 8 Comploto'v s..bm;"'v• • F.. "u'
ARN1E 5
SHORTY 6
3(B) Same as (A) after bilateral amygdalectomy had been performed on Dave. Note his drop to the bottom of the hierarchy.
FIGURE
in their syndrome. Our studies (Blum, Chow, and Pribram, 1950; Chow, 1951; Mishkin and Pribram, 1954; and Pribram, 1954) showed the agnosia to be due to resection of the lateral portions of the temporal lobe; however, these results make up a body of evidence which, though related to the issues being examined here, constitute a sufficiently separate domain to warrant presentation on another occasion (see, for example, Pribram, 1969, 1975) . The remaining part of the syndrome was obtained full-blown when lesions were restricted to the anterior limbic portions of the lobe-those comprising the temporal lobe portions of the mediobasal motor mechanism (Pribram and Bagshaw, 1953). Subsequent studies showed the entire mediobasal motor system to be involved (Pribram and Weiskrantz, 1957). Specifically, tests were performed to measure feeding, fleeing (avoidance), fighting (dominance), mating, and maternal (nesting) behavior (see reviews by Pribram, 1958, 1960). The fairly gross changes in
61
SELF-CONSCIOUSNESS AND INTENTIONALITY
~ RIVA' Oomm.ant, Not Th, .. tantd by Othtfs
<, .."
~
JNI BENNY 3
HIERARCHY AFTER ZEKE'S OPERATION
ZEKE 7 SubmluiwlII 10 Others,
Intarminenlly TOWlrd
0.."".
Aljlgu!'UI'I.
LARRY 6
.~
~ oo
-
SHORTY 5
DAve 8 Ctingtor. Avoids Int.rlKlion
3(C) Same as (A) and (B) except that both Dave and Zeke have received bilateral amygdalectomies.
FIGURE
these behaviors following lesions of the limbic motor systems are so well known by now that I want here to present more quantitative data. The effects of these lesions on fighting were observed in social situations (Rosvold, Mirsky, and Pribram, 1954). A group of eight preadolescent male monkeys were housed together until a more or less straight-line dominance hierarchy became stably established. Dominance was initially rated on the basis of order of obtaining food pellets inserted into the colony space, one by one through a metal tube. The dominance rating obtained in this way was then checked against quantitative observations of threatening gestures, actual fighting contacts, grooming contacts, and position displacements. Such observations were made not only in the colony when the group was together as a whole but also for all possible dyads. Then the clearly dominant monkey was submitted to psychosurgery. As expected, he fell to the bottom of the hierarchy. Interestingly, however, this drop was effected over a
62
KARL
H.
PRffiRAM
~ . . . . ,I!'
More Oomin.nt. Unpri!dictablV Aw ressiye and Vicioul
HIERARCHY AFTER RIVA'S OPERATION
ZEKE 7 Contin .... n In,ermi11enlJy Aggressh,"
Tow.rd O.....e
DAVE 8 Out
· 717'~ ~ ~
t
3(0) Final social hierarchy after Dave, Zeke and Riva have all had bilateral amygdalectomies. Note that Riva fails to fall in the hierarchy. Minimal differences in extent of locus of the resections do not correlate with differences in the behavioral results. The disparity has been shown in subsequent experiments to be due to Herby's nonaggressive "personality" in the second position of the hierarchy. FIGURE
48-hour period during which interactions with the other monkeys gradually lowered the "status" of the previously dominant monkey. This experience was essentially replicated when we performed surgery on the formerly Number 2 and now dominant animal. When, however, we attempted to repeat the procedure for a third time, the expected effects did not occur. In fact, the original Number 3 animal became, if anything, more aggressive and dominant. My colleagues in the study, Hal Rosvold and Alan Mirsky, of course, blamed inadequate surgery for this development, but histological verification failed to confirm their suspicions. In fact, the lesion of this last monkey extended further than that of one of the others, and all lesions encompassed the same anatomical structures. What, then, could account for our results? Briefly, examination of our data, especially of the 48 initial postoperative hours in the colony,
63
SELF-CONSCIOUSNESS AND INTENTIONALITY
and observations of other colonies of monkeys made us believe that lack of interaction between the operated subject and the original Number 4 monkey was responsible for the original Number 3 monkey's failure to fall in dominance. Many subsequent observations in dyadic situations confirmed this belief: Postoperative monkeys were especially sensitive to the way they were treated by their cage mates and handled by their caretakers. The immediate postoperative taming could be prolonged for years by gentling procedures, whereas ordinary neglect and occasional rougher treatment would produce either an excessively fearful or an unpredictably aggressive monkey. These results make it unlikely that some fundamental mechanism responsible for aggression had been excised; rather some brain process sensitive to the social environment seems to have been tapped.
70~~--~-----r-----r-----r-----------r----~----~
• ANTEROFRONTAL, FRONTOTEMPORAL. ME DIAL-FRO NT AL-CING ULATE, AND AMMON'S FORMATION o OCCIPITO-PARIETAL AND INFEROTEMPORAL 6SHAM OP.
5
0-10 TEST TRIALS - 1st DAY
TEST DAYS
Graph of the percentage of time spent by various groups of animals in the dark (previously shocked) compartment of a shuttle box during postoperative extinction of a preoperatively acquired conditioned avoidance. The scores for the first extinction day are recorded in lO-trial blocks; subsequent extinction trials are plotted in 50-trial blocks (one test day). Vertical bars indicate range of performance.
FIGURE 4(A)
64
KARL
H.
PRmRAM
60 I-
Z ~ 50 I0::
~
~ 40
u
'"C!i 0::
30
~ w
::;; f= 20
....
10
5 TEST TRIALS - 1st DAY
TEST DAYS
4(B) Graph of the percentage of time spent by the various groups of animals in the dark (previously shocked) compartment of a shuttle box during postoperative reextinction of postoperatively reacquired conditioned avoidance. Trial blocks divided as in 4(A). Bars indicate range of performance. FIGURE
Similar results were obtained when fleeing, (avoidance) was tested by Weiskrantz and myself (Pribram and Weiskrantz, 1957) in a classical shuttle box. Escape proved unaffected by limbic lesions, but learning, extinction, and relearning of conditional avoidance behavior were affected. It should be noted in passing that limbic and not lateral forebrain lesions (except for frontal in the case of extinction) produced such results; however, neither classical sensory and motor resections nor basal ganglion removals were included in these studies. Perhaps the clearest indication of what type of regulation is accomplished by the mediobasal motor mechanism came from our studies of feeding. Postoperatively, monkeys with lesions in these limbic areas often failed to eat for a time (usually not more than a week). Once they recovered, however, they might eat twice as much per day as preopera-
65
SELF-CONSCIOUSNESS AND INTENTIONALITY
tively, and this increased food intake usually lasted for months (Pribram and Bagshaw, 1953). Further analysis almost immediately uncovered a paradox. Despite this marked increase in feeding in an ad libitum situation, Schwartzbaum and I (Schwartzbaum, 1961) found that the monkeys with lesions were less sensitive to food deprivation or satiation when made to work for their food. They appeared "hungrier" when food was available but less "hungry" when they were deprived and food could be obtained only by the pressing of a lever to obtain pellets (on a variable interval schedule). But this was not all; the loss of sensitivity was not limited to variations in the internal state produced by the deprivation but extended as well to variations in the external characteristics of the food,
LU250 A
70- HRS. FOOD DEPRIVATION
!!)
~220~ ~
190
~160
II::
130
-
...J
~
5(A) Effect of food deprivation on number of responses emitted by normal and bilaterally amygdalectomized monkeys in a fixed-interval operant-conditioning situation. Note that the percentage change in total number of responses is plotted on the graph. Absolute values are indicated in the right lower comer. 5(B) The effects of the changing of reward size on bilaterally amygdalectomized and normal monkeys on response rate in a fixed-interval situation. Note that normal animals satiate rapidly in a session when the pellet size is increased.
00
gloo
FIGURE
2
Normals 175-480 Amygdalas 274-375
3
4
5
SESSION
~ z
B
~ 140
u
~ o
LU
~
II::
LU
~
120
~ ~o ..
o
······0
100
~ 80
ffl
LARGE REWARD
-Normals 0 Amygdalas
o
II::
4 3 5 2 10- MINUTE INTERVALS
6
66
KARL
H. PRIBRAM
such as the size of the pellets received as a consequence of lever pressing (Schwartzbaum, 1960a, 1960b). These results paralleled those that were being reported from analyses of feeding disturbances produced by ventromedial and far-lateral hypothalamic lesions. Miller, Bailey, and Stevenson (1950) found that rats who became obese in ad libitum feeding situations might starve if required to make the effort to cross a barrier to obtain their food. And Teitelbaum (1955) showed that rats who would starve in ad libitum situations could be induced to attend and eat if the sensory attractiveness of the food was sufficiently enhanced. Two major hypotheses derive from these observations, one with respect to the sensory, the other with respect to the motor processes regulated by the limbic (and hypothalamic) mechanisms: 1. Attention (Le., reaction) to external as well as to internal stimulation is involved in motivational and emotional feeling. 2. Effort, not drive (e.g., as defined by ad libitum feeding and lever pressing), is the critical variable determining motivational and emotional expression in behavioral responses. Let us examine the evidence related to each of these hypotheses in tum.
II. THE
ROLE OF ATTENTION IN MOTIVATIONAL AND
EMOTIONAL REACTIONS
A. Transfer of Training In order to bring the altered reactions to external stimulation of the lesioned monkeys into sharp focus, I decided to test them in a series of tasks which were as minimally related to motivation and emotion as I could find and yet might provide some indication of function. The point of departure for selecting these tasks was the dramatic change in dominance displayed in the social colony experiment. The lesioned monkeys behaved postoperatively as if they had never experienced the colony structure-they seemed to have to learn anew the repertoire of aggressive interactions that established their place in the hierarchy. They appeared not to transfer their prior experience to the postoperative situation.
67
SELF-CONSCIOUSNESS AND INTENTIONALITY
Thus a series of transfer-of-training tasks was devised. The first, undertaken with Schwartzbaum (Schwartzbaum and Pribram, 1960), used a transposition paradigm in which the monkeys were initially trained to choose the lighter of two gray panels and were then tested on a pair of panels of which the formerly lighter one was now the darker of the test pair. All control subjects continued to choose the lighter panel, but the lesioned animals behaved oddly. They hesitated and then chose randomly between the two panels. It appeared as if they perceived the test situation to be novel and proceeded accordingly. The results of a second experiment performed with Muriel Bagshaw (Bagshaw and Pribram, 1965) supported the findings of the first. In this experiment Heinrich Khiver's equivalence test (Kluver, 1933) was used. The monkeys were trained to choose the larger of two moderately sized squares and tested on a pair of smaller ones. Again the control subjects chose the larger panel throughout while the lesioned monkeys behaved as if the test panels presented them with an entirely novel situation. Both the transposition and the equivalence results could be attributed to an altered gradient of generalization of input following the lesion. Eliot Hearst and I (Hearst and Pribram, 1964a, 1964b) put this possibility to test and found generalization unimpaired. In neither positive nor negative reinforcing situations were the monkeys' generalization gradients changed by the lesion. Thus transfer of training appears to be dissociable from sensory generalization-and perhaps might be more appropriately thought of in terms of "motor generalization"-a view consonant with the fact that lesions of the mediobasal motor cortex were the responsible agent in producing the changes. TABLE 1 Number of Transposed Responses Made on Transposition Tests Normals
Amygdalectomized
Day
1 2 Total
439
441
443
447 Mdn.
6 5 11
5 5
6 5 11
6 6 12
10
11.0
397
405
438
442
Mdn.
2 3 5
5 6 11
2 2 4
4 2 6
5.5
The effects of amygdalectomy on transfer of training to a new but related task. Note that the amygdalectomized monkeys treat the task as completely novel. whereas their normal controls transpose their responses on the basis of their earlier experience.
68
KARL
H.
PRffiRAM
B. Psychophysiological Experiments The observations on response to novelty made in these neurobehavioral experiments were considerably enhanced by some obtained in psychophysiological studies. One possibility, consonant with the Jamesian hypothesis, would be that motivation involved the somatomotor system while emotion invoked visceroautonomic processes. Thus, despite the juxtaposition of their central mechanism, peripheral differences could account for the behavioral distinction. This possibility was put to direct test in a series of experiments which assayed the effects on visceroautonomic indicators of resections rather than stimulations of portions of the mediobasal, limbic motor mechanism. Such experiments allowed observations to be made under physiologically normal conditions for many months and even years in a variety of environmental circumstances that ordinarily produce visceroautonomic reactions. The question was asked as to which of these circumstances produced altered reactions or absence of reaction in the lesioned monkeys. To summarize a decade of experiments, Muriel Bagshaw, Daniel Kimble, and I (Bagshaw, Kimble, and Pribram, 1965; Kimble, Bagshaw, and Pribram, 1965) found that the forebrain lesions had, as might be expected, no effect on peripheral, reflexly produced visceroautonomic reactions. Galvanic skin responses (GSR), heart, and respiratory changes occurred in normal amount and frequency when the monkeys moved or when gentle electric shock was applied to the soles of their feet. We found (Bagshaw and J. Pribram, 1968), if anything, that the threshold for obtaining such reactions was lower than in normal subjects. By contrast, however, when response to novel stimulation or to conditioning was tested, visceroautonomic reactivity was grossly deficient. The visceroautonomic components of orienting and conditioning were markedly attenuated or completely eliminated by the lesions (Bagshaw and Benzies, 1968). Analysis showed that the deficit in conditioning was to some considerable extent due to a restriction in anticipatory reactions to the unconditional stimulus which occurred in control subjects (Bagshaw and Coppock, 1968). Thus the situations in which the deficit was manifest were those that demanded reactions to recurring events, not reactions to the events themselves. The fact that limbic forebrain structures, the mediobasal motor mechanisms, are critically involved in the reaction to novelty as well as in motivational and emotional reactions
69
SELF-CONSCIOUSNESS AND INTENTIONALITY
••- -..... N
80
x, ••••••••••••••• 'X
H
0···············0 IT .················AM
x ..
\~
..........
60
\,\
.....•...
\,
AVG PERCENT GSR
\ ..",...
.~..::>\:;:::.... ~ " u
40
\\
...... - -
••'!-•••••• ::.~: ••••••Ji....•.....•..••. .:!.
"..•., ,., ...
20
....
~.
\\.. \,. ............... 0 ..•................•.....................
...•......
o
234 10 TRIAL BLOCKS
..........•
5
FIGURE 6. Curves for overall percentage of GSR responses to the first 50 presentations of an irregularly presented 2-sec tone. Amygdalectomized (AM), hippocampectomized (H), control (IT), and unoperated (N) groups.
suggests a major modification of Jamesian theory: Emotional and motivational feeling comes about not by any direct bodily reaction to perceived events, but by a change in the sets, expectations, and anticipations produced by such events.
C. Habituation The simplest expression of such sets or expectations is habituation. And, in fact, habituation of a locomotor response in a repetitive situa-
70
KARL
H.
PRIBRAM
100~-------------------------------'
A
10
;. v/- - - - . ::: 60
•....
~
&"
................
•••••••/ . . ....~ ............ \
..
'
.~
\.
•
~
&.20.
'.
,:'
OL-------------------------------~ o .1-4 4·1 I· I 1-4 4 ·.1 o d
...
...... AM
60 B
,.........•..
'"z ...'"240 III
.:
.../
...z
:::20 III
=:
SHOCK
..
-N
......
...................
. .I·A
.4-.1
.1·1.2
1.2·1.6
1.6-2.0
IIA SHOCK FIGURE 7(A) Curves of percentage GSR generated by three runs of stimuli of ascending and descending intensity (in map) by the amygdalectomized (AM) and control (N) groups. 7(B) A finer breakdown of stimulus values from .1 to 1.0 mamp, pooled ascending and descending values.
tion has been repeatedly shown to be impaired by lesions of the mediobasal motor mechanism (Ruch and Shenkin, 1943). This failure to habituate has often been termed hyperactivity, though it was soon established that it was more truly a hyperreactivity (Mettler and McLardy, 1948). The hyperreactivity is, however, not so much an increased initial reaction (though there is some of this as seen in the GSR
SELF-CONSCIOUSNESS AND INTENTIONALITY
71
threshold to shock experiment) as it is a persistence of reactivity long after controls have become habituated. But here we meet a paradox. Behaviorally, the monkeys with lesions of the mediobasal motor system fail to habituate, i.e., they continue to orient long after control subjects react to a recurring situation as familiar (Schwartzbaum, Wilson, and Morrissette, 1961). As already described, however, when we looked to visceral-autonomic responses to indicate orienting, such responses could hardly be found (Bagshaw, Kimble, and Pribram, 1965). The evidence thus suggests that visceroautonomic responses are integral to habituation. Can visceroautonomic responses be integral to habituation and also be the determinants of emotional feeling? In man habituation precludes awareness. We are not ordinarily aware of wearing clothes that have become familiar, of movements that have become habitual, of digestive functions or heart beats that recur more or less regularly. Only when dishabituation takes place do we notice such objects and events. The feeling is therefore attendant on dishabituation-disruption of the current set or state. Bodily changes may accompany the disruption, and the ensuing visceral and somatosensory input may in fact contribute to the general emotional or motivational feeling. But, as I have reviewed elsewhere (Pribram, 1967a, 1967b, 1971, Chapter 11), the attribution of specific feelings to the change in state is as much a function of the situation in which the change occurs as it is of the visceral and somatic changes per se. As pointed out in the introduction, the feeling of upset can readily be distinguished from the disposition (e.g., being in love) from which the upset takes its origin. Habituation poses another problem. If we habituated in every recurring situation we would never be able to deal with such situations; we would never learn, would never be able to attend to now this, now that aspect of a situation. The organism must possess a mechanism which overrides habituation. In a recent review, Diane McGuinness and I (Pribram and McGuinness, 1975) spelled out the details of this system, which appears to depend on greater involvement of the somatomotor rather than of the visceroautonomic system. The data suggest that three separate but interacting neural systems govern the reaction to novelty and its habituation. One system controls arousal, which is defined in terms of phasic physiological responses to input. The arousal control circuits center on the amygdala, a core structure in the mediobasal motor system. A second system controls activation, which is de-
I .. • • • • o.
9.3
5.8
NORM
AHX
2.7
AMX
S S
8.2
14. 5~':
4.8
8.8n
3.3
7.3
10.3
3.5
6.2
3.9
3.9
7. Om':
6~t
OFF
5-10 SEC OFF
.. .. ..
A ON
>I
'5
6.3
7.0
4.3
4.5
6.3
7.0
10-15 SEC OFF
'0
L-CONTROL ~
5
10-15 SEC ON
..........
:' 5
FIGURE
8. Mean number of GSR occurring in 10-sec period of light on just preceding light offset (CS) in the first 40 and in the second 40 trials for each group. Conditioning paradigm presented above table.
-AI I 80
5. 7~'c*
3.2
AMX
NORM
3.7
NORM
-
....
5-10 SEC ON
..........
FIRST 40
SECOND 40
:::
• • • • • • • o. •••••• •••••• ••• ............ .... .o..................... ............ .... .. ••.••. .............. o... ...... ..... ..... ............ .... •.... o......... ........... .... .... ............ .. .. .
:~:::
... . ... ... .. I
5 :::::: :: ::::::
GRP
ON
A
o
TRIALS
SEC
E::
~
~
'"0
~
~
~
tv
'I
SELF-CONSCIOUSNESS AND INTENTIONALITY
73
fined in terms of tonic physiological readiness to respond. The readiness circuits center on the basal ganglia of the forebrain. A third system was discerned which coordinates arousal and activation. This coordinating activity apparently demands effort. Its circuitry centers on the hippocampus. Even at the hypothalamic level the distinction between an arousal and a readiness mechanism exists. In the feeding mechanism, for instance, the ventromedial region (the one involved in making obese rats in the ad libitum situation) has been shown to monitor the utilization of blood sugar during satiety, while the far lateral hypothalamic region functions reciprocally to intiate feeding when utilization has come to a halt. The satiety mechanism stops behavior; the feeding mechanism makes behavior go. And recent evidence (Ungerstedt, 1974; Fibiger, Phillips, and Clouston, 1973) has shown these far lateral hypothalamic effects to be due to disruption of basal ganglia circuits. Elsewhere (Pribram, 1971, Chapter 10) I have suggested that emotional arousal becomes organized around "stop" mechanisms and that motivational readiness is an elaboration of "go" mechanisms. As detailed in the review, there is ample evidence that the limbic forebrain (e.g., the amygdala and the hippocampus) participates in such processes. Here it is important to recall that the evidence also shows that this participation takes place by way of the organism's reactivity to novelty and familiarity. The evidence is thus consonant with that from the clinic where epileptic auras of deja and jamais vu preceding psychomotor seizures are considered pathognomonic of disturbances of the limbic mediobasal motor formations.
D. James Reconsidered In his opening paragraph on emotions, William James suggested that "emotional reaction usually terminates in the subject's own body whilst the instinctive (motivational) reaction is apt to go farther and enter into practical relations with the exciting object" (1950 Vol. II, p. 442). This distinction rather than the one more commonly attributed to James-that emotion is essentially viscerally determined-is borne out by our review of current data. James was overly impressed with Lange's argument because "reactions that terminate in one's own body" (Le., self-regulatory reactions) tend to display a larger visceral component
74
KARL
H.
PRffiRAM
than do "reactions" which "enter into practical relations with the exciting object." Entering into practical relations demands somatomotor activity. Yet emotional arousal also involves the somatomotor system, and somatomotor activity is accompanied by visceroautonomic changes. In short, we have a considerable amount of evidence which demands a modification of the James-Lange position that "bodily changes follow directly the perception of the exciting fact and our feeling of these same changes is the emotion." Feelings of familiarity, of elation and depression, of assertion and aggression, and of sleepiness and alertness have been shown to depend on brain processes (see Pribram, 1971, Chapters 10 and 15; Pribram and McGuinness, 1975; or Marshall and Teitelbaum, 1974, for a review of the evidence that relates neurochemical brain systems to these dispositional states). Bodily changes are initiated by these brain processes, but not, as James thought, by the processes that directly perceive. Rather, the bodily changes are induced by mechanisms which monitor the familiarity and novelty of situations. Bodily changes, both visceroautonomic and somatomotor, do appear to be integral to emotional and motivational expression, however, in that they do in fact help to distinguish their mechanisms of operation. But even here the distinction does not rest on which peripheral mechanism becomes activated, but rather on how they both are used. If the brain processes regulating bodily changes lead the organism into doing something about a situation, i.e., "entering into practical relations" with it, motivational mechanisms become active; when these brain processes result in reactions "terminating within the subject's own body," emotional mechanisms are set into operation. This, then, is the essential distinction between motivational and emotional expression: What mechanism sets one rather than the other process in motion?
III.
EFFORT AND THE EXPRESSION OF MOTIVATION AND EMOTION
A. Part Behaviors and Their Integration In attempting to answer this question, we need to examine the impact of the finding that effort, rather than drive, is the critical variable determining the behavioral response. Presumably, therefore, it
SELF-CONSCIOUSNESS AND INTENTIONALITY
75
is effort that resolves whether the organism will react emotionally (i.e., by attempted self-regulation) or motivationally (i.e., by entering into practical action). Effort is a measure of the resistance which must be overcome in order to do a certain amount of work in a specified time. It is analogous to force in physical systems and an expression of the capacity for exerting power, the rate of doing work. And work is a measure of the energy required to change the state of a system (see McFarland, 1971, p. 4, for the derivation of these definitions). The critical question is, therefore, What are the variables which constrain a system to resist a change in state? A homeostatic system, by definition, is one that resists change by virtue of its negative feedback. But additional constraints develop (hyperstability) when several such systems interact (Ashby, 1960). Thus a drop in basal temperature may result in shivering, in motor activity, in sleeping, or in eating. Motor activity and eating have been shown to be reciprocally related over short time periods-they appear constrained by the basal temperature variable. It therefore takes effort to attempt to eat during or immediately after exercising and vice versa (see Brobeck, 1963, for a thought-provoking and thorough review of these data). There is considerable evidence as to the neural organization that invokes such constraints. Electrical stimulations in the hypothalamic region, when carried out with small electrodes, give rise to only parts of behavioral acts, such as jaw or tongue movements, swallowing, pilo- or genital erection, head thrusts, etc. Further, these part behaviors appear to be more or less randomly interspersed with one another. Adjacent stimulations do not produce a completed behavior pattern (Roberts, 1969). When larger electrodes are used, a different pattern emerges. Now chewing, drinking, sexual, or aggressive behaviors are elicited fullblown. But interestingly, which behavior is elicited depends to some considerable extent on the environmental situation in which the stimulation occurs. Thus a rat, when initially stimulated, may drink every time the electrical current is applied to his brain. He is now left overnight in a cage with no opportunity to drink but with several pieces of wood to chew on. The brain stimulation is kept up intermittently all night. The next morning the rat is provided with the opportunity either to drink or to chew. Now, brain stimulation from the identical site will as often elicit chewing as drinking (Valenstein, Cox, and Kakolewski, 1969; Valenstein, 1970).
76
KARL
H.
PRffiRAM
B. The Precentral Motor Cortex and Action These results have led to a controversy similar to that which for many years centered on the functions of the classical precentral motor cortex. The issue concerned the nature of the motor representation: Is it punctate, representing discrete muscles or even parts of muscles, or are movements, sets of muscle contractions, flexibly represented? I have elsewhere (Pribram, 1971, Chapters 12 and 13) pointed out that neuroanatomically the representation is indeed punctate, that neurophys-
A
B
77
SELF-CONSCIOUSNESS AND INTENTIONALITY
iologically, i.e., with fairly gross electrical stimulations of awake animals and humans, movements rather than discrete muscle contractions are obtained, and that which movement is elicited depends on body and limb position, prior stimulation, etc. But I also have shown (Pribram, Kruger, Robinson, and Berman, 195~56) that neither of these views is sufficient to explain the results of neurobehavioral experiments. These show that resection of the classical motor cortex fails to interfere with any muscular contraction, or even with any set of muscular contractions. All movements can be shown to remain intact when
kg 14 12 10
8 C
8 6
6
4 -
2
4
2 0
-2 -4
-6 -8
nC
n/3
FIGURE 9(A) Subject in black costume with white tape. (Reprinted with permission from N. Bernstein, 1967.) 9(B) Cinematograph of walking. Movement is from left to right. The frequency is about 20 exposures per second. (Reprinted with permission from N. Bernstein, The Co-ordination and Regulation of Movements. © Pergamon Press Ltd., 1967.) 9(C) Force curves at the center of gravity of the thigh in normal walking. (Above) vertical components. (Below) Horizontal components. (Reprinted with permission from N. Bernstein, The Co-ordination and Regulation of Movements. © Pergamon Press, Ltd., 1967.)
78
KARL
H.
PRIBRAM
SELF-CONSCIOUSNESS AND INTENTIONALITY
79
they are examined in a sufficient range of situations. Yet the monkeys were defective in solving latch-box problems, and the deficiency was not due to any overt difficulty in sequencing the movements. I therefore came to the conclusion that the essential representation in motor cortex was neither of individual muscles nor of movements, but of actions, defined as environmental consequences of movements. Subsequent reports (Bernstein, 1967) have clarified the possible mechanisms by which a representation of environmental consequences could come about. For instance, cinematographic records are made of people performing relatively repetitious tasks, such as hammering a nail or running over rough terrain when dressed in black with white markings on their limbs. Such records display continuous wave forms which can be analyzed as if they were modulated sine waves. The use of such a Fourier analysis allows accurate prediction to be made of the extent of the next movement in the series, the next hammer blow or step. If this can be done by an investigator in this fashion, it is not farfetched to believe that it can be done in a similar way by the subject's motor system. The essential representation would therefore be the equivalent of the mathematical operation of Fourier analysis, a considerable saving in storage over representing each movement and sequence of movements that might ever be utilized. This program, or a similar stored set of mathematical rules, could readily assemble the more or less randomly dispersed part-functions which have been demonstrated with discrete stimulations, much as these rules have been used to make a computer-driven dot display that is interpreted by the observer as a running or dancing figure (Johansson, 1973). Other experiments (Evarts, 1967) have shown with microelectrodes that muscle length is not the relevant variable to which motor cortex neurons respond. These experiments were performed on fully awake 10(A) View of the lateral surface of the cerebral cortex showing the distribution of potentials evoked by the stimulation of a cutaneous or a muscular branch of an arm nerve. Plus (+) indicates a response of 100 microvolts or more; triangle (Il) indicates a response of from 50 to 100 microvolts; and open circles indicate response of from 0 to 50 microvolts. 10(B) Cortical responses evoked by sciatic nerve stimulation before resection of postcentral cortex and cerebellum. (1) Upper trace, postcentral; lower trace, precentral. Time: 10 msec. (2) Same immediately after resection of both cerebellar hemispheres. (3) Same immediately after resection of both cerebellar hemispheres. (3) Same following additional resection of anterior lobe of cerebellum. (4) Same after additional resection of both postcentral gyri. Note that postcentral record how registers only white matter response.
FIGURE
80
KARL
H.
PRIBRAM
monkeys taught to pull an adjustable counterweighted lever. The response of motor cortex units did not vary with the length of the excursion of the lever but with the effort necessary to move it-the force that had to be applied to overcome the resistance to movement due to the weight.
C. Effort and Volition Thus both in the hypothalamic experiments and in the experiments on classical motor cortex, situational stimulus variables are seen to be critically involved (this accounts for the findings of Teitelbaum, 1955, that enhanced attractiveness of food helps overcome the resistance to eating shown by the animals with lesions in the lateral hypothalamus). Both of these parts of the brain do, of course, receive rather direct inputs from exteroceptors. In the case of the classical motor cortex, our discovery (Malis, Pribram, and Kruger, 1953) of these paradoxical inputs to a "motor" region provided the original impetus for this line of investigation. In the case of the hypothalamus and mediobasal motor cortex, the existence of these direct inputs from exteroceptive receptors is just now beginning to be established with microelectrode and new neuroanatomical autoradiographic techniques (see, e.g., Cowan, Adamson and Powell, 1961; Cowan, Gottlieb, Hendrickson, Price, and Woolsey, 1972). However, some early neurophysiological results exist showing changes in electrical activity evoked by peripheral stimulation (Bailey and Sweet, 1940; Dell, 1952; Pribram and MacLean in Fulton, 1951, p. 57). The distributed representation of part behaviors is the neural substrate upon which effort variables (as induced by deprivation, for instance) critically operate to determine whether the expressed behavior is to be emotional or motivated. We have seen that even at the neural unit level, neurons in motor systems are directly sensitive to effort variables-i.e., they respond according to the constraints of the moment. The constraints that need to be overcome by effort have thus been shown to be externally as well as internally determined. This was especially clear in experiments on the classical motor cortex: Effort is what correlated with neural unit activity, not changes in muscle length. Similar correlations need to be established at the unit level in the mediobasal motor systems, but, as noted above, the indications from
SELF-CONSCIOUSNESS AND INTENTIONALITY
81
neurobehavioral data are that some sort of anticipatory mechanism based on the constraints developed by repetition (familiarity, habituation) rather than "bodily change" per se is involved. The upshot of these results is that motivation and emotion reflect the effort involved in changing bodily systems, not the changes themselves. Effort is a brain process (involving the hippocampal circuit-see Pribram and McGuinness, 1975, for a review of the evidence) that apppears to be critical in determining whether a reaction is to be motivated or emotional.
D. The Jamesian Theory of Will We need, therefore, to take a look at another domain of Jamesian theory. We have already noted the fact that some sort of appraisal of familiarity rather than a direct perception of a situation initiates the motivational-emotional process. We also reviewed the evidence that the distinction between emotional and motivational behavior was best stated by the Jamesian view that emotional reactions "terminate within the body" (i.e., are self-regulatory), while motivational reactions "enter into practical relations with the exciting object." The Langian portion of the James-Lange theory-that emotions are the feelings generated by visceral reactions-we found untenable. And finally we found that we must invoke "effort" as the critical variable which determines whether a reaction is to be emotional or motivated. Effort is discussed by James under the rubric of will. His definition of what leads to voluntary actions reads much as we have stated it here, if we interpret the words anticipatory image to mean the resultant of the mathematical operation from which the next movement in a series can be predicted: An anticipatory image, then, of the sensorial consequences of a movement, plus (on certain occasions) the fiat that these consequences shall become actual, is the only psychic state which introspection lets us discern as the forerunner of our voluntary acts. (1950, Vol. II, p. 501)
But again, James can be interpreted as taking a dual stance. He quoted at length from Ferrier, who attempted to show that input from muscular contraction (usually the holding of one's breath when other evidences of muscular contraction are missing) is necessary for the experience of effort. (Ferrier was making the argument against Wundt, whose views were that efferent rather than the afferent neural activity
82
KARL
H.
PRffiRAM
was perceived as effortful). James heartily endorsed Ferrier's views: [the experiments reviewed] prove conclusively that the consciousness of muscular exertion, being impossible without movement effected somewhere, must be an afferent and not an efferent sensation; a consequence, and not an antecedent, of the movement itself. An idea of the amount of muscular exertion requisite to perform a certain movement can consequently be nothing other than an anticipatory image of the movement's sensible effects. (Vol. II, p. 505)
Note carefully here what James was saying. Superficial reading makes the statement sound like another version of the James-Lange theory: James-Lange for emotion; James-Ferrier for motivation and will. But here James clearly stated that "an anticipatory image of the movement's sensible effects" is involved. Such an "image" must be a brain, not a peripheral, process. A careful reading of this passage makes one wonder whether the Jamesian theory of emotions, interpreted as peripheralist, has not been grossly misinterpreted as well. James implicitly and explicitly always had a brain process in mind whenever discussing mind (thoughts, feelings, consciousness, attention, etc.). In summarizing his chapter on emotions, James was discussing brain processes, not peripheral ones: To sum up, we see the reason for a few emotional reactions; for others a possible species of reason may be guessed; but others remain for which no plausible reason can even be conceived. These may be reactions which are purely mechanical results of the way in which our nervous centres are framed, reactions which, although permanent in us now, may be called accidental as far as their origin goes. In fact, in an organism as complex as the nervous system there must be many such reactions, incidental to others evolved for utility's sake, but which would never themselves have been evolved independently, for any utility they might possess. Sea-sickness, the love of music, of the various intoxicants, nay, the entire aesthetic life of man, we have already traced to this accidental origin. It would be foolish to suppose that none of the reactions called emotional could have arisen in this quasi-accidental way. (Vol. II, p. 484)
We note, therefore, that the contemporary view of the theory of motivation and emotion proposed by William James is in one respect grossly misleading. While James wrote that emotional feeling was based on visceral sensations, he also wrote that such feeling was coordinate with a brain process resulting from the visceral sensation. This central aspect of Jamesian theory becomes even more clearly stated with respect to motivation and has been little appreciated by James's critics.
SELF-CONSCIOUSNESS AND INTENTIONALITY
83
On the other hand, James was in error in suggesting that emotion depended on immediate visceral sensation (or that motivation depended on immediate sensations derived from the somatic musculature). Cannon's classic experimental demonstrations that an organism is capable of emotional responses despite visceral deafferation have been the source of the major rebuttal to James's position, although exceptions to the validity of Cannon's claims have also been voiced (e.g., Beebe-Center, 1971; Schachter, 1967; Mandler, 1967). Reviewed here have been additional experiments that make it necessary to modify Jamesian theory of emotion. James clearly wrote of "an anticipatory image" when dealing with motivation. On the basis of the experimental results on limbic system function-which have shown that visceral responsiveness follows the appreciation of novelty and the appraisal of changes in sets, expectations, and anticipations, not directly on perceived events per se--it is now mandatory to think in like fashion about emotion. However, we also reviewed an aspect of Jamesian theory which is acceptable today and accounts for James's overemphasis on immediate bodily sensations while at the same time providing a useful distinction between emotion and motivation: Emotional expression tends to terminate within the organism while motivations enter into practical relations with the exciting event. Entering into practical relations often involves effort or will-thus the distinction still used in the neurological clinic of the apposition of emotional to voluntary behavior.
IV. A
CONTROL-THEORY MODEL OF SELF-REGULATION AND
SELF-CONSCIOUSNESS
A. The Model In the introduction I suggested that the scientific study of selfconsciousness would show rapid progress provided a technique and a brain model were made available. The technique of biofeedback leading to the self-regulation of dispositional states appears to fill part of this need. The present manuscript has attempted to show that Jamesian theory might be useful in launching the necessary brain model.
84
KARL
H.
PRffiRAM
What sort of model, then, can be constructed from these elementary observations? What form of discourse is available to describe the model? What type of "in vitro" simulations can aid our understanding of the processes and mechanisms involved? I want to propose that contemporary control theory can and does provide the model, the language, and initial understanding of the brain processess and mechanisms involved in the interrelated domains of motivation and emotion, of effort and will, and of self-regulation and self-consciousness. Specifically, the concepts of feedback and feedforward as they describe closed and open (helical) loop systems are useful in the formulation of a testable model of this domain of inquiry in precise, scientifically useful terms. Control theory is no foreigner to biological and neurological explanation. The term regulation is as often used as the term control, but biological principles are almost universally regulatory principles-i.e., principles invoking mechanisms of control. With regard to the issues under consideration, homeostasis and homeorhesis are thoroughly tested conceptions and biofeedback has become a household word. And to view the nervous system as an information-processing mechanism is now standard practice among neurophysiologists. The proposal derived directly from Jamesian theory states simply that emotion is essentially based on closed-loop feedbacks, while motivations go beyond these and "enter into practical relations" by way of information-processing, open-loop, feedforward mechanisms. The maintenance of "practical relations" demands repeated changes (biasing) in the constraints (the feedbacks) operating on the system; thus voluntary effort is a necessary concomitant of open-loop processes. The most generally known innovation in control theory has been the formal description of the concept of feedback (e.g., Miller, Galanter, and Pribram, 1960), a circular process initiated by a test, a matching of two settings. When there is mismatch, one of the settings becomes fixed, while the other triggers an operation which continues until a match is produced. Thus a test-operate-test-exit sequence, a TOTE, characterizes the feedback: For example, if the setting of a thermostat and that of room temperature are incongruent (mismatch) a furnace is turned either on or off until congruence is established. More recently another, equally useful conception-feedforward (e.g., see MacKay, 1969; Mittelstaedt, 1968; Pribram, 1971)-has been found important. In feedforward control, an operation procedes to a
85
SELF-CONSCIOUSNESS AND INTENTIONALITY
Load
Control centre
Motor outflow
Extraocular muscle control system
Eyeball
Load
B
Control centre
Motor outflow
Limb muscle control system
Limb
:0
Position
Joint receptors FIGURE l1(A) Eyeball position control system: (open-loop, feedforward) and (B) limb position control system: (closed-loop, feedback). (Reprinted with permission from D. J. McFarland, Feedback Mechanisms in Animal Behavior. © Academic Press, 1971.)
predetermined end point. For example, in most apartments, the furnace continues to operate for fixed periods, irrespective of local temperature conditions. The distinction between feedback and feedforward has been extremely useful in the analysis of engineering and biological systems, which ordinarily are composed of complex combinations of feedback and feedforward processes. Two types of combinations have been extensively studied. In one, feedback processes become associated or multiply linked with each other, producing an extremely stable system resistant to change (i.e., they exhibit equilibrium and inertia). An engineering example of such a system is the multilinking of power plants in the northeastern United States, which guards against frequent local disruptions, though it is vulnerable to occasional massive failure. Biologically, physiological drive systems have been found to display this type of organization. Thus food intake, muscular activity, temperature regulation, and water metabolism are interdependent regulatory mechanisms which, as a rule, operate to maintain basal temperature
86
KARL
H.
PRffiRAM
constant. The associative links which make up this and similar systems have been studied extensively (e.g., see review by Brobeck, 1963) and their operating characteristics thoroughly analyzed (Ashby, 1960). Combinations based primarily on feedforward processes are ubiquitous; they constitute our computer technology. For the most part, such combinations contain feedbacks as well. Biologically, combinations of feedforwards occur in parallel, processing signals simultaneously by virtue of overlapping neighborhood interactions, and constitute one class of cognitive processes (see Neisser, 1967; Eccles, 1967; Pribram, 1971). When feedback loops are included, hierarchical sequential arrangements called plans or programs are constituted (Miller, Galanter, and Pribram, 1960). Parallel and hierarchical processing mechanisms provide the foundations of contemporary cognitive theory. In biology, homeostatic processes, oscillating phenomena such as biological rhythms and clocks, and load-adjusting mechanisms such as those regulating muscular contraction have all been shown dependent upon feedback organization (Pribram, 1971). The essential characteristic of such systems is that they depend upon a match between two settings. A mismatch produces an error signal which controls the operation of the system until equilibrium-match-is reestablished. This homeostatic conservation of equilibrium is akin to that described by the first law of thermodynamics, which states that the conservation of energy is maintained because change elicits an "equal and opposite reaction." Energy concepts are therefore appropriate to a description and an understanding of feedback organizations and, in fact, are regularly used, as for example in the description of the "effort" or "work" involved in load-adjusting mechanisms. Information concepts, by contrast, have often been linked to the second law, and in fact, information has often been termed (e.g., by Brillouin, 1962) neg-entropy (see also von Foerster, 1965). Confusion has arisen because there has been a tendency to label "error" (the mismatch signal) information. But "error" has nothing in common with this type of "information": The amount of information contained in a message does not depend on the processing of its errors. Ashby (1963) details the distinction in terms of the constraints (limits on the independence of the functioning parts) operating on the processing system, the constraints on variety. Information is a measure of variety; redundancy (repetition), a measure of constraints. (For a comprehen-
87
SELF-CONSCIOUSNESS AND INTENTIONALITY
'OPU'-rl I I I I I
I I I I I
I I I I
I
l_1
Tcst
r
I
I I
I I
...
I
reedback
Operatc
1
l-r"" I
reedf'orward
1 I
I-j I
reedbaC}.J
Bia5
I
I
I
12. The TOTE servomechanism modified to include feedforward. Note the parallel processing feature of the TOTE.
FIGURE
sive discussion of the nature of constraint in physics and biology, see Pattee, 1971). Information thus refers to the content of a communication, while redundancy reflects the context or code in which information is communicated. Feedback organizations constrain systems to equilibrium. Thus error becomes a term denoting redundancy or lack thereof, not information. When a long-distance conversation is interrupted by a periodic whoosh, the constraint (the context in which the information is relayed) becomes disturbed-the conversation becomes unintelligible and a mismatch is conveyed to the sender, who then repeats the same information more slowly with greater emphasis and perhaps several times, changing the structure of the constraints operating during the conversation without altering its content (the amount of information). From this it follows that the invoking of biofeedback procedures accomplishes its purpose by providing an external bias on the internal feedbacks that maintain the ordinary homeostases operating in the system. The bias, maintained with effort, produces conscious voluntary control on the system, which now is an information-processing, feedforward, open-loop, helical mechanism rather than just an unconscious error-processing, feedback, closed-loop system. The outline of a model for self-regulation and self-consciousness
88
KARL
H.
PRmRAM
thus appears to be relatively easy to discern. Even the neural mechanisms involved are to some extent becoming known. With regard to the effort involved in the coordination of internal control with external demand, the evidence of the critical role of hippocampal circuit has been reviewed elsewhere (see Pribram, 1971, Chapter 15; Pribram and McGuinness, 1975). With respect to the coordinations involved in t~e maintenance of "practical relations" with the exciting event, the cerebellar circuit appears critical (Pribram, 1971, Chapter 13). Both hippocampal and cerebellar mechanisms, based on somewhat comparable anatomical structure, can be thought to perform rapid calculations of probable future states from extrapolations of present and immediately past circumstances. The remainder of the system can thus change its operations to achieve or preclude that particular estimated future state's occurring. New calculations then take place and the process is repeated, monitoring and extrapolating continuously the changes, or lack thereof, which result.
B. Attention Span and Self-Consciousness While the proposal of a plausible model of self-regulation and selfconsciousness is thus feasible, understanding the genesis of self-consciousness poses greater difficulties. Two points are clear: A change from feedback to feedforward organization effected through biofeedback procedures leads to conscious, voluntary control; we exercise this control by "paying" attention. Thus the key to understanding the genesis of self-consciousness is attention, and specifically the set of problems psychologists deal with under the rubrics of attention span and central capacity. James, in discussing the span of attention, reviewed (1950, Vol. I, pp. 427-435) the reaction-time experiments of Wundt, Exner, and Munsterberg. He concluded that the results indicate that shorter times are elicited by the following mechanisms: (1) The accommodation or adjustment of the sensory organs; and (2) The anticipatory preparation from within of the ideational centres concerned with the object to which the attention is paid ... The two processes of sensorial adjustment and ideational preparation probably coexist in all our concrete attentive acts. (Vol. I, p. 434).
Again, attention consists of: a collection of activities physiologically in no essential way different from
SELF-CONSCIOUSNESS AND INTENTIONALITY
89
the overt acts themselves. If we divide all possible physiological acts into adjustments and executions, the nuclear self would be the adjustments collectively considered. (Vol. I, p. 302)
We note here the recurring theme which differentiates "adjustments" that end within the organism's body and "executions," which go beyond into practical actions. Further, a "nuclear self" can be ascertained by consideration of the collection of adjustments which shorten reaction time. Today, reaction-time experiments have once more raised the issue of attention span and its dependence on some sort of nuclear self or competency to adjust central capacity. The issue is handled in the Pribram-McGuinness review (1975) in terms of control theoretic concepts and is worth repeating here because of its relevance to the problem of the mechanism which brings about self-consciousness.
C. Central Competency In living systems, an arousing stimulus often increases the uncertainty of the organism by its novelty. This effect of input information is contrary to that obtained in nonliving communication systems, where information conveyed always reduces uncertainty. The difference between living and nonliving systems can be conceptualized in terms of the channel over which the communication takes place. In nonliving communication systems the channel is akin to a sensorimotor channel which is fixed in capacity and does not alter with the communication. Living systems (and also computers) have the capability of memory, which alters the competence with which they process information (Pribram, 1971, Chapters 14 and 16). This is produced by the alteration of channel redundancy and superficially resembles a change in the number of channels with fixed capacity. The increase in competence is the result of an increase in the complexity of the neuronal model, an encoding process described as "chunking" the information (Miller, 1956; Simon, 1974). This and similar mechanisms in human information-processing effect a change in central processing very different from that produced by a simple increase in the number of fixed-capacity channels available. The evidence that information-processing competency can be changed in living organisms comes from a variety of problem-solving situations. Kahneman (1973), in reviewing several such studies from
90
KARL
H.
PRIBRAM
the psychophysiological literature, suggested that" arousal" is in fact an indicator of a change in capacity-"the allocation of spare capacity"much as this is changed in nonliving systems by an increase in the number of channels available. He also goes on to equate "arousal" and "capacity" with" effort" and" attention" in a global fashion. As noted, however, emotional arousal is indicative of but one sort of attention, and effort is involved only when the situation demands the regulation of arousal and motivational readiness to produce a change in information-processing competency. The way in which competency is controlled by brain systems in the living primate is demonstrated by the finding that removal of the area of the brain usually called sensory or posterior intrinsic or "association" cortex reduces the sampling of novel alternatives. The opposite effect is obtained when the lateral frontal cortex is resected. Removal of this same frontal cortex leads to an increase in behavioral orienting and an abolition of the viscerautonomic components of orienting. There thus appear to be opposite effects (posterior and frontal) on the number of alternatives sampled in a situation. This was interpreted to indicate a dual control mechanism determining the ability to sample (Pribram, 1960). Supportive behavioral evidence came from an experiment by Butter (1968, 1969), in which he investigated the number of features usually attended by monkeys while discriminating between two cues. He did this by eliminating each feature in turn in various combinations. He found that resection of the same brain region (the posterior cortex) that produced a restriction in the number of alternatives sampled also produced a restriction in the number of features used to make the discrimination. Electrophysiological evidence has been obtained that the posterior and frontal cortex contribute opposing controls on sensory channels. This evidence is based on changes produced in the recovery cycles of the system (the speed with which the system recovers to its full capacity after a sudden, intense stimulus) and the alterations produced in the shape of visual receptive fields (Spinelli and Pribram, 1966, 1967). These changes in sensory channels were, however, not attributed to a simple change in the number of channels of fixed capacity, as the effects of surgical resection have shown that as little as a few percent of an anatomically defined sensory channel is sufficient for ordinary discrimination learning, performance, and transfer (Lashley, 1929; Galam-
SELF-CONSCIOUSNESS AND INTENTIONALITY
91
bos, Norton, and Frommer, 1967; Chow, 1970). The remainder of any input channel appears to be redundant, spare channel capacity, under most circumstances. The results on the control of input channels by posterior and frontal cortex were therefore interpreted (Pribram, 1967c) as influencing redundancy, not sensory capacity in the usual information theoretic sense. Specifically, it was suggested that the input systems acted as channels in which spatial and temporal multiplexing could occur, a suggestion similar to that put forward by Lindsay (1970). On the basis of the data reviewed above, Kahneman's (1973) concept that arousal involves an increase in the number of sensory channels available can be generalized to include constraints involving the redundancy characteristics (the competency) of that capacity. Kahneman's discussion approached such a generalization when he spoke of changes in "structural connections between components." In technical language, such changes in competency would be reflected in changes in the equivocation of the channel (defined as the sum of noise and redundancy). Competency is the reciprocal of equivocation. Effort can then be defined as the measure of the attention "paid" to increase or maintain efficiency by reducing equivocation, i.e., enhancing competency.
D. External Versus Internal Constraint Gamer (1962) in his analysis of the structure of redundancy has shown that the total amount of constraint operating in any system of variables can be divided into internal and external components. Internal constraints refer to the relationships among the system of variables under consideration, while external constraints refer to the relationship between these variables and some external referent system of variables. In our neurophysiological experiments we considered the constraints that describe the central operation of the channel as internal and the constraints that refer to operations on the environmental situation which control its sensory input as external. In addition, it was found important to distinguish between temporal (repetition of the use of the channel or variable over time) and spatial (replication of the variable over space) redundancy for each of Gamer's categories. Specifically, it was suggested (Pribram, 1967c) that when the frontal system becomes involved in the orienting reaction, the internal
92
KARL
H.
PRIBRAM
redundancy in the input channel is increased so that all of the information being simultaneously processed becomes uchunked" into one unit. By contrast, when the posterior cortex becomes involved in the attentional process, internal redundancy in the input channels is decreased, separating the bits of information in each channel from each other. This is concomitant with enhancement of external redundancy, which according to Gamer's findings enhances the ability to make discriminations, i.e., to categorize input. In short, the controls on emotion and motivation operate on the mechanisms of redundancy, on the constraints operating within and between channels, rather than on the uinformation" being processed. These constraints involve a neuronal model and may be conceived of as operating on memory rather than on input information. Another way of stating this is to say that the controls operate on the representational context in which the informational content is processed. A good deal of additional evidence can be cited to show that competency rather than sensory channel capacity per se is controlled by the attentional systems discussed here. For instance, the studies of Anderson and Fitts (1958) cited by Gamer (1962) show that as much as 17 bits of sensory information can be simultaneously processed. The work of Lindsay (1970), which demonstrated the relationship between sensory discriminability (difficulty in distinguishing between inputs) and central processing competency, has already been mentioned. Pribram, Lim, Poppen, and Bagshaw (1966) and Mishkin and Pribram (1955), using various forms of the delayed alternation tasks, attributed the differential effects obtained after resections of two reciprocal frontoamygdala systems as due to selective alterations in the structure of internal redundancy (spatial and temporal, respectively) of the remaining processing competency. Further, Pribram and Tubbs (1967) have shown that when the delayed alternation task, the nemesis of monkeys with frontal-lobe resections, is externally parsed or chunked as a result of making the intertrial intervals asymmetric, the deficit is completely overcome. Similarly Wilson (1968) analyzed the trade-off between tasks involving external temporal and spatial redundancy in reciprocal mechanisms (anterior and posterior inferotemporal cortex) which have been delineated within the posterior system. Thus, both Kahneman's (1973, pp. 8, 9, 15) and this analysis attribute the control of attention to alternations in information-processing channels, not the direct control on information and uncertainty
SELF-CONSCIOUSNESS AND INTENTIONALITY
93
per se. We differ in that Kahneman focused on the problem of increasing the number of channels of fixed capacity-the" allocation of spare capacity"-while this analysis emphasizes the broader issue of competency, defined by any constraints operating on the structure of channel redundancy. We also differ in separating readiness from arousal and in that we do not identify attention, arousal, readiness, and effort as different names for the same process. Finally, the model put forward here specifies that effort accompanies only those attentional processes which result in a change in the representational organization of the information-processing mechanism. Part of the mechanism detailing how and when effort is expended during attention has been revealed by studies measuring peripheral autonomic and somatic changes. A way to picture this somewhat technical account of the model is as follows. Most psychologists today view the limitation on central processing to be due to a fixed frame (i.e., "frames of consciousness"), which limits the momentary capacity of a channel, much as does the exoskeleton of a crustacean. The model proposed here is that the limits, (the constraints) are not exoskeletal but endoskeletal-they operate by virtue of the internal structure of the channel, not by some outer shell, or "frame," that encases it. Further, the evidence from brain research as well as from behavioral research indicates that the internal skeleton is flexible: It can be reorganized into a variety of configurations. Organization involves "paying" attention and comes about in two ways: through purely mnemonic internal emotional "adjustments/l (control of internal redundancy) or through the motivational "execution of practical relations" with external events (control of external redundancy). A good deal remains to be explained. Do these observations dealing with overall central capacity and competency also apply to how the attentional mechanism becomes intentional? That is, what brain processes allow the act and actor, percept and perceiver, to be simultaneously attended? Does self-consciousness accrue simply as a dividend from the fact that central competence as a whole fluctuates around the "magical number 7/1 (Miller, 1956), or is a higher-order constraint necessary to its genesis? Is, as suggested here, the change from a homeostatic, error-processing feedback mechanism to the parallel processing of information in an open-loop mechanism sufficient explanation, or is the change from a holonomically constrained system (described by integrable differential equations) to a nonholonomic
94
KARL
H.
PRffiRAM
system the essential development? What role do the limbic forebrain structures (involved in psychomotor seizures) have in making parallel or nonholonomic processes possible? And are either parallel or nonholonomic attentional mechanisms the essence of such typically human information-processing abilities as practical skills and linguistic communication and the memory mechanisms associated with such abilities? This is as far as the outlines of the model can take us today. Further neuropsychological and neurophysiological research and even more precise formulation of the brain mechanisms of intentionality, attention, emotion, and motivation in terms of control theory are needed. But such formulations need not begin de novo. A beginning was made by William James, as we have seen. In discussing certain clinical observations, he clearly foreshadowed the endoskeletal model of competency developed here and its relationship to self-consciousness: If we speculate on the brain-condition during all these different perversions of personality, we see that it must be supposed capable of successively changing all modes of action, and abandoning the use for the time being of whole sets of well organized association-paths. In no other way can we explain the loss of memory in passing from one alternating condition to another. And not only this, but we must admit that organized systems of paths can be thrown out of gear with others, so that the processes in one system give rise to one consciousness, and those of another system to another simultaneously existing consciousness. Thus only can we understand the facts of automatic writing, etc., whilst the patient is out of trance, and the false anaesthesias and amnesias of the hysteric type ... Each of the selves is due to a system of cerebral paths acting by itself. If the brain acted normally, and the dissociated systems came together again, we should get a new affection of consciousness in the form of a third "Self" different from the other two, but knowing their objects together, as the result ... Some peculiarities in the lower automatic performances suggest that the systems thrown out of gear with each other are contained one in the right and the other in the left hemisphere. The subjects, e.g., often write backwards, or they transpose letters, or they write mirror-script. All these are symptoms of agraphic disease. The left hand, if left to its natural impulse, will in most people write mirror-script more easily than natural script ... On Hughlings Jackson'S principles, the left hemisphere, being the more evolved organ, at ordinary times inhibits the activity of the right one; but Mr. Myers suggests that during the automatic performances the usual inhibition may be removed and the right hemisphere set free to act by itself. This is very likely to some extent to be the case. But the crude explanation of "two" selves by "two" hemispheres is of course far from Mr. Myer's thought. The selves may be more than two, and the brain systems
severally used for each must be conceived as interpenetrating each other in very minute ways. [Italics mine) (1950, Vol. I, pp. 399-400)
SELF-CONSCIOUSNESS AND INTENTIONALITY
95
ACKNOWLEDGMENTS
The research involved was supported by NIMH Grant # MH1297009 and NIMH Career Award #MH15214-13 to the author.
REFERENCES ANDERSON, N. S., AND FITTS, P. M. Amount of information gained during brief exposures of numerals and colors. Journal of Experimental Psychology, 1958,56, 362-369. ASHBY, W. R. Design for a brain: The origin of adaptive behavior (2nd ed.). New York: Wiley, 1960. ASHBY, W. R. An introduction to cybernetics. New York: Wiley, 1963. BAGSHAW, M. H., AND BENZIES, S. Multiple measures of the orienting reaction and their dissociation after amygalectomy in monkeys. Experimental Neurology, 1968,20,175187. BAGSHAW, M. H., AND COPPOCK, H. W. Galvanic skin response conditioning deficit in amygdalectomized monkeys. Experimental Neurology, 1968,20, 188-196. BAGSHAW, M. H., KIMBLE, D. P., AND PRIBRAM, K. H The GSR on monkeys during orienting and habituation and after ablation of the amygdala, hippocampus and inferotemporal cortex. Neuropsychologia, 1965,3, 111-119. BAGSHAW, M. H., AND PRIBRAM, J. D. Effect of amygdalectomy on stimulus threshold of the monkey. Experimental Neurology, 1968,20, 197-202. BAGSHAW, M. H., AND PRIBRAM, K. H. Effect of amygdalectomy on transfer of training in monkeys. Journal of Comparative and Physiological Psychology, 1965,59, 118-121. BAILEY, P., AND SWEET, W. H. Effects on respiration, blood pressure and gastric motility of stimulation of orbital surface of frontal lobe. Journal of Neurophysiology, 1940,3, 276-281. BAILEY, P., VON BONIN, G., AND MCCULLOCH, W. S. The isocortex of the chimpanzee. Urbana: The University of Illinois Press, 1950. BEEBE-CENTER, J. G. The psychology of feeling. In Encyclopedia Brittanica, (Volume 9). Chicago: Benton, 1971 edition. BERNSTEIN, N. The co-ordination and regulation of movements. New York: Pergamon, 1967. BLUM, J. S., CHOW, K. L., AND PRIBRAM, K. H. A behavioral analysis of the organization of the parieto-temporo-preoccipital cortex. Journal of Comparative Neurology, 1950,93, 53-100. BRENTANO, F. Psychologie vom empirischen Stanpunkt (3rd ed.). Leipzig: 1925. BRETANO, F. The distinction between mental and physical phenomena. In R. M. CHISHOLM (Ed.), Realism and the background of phenomenology. New York: Free Press, 1960, pp. 39-61. BRILLOUIN, L. Science and information theory (2nd ed.). New York: Academic Press, 1962. BROBECK, J. R. Review and synthesis. In M. A. B. BRAZIER (Ed.), Brain and behavior (Vol. 2). Washington: American Institute of Biological Sciences, 1963, pp. 389-409. BuCY, P. c., AND PRIBRAM, K. H. Localized sweating as part of a localized convulsive seizure. Archives of Neurology and Psychiatry, 1943,50, 456-461. BUTTER, C. M. The effect of discrimination training on pattern equivalence in monkeys with inferotemporal and lateral striate lesions. Neuropsychologia, 1968, 6, 27-40.
96
KARL
H.
PRIBRAM
BUTTER, C. M. Impairments in selective attention to visual stimuli in monkeys with infero-temporal and lateral striate lesions. Brain Research, 1969,12, 374--383. CANNON, W. B. The James-Lange theory of emotions: A critical examination and an alternative theory. American Journal of Psychology, 1927,34, 106-124. CHOW, K. L. Effects of partial extirpations of the posterior association cortex on visually mediated behavior. Comparative Psychology Monographs, 1951,20, 187-217. CHOW, K. L. Integrative functions of the thalamocortical visual system of cat. In K. H. PRIBRAM AND D. BROADBENT (Eds.), Biology of memory. New York: Academic Press, 1970, pp. 273-292. . COWAN, W. M., ADAMSON, L., AND POWELL, T. P. S. An experimental study of the avian visual system. Journal of Anatomy (London!, 1961,95, 545-563. COWAN, W. M., GOTTLIEB, A. E., HENDRICKSON, J. L., PRICE, J. L., AND WOOLSEY, T. A. The autoradiographic deminstration of axonal connection in the central nervous system. Brain Research, 1972,37, 21-51. DELL, P. J. Correlations entre Ie systeme vegetatif et Ie systeme de la vie de relation; mesencephale, diencephale et cortex cerebral. Journal de Physiologie (Paris), 1952, 44, 471-557. ECCLES, J. c. Postsynaptic inhibition in the central nervous system. In G. C. QUARTON, T. MELNECHUK, AND F. O. SCHMITT (Eds.)' The neurosciences. New York: Rockefeller University Press, 1967, pp. 408-427. EVARTS, E. V. Representation of movements and muscles by pyramidal tract neurons of the precentral motor cortex. In M. D. YAHR AND D. P. PURPURA (Eds.), Neurophysiological basis of normal and abnormal motor activities. Hewlett, N.Y.: Raven Press, 1967, pp. 215-254. FIBIGER, H. c., PHILLIPS, A. G., AND CLOUSTON, R. A. Regulatory deficits after unilateral electrolytic or 6-0HDA lesions of the substantia nigra. American Journal of Physiology, 1973,225(6), 1282-1287. FULTON, J. F. Frontal labotomy and affective behavior: A neurophysiological analysis. New York: Norton, 1951. GALAMBOS, R., NORTON, T. T., AND FROMMER, C. P. Optic tract lesions sparing pattern vision in cats. Experimental Neurology, 1967,18, 8-25. GARNER, W. R. Uncertainty and structure as psychological concepts. New York: Wiley, 1962. HEARST, E., AND PRIBRAM, K. H. Appetitive and aversive generalization gradients in amygdalectomized monkeys. Journal of Comparative Physiological Psychology, 1964a, 58, 296-298. HEARST, E., AND PRIBRAM, K. H. Facilitation of avoidance behavior by unavoidable shocks in normal and amygdalectomized monkeys. Psychological Reports, 1964b, 14, 39-42. HUSSERL, E. Logische Untersuchungen (4th ed.). Halle, 1928. JAMES, W. Principles of psychology (Vols. 1 and 2 complete). New York: Dover, 1950. JOHANSSON, G. Visual perception of biological motion and a model for its analysis. Perception and Psychophysics, 1973,14(2), 201-211. KAADA, B. R. Somato-motor, autonomic and electro-corticographic responses to electrical stimulation of "rhinencephalic" and other structures in primates, cat and dog. Acta Physiologica Scandinavica, 1951, 23(Suppl. 83), 285. KAADA, B. R., PRIBRAM, K. H., AND EpSTEIN, J. A. Respiratory and vascular responses in monkeys from temporal pole, insula, orbital surface and singulate gyrus: A preliminary report. Journal of Neurophysiology, 1949,12, 347-356. KAHNEMAN, D. Attention and effort. Englewood Cliffs, N.J: Prentice-Hall, 1973, KIMBLE, D. P., BAGSHAW, M. H., AND PRIBRAM, K. H. The GSR of monkeys during
SELF-CONSCIOUSNESS AND INTENTIONALITY
97
orienting and habituation after selective partial ablations of the cingulate and frontal cortex. Neuropsychologia, 1965,3, 121-128. KLUvER, H. Behavior mechanisms in monkeys. Chicago: University of Chicago Press, 1933. KLUvER, H., AND BuCY, P. C. "Psychic blindness" and other symptoms following bilateral temporal lobectomy in rhesus monkeys. American Journal of Physiology, 1937, 119, 352-353. LASHLEY, K. S. Brain mechanisms and intelligence. Chicago: University of Chicago Press, 1929. LINDSAY, P. H. Multichannel processing in perception. In D. I. MOSTOFSKY (Ed.), Attention: Contemporary theory and analysis. New York: Appleton-Century-Crofts, 1970, pp. 149-17l. LIVINGSTON, R. B., FULTON, J. F., DELGADO, J. M., SACHS, E., JR., BRENDLER, S. J., AND DAVIS, G. Stimulation and regional ablation of orbital surface of frontal lobe. Research Publications Association for Research in Nervous and Mental Disease, 1948,27, 405-420. MACKAY, D. M. Information, mechanism and meaning. Cambridge, Mass.: MIT Press, 1969. MACLEAN, P. D. Psychosomatic disease and the "visceral brain": Recent developments bearing on the Papez theory of emotion. Psychosomatic Medicine, 1949, 11, 33~353. MACLEAN, P. D., AND PRmRAM, K. H. A neuronographic analysis of the medial and basal cerebral cortex: I. Cat. Journal of Neurophysiology, 1953,16, 312-323. MALIS, L. I., PRIBRAM, K. H., AND KRUGER, L. Action potentials in "motor" cortex evoked by peripheral nerve stimulation. Journal of Neurophysiology, 1953,16, 161-167. MANDLER, G. The conditions for emotional behavior. In D. C. GLASS (Ed.), Neurophysiology and emotion. New York: Rockefeller University Press and Russell Sage Foundation, 1967, pp. 96-102. MARSHALL, J. F., AND TEITELBAUM, P. Further analysis of sensory inattention following lateral hypothalamic damage in rats. Journal of Comparative and Physiological Psychology, 1974,86(3), 375-395. McFARLAND, D. J. Feedback mechanisms in animal behavior. New York: Academic Press, 1971. METTLER, A., AND McLARDY, T. Posterior cuts in prefrontal leucotomy: A clinicalpathological study. Journal Mental Science, 1948,94, 555-564. MILLER, G. A. The magical number seven, plus or minus two, or, some limits on our capacity for processing information. Psychological Review, 1956,63(2),81-97. MILLER, G. A., GALANTER, E. H., AND PRIBRAM, K. H. Plans and the structure of behavior. New York: Holt, 1960. MILLER, N. E., BAILEY, C. J., AND STEVENSON, J. A. Decreased "hunger" but increased food intake resulting from hypothalamic lesions: Science, 1950, 112, 256-259. MISHKIN, M., AND PRIBRAM, K. H. Visual discrimination performance following partial ablations of the temporal lobe: I. Ventral vs. lateral. Journal of Comparative Physiological Psychology, 1954,47, 14-20. MISHKIN, M., AND PRmRAM, K. H. Analysis of the effects of frontal lesions in monkey: I. Variations of delayed alternation. Journal of Comparative and Physiological Psychology, 1955,48, 492-495. MITTELSTAEDT, H. Discussion. In D. P. KIMBLE (Ed.), Experience and capacity. New York: The New York Academy of Sciences, Interdisciplinary Communications Program, 1968, pp. 46-49. NEISSER, U. Cognitive psychology. New York: Appleton-Century-Crofts, 1967. ORNSTEIN, R. E. The psychology of consciousness. San Francisco: Freeman, 1972. ORNSTEIN, R. E. The nature of human consciousness: A book of readings. San Francisco: Freeman, 1973.
98
KARL
H.
PRIBRAM
PATTEE, H. H. Physical theories of biological coordination. Quarterly Review of Biophysics, 1971,4,2, & 3, 255-276. PRIBRAM, K. H. Toward a science of neurophysiology: (Method and data). In R. A. PATTON (Ed.), Current trends in psychology and the behavioral sciences. Pittsburgh: The University of Pittsburgh Press, 1954, pp. 115-142. PRIBRAM, K. H. Comparative neurology and the evolution of behavior. In A. ROE AND G. G. SrMPSON (Eds.), Behavior and evolution. New Haven, Conn: Yale University Press, 1958, pp. 140-164. PRIBRAM, K. H. A review of theory in physiological psychology. In Annual Review of Psychology (Vol. 11). Palo Alto, Calif.: Annual Reviews, Inc., 1960, pp. 1-40. PRIBRAM, K. H. Emotion: Steps toward a neuropsychological theory. In D. C. GLASS Ed.), Neurophysiology and emotion. New York: Rockefeller University Press and the Russell Sage Foundation, 1967a, pp. 3-39. PRIBRAM, K. H. Memory and the organization of attention. In D. B. LINDSLEY AND A. A. LUMSDAINE (Eds.), Brain Function (Vol. 4). Berkeley: University of California Press, 1967b, pp. 79-122. PRIBRAM, K. H. The new neurology and the biology of emotion. American Psychologist, 1967c, 10, 830-838. PRIBRAM, K. H. The amnestic syndromes: Disturbances in coding? In G. A. TALLAND AND N. C. WAUGH (Eds.), Pathology of memory. New York: Academic Press, 1969, pp. 127157. PRIBRAM, K. H. Languages of the brain: Experimental paradoxes and principles in neuropsychology. Englewood Cliffs, N.J.: Prentice-Hall, 1971. PRIBRAM, K. H. How is it that sensing so much we can do so little? In F. O. SCHMITT AND F. G. WORDEN (Eds.), The Neurosciences Study Program, III. Cambridge, Mass.: MIT Press, 1974, pp. 249-261. PRIBRAM, K. H. The isocortex. In D. A. HAMBURG AND H. K. H. BRODIE (Eds.), American handbook of psychiatry (Vol. 6). New York: Basic Books, 1975, pp. 107-129. PRIBRAM, K. H. Holonomy and structure in the organization of perception. In Proceed-
ings of the Conference on Images, Perception and Knowledge, University of Western Ontario, 1976a. PRIBRAM, K. H. Problems concerning the structure of consciousness. In G. GLOBUS, G. MAXWELL AND 1. SAVODNIK (Eds.), Consciousness and the Brain. New York: Plenum Press, 1976b. PRIBRAM, K. H., AND BAGSHAW, M. H. Further analysis of the temporal lobe syndrome utilizing frontotemporal ablations in monkeys. Journal of Comparative Neurology, 1953, 99, 347-375. PRIBRAM, K. H., KRUGER, L., ROBINSON, R., AND BERMAN, A. J. The effects of precentral lesions on the behavior of monkeys. Yale Journal of Biology and Medicine, 1955-56,28, 428-443. PRIBRAM, K. H., LENNOX, M. A., AND DUNSMORE, R. H. Some connections of the orbitofronto-temporal, limbic and hippocampal areas of Macaca mulatta. Journal of Neurophysiology, 1950,13, 127-135. PRIBRAM, K. H., LIM, H., POPPEN, R., AND BAGSHAW, M. H. Limbic lesions and the temporal structure of redundancy. Journal of Comparative and Physiological Psychology, 1966,61, 365-373. PRIBRAM, K. H., AND MAcLEAN, P. D. A neuronographic analysis of the medial and basal cerebral cortex: II. Monkey. Journal of Neurophysiology, 1953, 16, 324-340. PRIBRAM, K. H., AND MCGUINNESS, D. Arousal, activation and effort in the control of attention. Psychological Review, 1975,82,116-149.
SELF-CONSCIOUSNESS AND INTENTIONALITY
99
PRIBRAM, K. H., AND TUBBS, W. E. Short-term memory, parsing, and the primate frontal cortex. Science, 1967,156, 1765-1767. PRIBRAM, K. H., AND WEISKRANIZ, 1. A comparison of the effects of medial and lateral cerebral resections an conditioned avoidance behavior of monkeys. Journal of Comparative and Physiological Psychology, 1957,50, 74-80. ROBERTS, W. W. Are hypothalamic motivational mechanisms functionally and anatomically specific? Brain, Behavior and Evolution, 1969,2, 317-342. ROSVOLD, H. E., MIRSKY, A. F., AND PRIBRAM, K. H. Influence of amygdalectomy on social interaction in a monkey group. Journal of Comparative and Physiological Psychology, 1954, 47, 173--178. RucH, T. c., AND SHENKIN, H. A. The relation of area 13 on orbital surface of frontal lobes to hyperactivity and hyperphagia in monkeys. Journal of Neurophysiology, 1943, 6, 349-360. SCHACHTER, S. Cognitive effects on bodily functioning: Studies of obesity and eating. In D. GLASS (Ed.), Neurophysiology and emotion. New York: Rockefeller University Press and Russell Sage Foundation, 1967, pp. 117-144. SCHWAR1ZBAUM, J. S. Changes in reinforcing properties of stimuli following ablation of the amygdaloid complex in monkeys. Journal of Comparative Physiological Psychology, 1960a, 53, 388--395. SCHWAR1ZBAUM, J. S. Response to changes in reinforcing conditions of bar-pressing after ablation of the amygdaloid complex in monkeys. Psychological Reports, 1960b, 6,215221. SCHWAR1ZBAUM, J. S. Some characteristics of amygdaloid hyperphagia in monkeys. American Journal of Psychology, 1961,74, 252-259. SCHWAR1ZBAUM, J. S., AND PRIBRAM, K. H. The effects of amygdalectomy in monkeys on transposition along a brightness continuum. Journal of Comparative and Physiological Psychology, 1960,53, 396-399. SCHWAR1ZBAUM, J. S., WILSON, W. A., JR., AND MORRISSETTE, J. R. The effects of amygdalectomy on locomotor activity in monkeys. Journal of Comparative and Physiological Psychology, 1961, 54(3), 334-336. SIMON, H. A. How big is a chunk? Science, 1974,183, 482-488. SPINELLI, D. N., AND PRIBRAM, K. H. Changes in visual recovery functions produced by temporal lobe stimulation in monkeys. Electroencephalography and Clinical Neurophysiology, 1966,20,44-49. SPINELLI, D. N., AND PRIBRAM, K. H. Changes in visual recovery functions and unit activity produced by frontal and temporal cortex stimulation. Electroencephalography and Clinical Neurophysiology, 1967,22, 143--149. TEITELBAUM, P. Sensory control of hypothalamic hyperphagia. Journal of Comparative and Physiological Psychology, 1955,48, 156-163. UNGERSTEDT, U. Brain dopamine neurons and behavior. In F. G. SCHMITT AND F. O. WORDER (Eds.), The Neurosciences Study Program, III. Cambridge, Mass.: MIT Press, 1974, pp. 695-704. VALENSTEIN, E. S. Stability and plasticity of motivation systems. In G. C. QUARTON, T. MELNECHUK, AND G. ADELMAN (Eds.), The neurosciences. New York: Rockefeller University Press, 1970, pp. 207-217. VALENSTEIN, E. S., Cox, V. c., AND KAKOLEWSKI, J. W. The hypothalamus and motivated behavior. In. J. T. TAPP (Ed.), Reinforcement and behavior. New York: Academic Press, 1969, pp. 242-285. VON FORESTER, H. Memory without record. In D. P. KIMBLE (Ed.), The anatomy of memory. Palo Alto, CaU£.: Science and Behavior Books, 1965, pp. 388-433.
100
KARL
H.
PRIBRAM
P. D., AND PRmRAM, K. H., Trigeminal neurotomy and blood pressure responses from stimulation of lateral cerebral cortex of Macaca mulatta. Journal of Neurophysiology, 1950,13, 409--412. WARD, A. A., JR. The cingular gyrus: Area 24. Journal of Neurophysiology, 1948,11, 13-23. WILSON, M. Inferotemporal cortex and the processing of visual information in monkeys. Neuropsychologia, 1968, 6, 135-140.
WALL,
3
Self-Regulation of Stimulus Intensity: Augmenting/Reducing and the Average Evoked Response MONTE BUCHSBAUM
I.
INTRODUCTION
Clinical practice in an emergency room quickly dramatizes individual differences in pain tolerance. I remember a Swedish carpenter who, declining analgesia, stoically allowed me to dig out a splinter from under his fingernail with a scalpel as he gaily discussed baseball. Holy men rest on their beds of nails; Lesch-Nyhan patients mutilate themselves; rock groups blast listeners with sound above normal auditory pain threshold-all of which raises the question, how do combinations of experimental and neurophysiological mechanisms work together to produce these variations in tolerance of extreme intensities of sensory input? This article will review individual difference research in sensory overload with especial interest in the possible electrophysiological correlates of these differences.
II.
SENSORY EXPERIENCE AND AUGMENTING/REDUCING
A. Petrie and Kinesthetic Figural Aftereffects Petrie (1960), following studies of individual differences in pain tolerance, classified subjects into "augmenters" and "reducers" based on their performance on a kinesthetic figural aftereffects task (KF A). Augmenters were those individuals who typically judged the magniMONTE BUCHSBAUM . Unit on Perceptual and Cognitive Studies, Adult Psychiatry Branch, Division of Clinical and Behavioral Research, Intramural Research Program, National Institute of Mental Health, U.S. Department of Health, Education and Welfare, Bethesda, Maryland.
101
102
MONTE BUCHSBAUM
tude of a standard stimulus as larger after kinesthetic stimulation; reducers judged the standard as markedly reduced. Augmenters were found to evidence a special sensitivity to pain. In contrast, reducers had a notable tolerance for pain (Petrie, Collins, and Soloman, 1958; Petrie, 1960). Several studies have tended to support this concept (Blitz, Dinnerstein, and Lowenthal, 1966; Sweeney, 1966; Ryan and Foster, 1967). Petrie (1960) also suggested that if the tolerance of reducers for pain is partially due to their tendency to diminish the intensity of stimulation they receive, they should be at a disadvantage in sensory deprivation, where they would further reduce the already limited stimulation available. She found that augmenters could tolerate more hours in a tanktype respirator than reducers, who could not tolerate this stress. Augmenters could be pushed in the reducing direction by bombardment with continuous loud noise (Petrie, Holland, and Wolk, 1963). Since auditory stimulation could cause kinesthetic augmentation, she concluded that some central nervous system (CNS) mechanism regulated levels of sensory input. Petrie's KFA procedure was adapted from an apparatus used by Kohler and Dinnerstein (1947) to study effects of what they termed satiation-the tendency for persons to judge a stimulus as less intense after they had been stimulated by a more intense stimulus. Petrie compared size judgments of the width of a standard block of wood before and after subjects stimulated their fingers by rubbing either a wider or a narrower block of wood. Finding that the reducer would tend to judge the standard block of wood as smaller not only after rubbing the wide block of wood but after rubbing a narrow block of wood as well, she concluded that satiation theory alone could not explain the observed patterns of individual differences, and the augmenter/reducer construct was developed. Reducers thus were defined as consistently underestimating the size of the standard after stimulating their hands with any size block, and augmenters were defined as consistently overestimating the size of the standard. Careful consideration of the Petrie task from today's vantage point of research on individual differences reveals a potential both for unreliability and for measuring several types of individual differences at once. First, initial size or width judgments themselves, before stimulation, usually show individual tendencies to over- or underestimate. These size-estimation individual differences show moderate reliability across sessions (r = 0.76, Platt, Holzman, and Larson, 1971), but differ-
SELF-REGULATION OF STIMULUS INTENSITY
103
ences between values of such reliability will have much lower reliability (Platt et al., 1971). Second, individual differences in contrast effects (judging the standard as smaller after rubbing the large block and larger after rubbing the small block) are probably much more prominent than Petrie credited (1967). Hilgard, Morgan, and Prytulak (1968) found large-group contrast effects, not augmenting/reducing, predominating in their sample. Broadhurst and Millard (1969) and Schooler and Silverman(1971) found nonsignificant correlations between large- and smallblock augmenting/reducing scores. Since individuals with large contrast effects would show negative correlations and individuals with small contrast effects would show positive ones, these negligible correlations are expected. In our own studies, individual consistencies in large-blocklsmall-block contrast effects and in ascending/descending starting-position effects were much more reliable than the augmenting/ reducing (prestimulation minus poststimulation) measure. Silverman's (1964) KFA procedure was similar to the Petrie in this respect. Contrast effects also seem to operate across days (Hilgard et al., 1968), further interfering with large-blocklsmall-block differences. While it might be argued that contrast effects are balanced out from the augmenting/ reducing score by the opposite directions of contrast in the use of the small and large blocks, the assumptions that contrast effects are entirely linear, equivalent in both size directions, and independent of augmenting/reducing are probably unjustified. Thus since size estimation and contrast effects were tapped as well as the hypothetical augmenting/reducing dimension, the Petrie procedure might give very different scores or reliabilities across patient groups or experimental treatment depending upon the relationships of these perceptual styles. These problems may help to explain the failure of Morgan, Lezard, Prytulak, and Hilgard (1970) to confirm the pain tolerance of reducers and the failure of Peters, Benjamin, Helvey, and Albright (1963) to confirm the sensory deprivation tolerance of augmenters.
B. Evoked Responses and Augmenting/Reducing The concept of augmenting/reducing still remained as an intriguing source of predictions about individual differences in neurophysiological data on sensory response to stimuli at different intensities. Average
104
MONTE BUCHSBAUM
FIGURE 1. Average evoked responses (AERs) to four intensities of flash: top to bottom, dim to bright. The individual on the left shows "augmenting"-increasing amplitude with increasing stimulus intensity for component PI (PI00)-Nl (N140). The individual on the right shows "reducing"-decreasing amplitude with increasing stimulus intensity for PI-Nl (from Buchsbaum and Pfefferbaum, 1971).
evoked response (AER) amplitudes showed striking differences in their behavior at high intensities, certain individuals showing marked drops as intensity increased (Figure 1). Was this AER reducing in any way analogous to Petrie's concept of a CNS mechanism regulating sensory input? Early studies tended to show correlations between AER intensity effects and KFA (Buchsbaum and Silverman, 1968; Spilker and Callaway, 1969; Borge, 1973), and AER measures of augmenting/reducing have been used by Soskis and Shagass (1974), Hall, Rappaport, Hopkins, and Griffin (1973b), and Zuckerman, Murtaugh, and Siegel (1974). Studies in cats (Hall, Rappaport, Hopkins, Griffin, and Silverman, 1970) and monkeys (Redmond, Borge, Buchsbaum, and Maas, 1975) have been possible using AER techniques. Parallels between Petrie's KFA results and AER results also appeared. AER reducers were more tolerant of electric shock (see Section VI.A). Women were more augmenting than men (Petrie, 1960; Buchsbaum and Pfefferbaum, 1971).
SELF-REGULATION OF STIMULUS INTENSITY
105
Schizophrenics were reducers (Petrie et a/., 1963; Buchsbaum and Silverman, 1968; Landau, Buchsbaum, Carpenter, Strauss, and Sachs, 1975). These and other studies detailed in the following sections suggest some predictive utility in the augmenter/reducer dimension. Recently, even Petrie herself (1974) has hailed the use of AER techniques for measuring augmenting/reducing. We now tum to focus on the AER as a possible indicator of the profound differences in sensory response and its regulation.
III.
AMPLITUDE/INTENSITY RELATIONSHIPS IN MAN
A. Visual AERs The amplitude of the visual AER tends to increase with increasing stimulus intensity up to some intermediate intensity. Further increases in stimulus intensity may yield no further increase in amplitude or actual decreases in amplitude. This tendency for amplitude of visual evoked potentials to saturate or reduce was reported as early as 1937 by Cruikshank in a subject whose responses were big enough to observe without averaging. In one of the first systematic studies of the properties of visual AERs, Tepas and Armington (1962) found that "maximum responses were not obtained with the strongest stimuli" for large stimulus fields. Similarly, reducing was observed by many authors (Tepas, Armington, and Kropfl, 1962; Rietveld, 1963; Wicke, Donchin, and Lindsley, 1964; Rietveld and Tordoir, 1965; Vaughan and Hull, 1965; Vaughan, 1966; Buchsbaum and Silverman, 1968) and in the AERs of some authors themselves (Shipley, Jones, and Fry, 1966). Prominent individual differences were noted by Rietveld (1963) and first systematically studied by Shagass, Schwartz, and Krishnamoorti (1965) who found higher amplitude/intensity slopes in psychiatric patients. Decreases in amplitude with increasing stimulus intensity appear to be related to central factors rather than merely reflecting poor stimulus control or peripheral adjustment mechanisms. Since subjects can show quite linear amplitude increases in the electroretinogram while showing reducing in the AER (Armington, 1964a, 1964b), we know that the intensity information can register at the retina and still yield nonlinear AER amplitude/intensity functions. Similarly, DeVoe, Ripps, and Vaughan (1968) reported linear latency decreases for PlOD (PI), together
106
MONTE BUCHSBAUM
with reducing in the amplitude of the P100-N140 (P1-N2) component, and other authors have also noted this latency/amplitude difference (e.g., Wooten, 1972; Clynes, Kohn, and Lifshitz, 1964; Vaughan and Hull, 1965; Figure 6, P100 (P1)-N140 (N2) in Vaughan, 1966). Peripheral factors such as pupillary diameter do not seem to explain reducing since the vertex AER is uninfluenced by pupillary diameter in humans (Kooi and Bagchi, 1964; Richey, Kooi, and Waggoner, 1966). Visual AER-reducing is unchanged by pilocarpine instilled in the eye to fix pupillary diameter (Soskis and Shagass, 1974), and in animals such as the pigeon, no AER changes with pupillary diameter were observed (Samson and Young, 1973). Studies in our laboratory of abeltalipoproteinemia, which causes retinal dysfunction, also failed to show vertex AER changes either on baseline or on treatment with vitamin A. Maneuvers such as averting or covering the eyes do not seem to be involved. The finding of AER reducing from vertex but not from occipital leads and in PlOD more than in P200 (Buchsbaum and Pfefferbaum, 1971), as well as the high correlation between amplitude/ intensity slopes obtained with randomized and blocked presentation of stimulus intensities (Buchsbaum, Landau, Murphy, and Goodwin, 1973), all tend to mitigate against the influence of such peripheral factors. Even under the rigid stimulus control of flaxedil administered to cats wearing corneal contact lenses with artificial pupils, single-unit recordings from the lateral geniculate showed some units decreasing with increasing stimulus intensity (Hamasaki and Winters, 1973). AER differences between augmenters and reducers do not appear to result from differences in habituation rates or interstimulus interval effects. Reducers do not habituate their AERs faster to high-intensity stimuli; if anything they habituate more slowly. Reducing the interstimulus interval also does not diminish augmenter/reducer differences (Buchsbaum and Pfefferbaum, 1971). As with discrete stimuli, when rapid sinusoidal modulation of light is used, individual differences in amplitude/intensity or amplitude/depth of modulation relationships are prominent (Kitajima, 1967; Pfefferbaum and Buchsbaum, 1971). The cortical response may decrease as the intensity of stimulation increases (Il'yanok, 1961; van der Tweel and Verduyn Lunel, 1965; Montagu, 1967; Regan and Beverley, 1973). This reducing does not seem to be due to poor stimulus control. Regan and Beverely (1973) presented their stimuli in a maxwellian-view opti-
SELF-REGULATION OF STIMULUS INTENSITY
107
cal system and still found saturation for first harmonic components and marked reducing for second harmonic components.
B. Auditory AERs As was found for visual AERs, auditory AERs generally increase in amplitude with increasing stimulus intensity (e.g., Abe, 1953; Keidel and Spreng, 1965; Suzuki and Taguchi, 1965). There are, however, great individual differences in the rate of increase (Davis, Mast, Yoshie, and Zerlin, 1966; Rapin, Schimmel, Tourk, Krasnegor, and Pollak, 1966; Davis and Zerlin, 1966; Davis, Bowers, and Hirsh, 1968; Rothman, 1970). Many investigators have reported that some individuals show a maximum response at around 70-75 dB with further increase in stimulus intensity causing a decrease in AER amplitude (Davis and Zerlin, 1966; Beagley and Knight, 1967; Moore and Rose, 1969; Picton, Good. man, and Bryce, 1970; Kollar, 1971; Marco, 1972; Khechinashvili, Kevanishvili, and Kajaia, 1973) or a leveling in amplitude (Butler, Keidel, and Spreng, 1969). Rose and Ruhm (1966) even found some amplitude decreases at levels as low as 40 dB. Picton et al. (1970) recorded stapedius muscle reflexes and concluded that this paradoxical reduction in AER amplitude did not result from peripheral factors but that the most probable explanation seemed to involve either the ipsilateral cochlear efferent, the uncrossed efferent olivocochlear, or the central descending auditory inhibitory system. Khechinashvili et al. (1973) also argued that middle-ear muscle contraction was not important in AER amplitude decrease at high intensities since the decrease became most marked at higher tone frequencies (e.g., 4000 Hz), whereas contractions of the stapedius muscle predominantly affect the conduction of lower frequencies. The absence of any significant difference in amplitude/intensity slope between serial and randomly presented tone intensities (Henry and Teas, 1968) also argues against a peripheral mechanism. Reducing appears with the good stimulus control available with earphones (e.g., Khechinashvili et al., 1973) or bone conduction (Liebman and Graham, 1967). Even in anesthetized cats, single units in the cochlear nucleus were found where the "greatest excitation was provided by tones of moderate intensity and further increases in intensity resulted in a marked decline in the response" (Lavine, 1971).
108
MONTE BUCHSBAUM
3
Stage 3
~
z z
2
0
~
Stage 4
<{
:>w 0 Z
Awake
<{
w :!:
Rem " . ; : Stage 2 Stage 1 -o~~------~------~----~~
50
60
70
80
STIMULUS INTENSITY IN dB 2. Mean amplitude (deviatiori in microvolts) for P200 (168-248 ms) for four click intensities for awake, REM sleep, and Stages 1-4.
FIGURE
The great decrease of reducing that we observed (Buchsbaum, Gillin, and Pfefferbaum, 1975) during Stages 3 and 4 sleep is of interest because of the well-known loss of inhibitory mechanisms during deep sleep. Click AERs to four intensities were recorded during all-night sleep recordings. AERs collected over I-minute epochs were sorted by sleep stage in nine normal subjects. Figure 2 illustrates that awake, Stage 1, REM, and Stage 2 showed little increase in AER amplitude with increasing stimulus intensity, whereas Stages 3 and 4 showed a marked increase. Tepas, Boxerman, and Anch (1972) suggested that the "failure" of some studies to produce increasing linear relationships between AER amplitude and stimulus intensity results from the failure to control stimulus conditions and parameters and not from individual differences. However, even in their own report of carefully collected data for the B-C component for bipolar lead Fz-Pz (similar latency to P100N140) one of their three subjects showed a correlation between amplitude and stimulus intensity of + .87 (augmenting) and one of -.64
SELF-REGULATION OF STIMULUS INTENSITY
109
(reducing). They further suggested that lead Cz-Oz produces the most linear amplitude/intensity functions, but this was not borne out in another report from the same laboratory (Klingaman and Anch, 1972). Data pooled across three subjects showed a 15% drop in AER B-C amplitude (about P80-N110) as stimulus intensity increased from 75 to 85 dB. Similarly, in a recent report (Schweitzer and Tepas, 1974), two of the three subjects showed trivial (.21 and .10) correlations between stimulus intensity and AER amplitude for component B-C, for lead Cz-Oz.
C. Somatosensory AERs As with visual and auditory AERs, reducing or saturation has been observed in both human and animal recordings. Since no mechanical peripheral adjustment exists for the somatosensory system analogous to the pupil or the stapedius muscle, and since electrical stimuli are more easily delivered with great precision, the presence of reducing should provide strong evidence for central inhibitory processes. Mark and Steiner (1958) recorded from cat somatosensory receiving areas and the superficial radial nerve: they found that the cortical response reached a ceiling at only 40% of the radial nerve volley and suggested that "some peculiarity of the complex cortical responding mechanism may be involved." Early human studies by Uttal and Cook (1964) also showed AER amplitude reaching maximal levels with relatively weak stimuli. Rosner and Goff (1967) reported that the amplitude/intensity relationship for somatosensory AERs was best fitted by two straight-line functions and that the slope of the second line was nearly zero for certain components (perhaps analogous to our time bands 76-112 and 168-248 ms). Beck and Rosner (1968) later reanalyzed these data and concluded that a single-power function with a threshold correction was a more parsimonious explanation. However, prominent individual differences in amplitude/intensity slopes, with some subjects showing reducing, were reported by Uttal and Cook (1964), Shagass and Schwartz (1964), Debecker and Desmedt (1964), Ikuta (1968), Shagass, Overton, Bartolucci, and Straumanis (1971), and Mushin and Levy (1974). Saturation or reducing phenomena also appear in single-unit recordings in the dorsal hom of the spinal cord (Wall, 1967; Hillman and Wall, 1969; Lundberg and Oscarsson, 1961), where the response to
110
MONTE BUCHSBAUM
higher intensities is apparently reduced by descending inhibitory impulses originating in the brain. Such descending impulses may originate in the cerebral cortex (Hagbarth and Kerr, 1954; Anderson, Eccles, and Sears, 1964), the diencephalon (Calma, 1966), or the reticular formation (Melzack, 1973) and may act upon thalamic relay nuclei as well as spinal levels (Mountcastle, 1974). Such descending inhibitory mechanisms would be capable of modulating nociceptive afferent input before it reaches the cortex and are a part of the well-known gate-control theory of pain advanced by Melzack and Wall (1970).
D. Summary of Amplitude/Intensity Relationships Unfortunately the plethora of recording and stimulating techniques in the studies discussed above and the absence of cross-modal studies within the same individual make a comprehensive summary very difficult. Nevertheless in a general way for visual, auditory, and somatosensory stimuli, AER amplitudes initially increase with increasing intensity; further increases in intensity often bring about a decrease in amplitude in many individuals. These decreases seem not to reflect poor stimulus control, peripheral adjustments, or transducer response, but rather result from underlying CNS mechanisms. The P100-N140 component frequently reveals these paradoxical intensity effects. Parallel nonlinearities are recordable from single units in the lateral geniculate, cochlear nuclei, and dorsal horn of the spinal cord. Three types of neural pathways may be involved: descending inhibitory, nonspecific arousal, and cortical-cortical. Descending inhibitory pathways have been implicated in the somatosensory system and are certainly active in the auditory system. However, if very early signals from primary sensory areas do not show reducing but later components do, then descending inhibitory pathways seem less likely to be the major cause. With auditory signals there is some evidence that an almost linear function may exist for a Pll-N21 ms component (Madell and Goldstein, 1972) with far less linearity for later components. Early tactile stimulus AERs also apear to produce quite log-linear amplitude/intensity functions (Franzen and Offenloch, 1969) but electrical stimuli (Mushin and Levy, 1974) seem clearly to show reducing. Thus descending inhibitory systems appear to be involved in somatosensory AER-reducing and are possibly less prominent in auditory AERs.
SELF-REGULATION OF STIMULUS INTENSITY
111
Nonspecific arousal systems, including subcortical structures, were suggested as a possible mechanism in AER augmenting/reducing by Zuckerman, Murtaugh and Siegel (1974). They pointed out that fibers originating in the cortex may regulate brainstem arousal systems (Koella and Ferry, 1963; Dell, Bonvallet, and Hugelin, 1961). These systems could then modulate AER amplitude in the cortex (Knispel and Siegel, 1972, 1973). Cortical-cortical systems as a hypothetical cause of AER reducing also cannot be ruled out. The many experiments in humans showing the influence of complex logical tasks on AER amplitude certainly suggest the operation of such systems. Descending inhibitory, nonspecific arousal, and cortical-cortical systems may work together to modulate sensory input levels depending on the vigilance, arousal, and other needs of the organism.
IV.
AUGMENTING/REDUCING RELIABILITY AND THE MEASUREMENT OF THE AER
The amplitude/intensity slope appears to be a relatively stable individual characteristic. Exactly how reliable it is depends on the method of AER measurement. Most authors have relied upon the visual identification of a triphasic wave in the vertex AER to visual stimuli; a positive peak at 70-110 ms (P100 or P1), a negative peak at 110-160 (N140 or N1), and a positive peak at 180-250 (P200 or P2). The N140P200 component in our experience is usually the most unambiguous, with the P100 often split into two peaks, or appearing as a point of inflection between some early (30-50 ms) component and N140. Since we have generally had one or more sets of AERs collected at four intensities, we have tried to identify P100-N140-P200 as components most reliably appearing in the largest number of AER curves. The latency of these components is then used to decide peak identification for the most ambiguous curves. Using such a strategy, we examined AER amplitude/intensity slope reliability across two sessions about 2 weeks apart in various groups. Test-retest correlations for peak-to-trough measured P100-N240 were .67 in a group of psychiatriC patients (Buchsbaum, Goodwin, Murphy, and Borge, 1971) and .52 in a group of 120 normal adult twins (Buchsbaum, 1974).
112
MONTE BUCHSBAUM
Somewhat higher reliabilities (.72-.91) were reported by Stark and Norton (1974), who, after comparing reliabilities of latency, amplitude, and amplitude/intensity slope, concluded that slope was "clearly the most reliable AER parameter employed." Soskis and Shagass (1974) found within-session reliability of .66 and concluded that eye movement, blinks, or pupillary diameter had little influence on AER augmenting/reducing. Despite these reasonable reliabilities, however, AER component identification is troubled by missing or aberrant components, especially when large groups of subjects are studied and analogous components must be identified in four or more AERs within a person. Hall, Rappaport, Hopkins, and Griffin (1973a) suggested a helpful computer algorithim for peak identification but were still troubled by peaks disappearing on one or more of a set of AERs. They noted that a point of inflection was frequently present at exactly the latency that a peak Was clearly present for a stimulus of greater or lesser intensity. This suggests that numerical techniques (not based on visual peak identification) developed by Shagass (1972, area integration) and Dustman and Beck (1965, perimeter) might be useful. For studies of individual differences in perceptual style, a measure should be reliable over time. Because of our interest in finding possible genetic markers of vulnerability to psychiatric illness, we also wanted measures showing similarities across genetically identical individuals-i.e., monozygotic twins. A variety of AER measurement techniques were assessed in a group of 64 normal adult twin pairs-32 monozygotic (Mz) and 32 dizygotic (Dz). Each subject was tested on the four-intensity AER procedure on two separate occasions, generally 2 weeks apart. This consisted of presenting four intensities of light in a randomized order behind a translucent plastic screen. The stimuli were generated by fluorescent tubes under computer control and had a rise time of 3 ms with a duration of 500 ms so that only "on" responses were studied. AER peaks were identified visually and by use of several numerical techniques. We calculated the amplitude/intensity slope by fitting a straight line to the amplitude measures (however obtained) using least-squares technique. The visually identified peaks yielded either peak-to-trough measures, P100-N140 and N140-P200, or prestimulus baseline-to-peak measures, P100, N140, and P200. For the numerical techniques, we chose three time intervals centered on the three peak:; identified on visual inspection: P100 (76-112 ms), N140 (116-152
a
Peak-la-trough PIOO-N140 and N140-P200.
Visual inspection" Peak-to-trough Baseline-to-peak Area Absolute area relative to mean Signed area relative to pre stimulus baseline RMS Perimeter Point at 100, 140, and 200 ms .51 .29 .63 .45 .52 .35 .71
.63 .46 .61 .18 .43
MZ twins
.52 .28
Test-retest
P100 76-112 ms
.53 .47 .65 .58 .65 .55 .63
-.01 -.10 .62 .60 .69
.37 .32 .36 .54 .37
Test-retest
.05
MZ twins
.72
.53 .36
.60 .55
.56 .28
MZ twins
P200 168-248 ms
.25
Test-retest
N140 116-156 ms
1 Amplitude/Intensity Slope Test-Retest and Mz-Twin Pair Correlations TABLE
t'i
CIl
....... ....... VJ
~
'"
z
![
CIl
..,
o>,j
z
S o
c::
~
.... 'l'
114
MONTE BUCHSBAUM
ms), and P200 (168-248 ms). They correspond fairly well to time intervals chosen by Lewis, Dustman, and Beck (1972), and the critical P100 band is similar to that used by So skis and Shagass (75-150 ms). Two techniques of areal integration were compared. For the first, termed "absolute area relative to mean," the mean value of each AER, used as a baseline, was subtracted from each AER coordinate; within each time band, the mean of the absolute values of successive coordi-
0.7
!z 0.6
LU
u 1-Li:0.5 "' ..... LULU t;J 8 0.4 a::;z
~~0.3 "'« LU....J I-
~ 0.2 a::
-p<0.05
80.1
0 I-
'" ~
§; 3 0
a::
u
:i ~
a::
LU (J)
::;:
«
....J
I-
0
~
2
0
LU
....J
....J
;;
'" ~
~ -I
~-I a:: u
~
LJ.J
::;:
0
0
:E -2
3. Analysis of AER amplitude and slope for consecutive 4 ms time intervals. The actual AER value for the highest-intensity AER and the slope for each coordinate (calculated across four intensities) were obtained for each of two AER sessions on 128 subjects. Data are shown for 300 ms following the stimulus mean across the 128 subjects (lower), and test-retest correlations across the two sessions (upper) are presented for amplitude data (left) and slope data (right). Thus the curve in the lower left is the mean AER for the group. Circled points are the mean latencies (across intensities) for P100, N140, and P200 for the entire group as determined by visual inspection. Note that P100 and P200 fall on maxima for test-retest reliability (above) both for raw amplitude (left) and slope (right). P100 falls on the minimum for the slope function (lower right) and P200 on the maximum; note that the P100 slope value is actually negative as well as reliable. (Data from Buchsbaum, 1974.) FIGURE
SELF-REGULATION OF STIMULUS INTENSITY
115
nates was calculated. For the second, "signed area relative to prestimuIus baseline," the pre stimulus baseline was subtracted from each AER coordinate and the mean of the signed values of successive coordinates was calculated. A root-mean-square measure was also calculated. The line-length measure was the sum of the absolute values of point-topoint differences. In addition, the individual AER points (after removal of AER mean) were used. Table 1 compares the test-retest reliability and Mz-twin similarity for the seven techniques. All techniques showed statistically significant test-retest correlations for PIOO and P200 (for r > .16, p < .05, 1 tailed). Positive peaks PIOO and P200 appear to have more reliable amplitude/intensity slopes than negative peak N140. Measurements referred to the mean level of the AER appear to be more reliable than those referred to the prestimuIus baseline. Generally reliability and Mz-twin similarity are in high agreement. Overall, the absolute area relative to the mean technique (closest to that of Shagass, 1972) seems to have the highest reliability and heritability, and we have adopted this technique. The data from the point-by-point analysis on the total group of 128 subjects indicate that considerable information is carried in the single AER coordinates (Figure 3). The circled dots are the mean latencies (as visually identified) for PIOO, N140, and P200. For both amplitude and slope, the pattern of higher test-retest correlations for positive peaks is seen. The plot for the mean slope illustrates that the minimum slope on the entire curve coincides exactly with PIOO--and is actually negative, indicating that AER values were lower for high-intensity flashes than for low ones. The high positive peak at P200 is consistent with our reports and those of others (see Section IILA) that P200 shows far higher slopes than PIOO.
v.
GENETIC FACTORS IN AUGMENTING/REDUCING
A. Twin Studies For the visual AER, amplitude/intensity slopes appear to be partially set genetically as we have seen above and as is illustrated in Figure 4. Heritability estimates for the AER amplitude/intensity slope on a group of 30 Mz and 30 Dz normal adult twins ranged from .52 for the PIOO--Nl40 component measured visually to .68 for certain quantita-
116
TWIN PAIR I
MONTE BUCHSBAUM
TWIN PAIR 2
FIGURE 4. Visual AERs in two pairs of Mz twins. Illustrated are AERs to four intensities of light, as in Figure 1. For each pair, the AER is shown as a solid line for one twin and as a dotted line for the co-twin. Note similarity in latency of peaks, waveform, and changes with stimulus intensity. For Pair 1 on the left, component P100-N140 increases markedly with stimulus intensity (augmenting) and for Pair 2 the component decreases with stimulus intensity (reducing) (Buchsbaum, 1974).
tive measures (Buchsbaum, 1974). These heritabilities reflect relatively high intraclass correlations in Mz twins and often absent or negligible correlations in Dz twins. Other investigators studying amplitude measure in the visual AER have reported similar results (Lewis, Dustman, and Beck, 1972). Osborne (1970) also found very low (.08 to .12) intraclass correlations in Dz twins for certain AER measures, together with much higher correlations in Mz twins. Similarly for EEG frequency spectra, Lykken, Tellegen, and Thorkelson (1974) found intraclass correlations of about .8 for Mz twins and near zero for Dz twins. These findings tend to suggest that the augmenting/reducing similarities between Mz twins are not primarily due to additive genetic or dominance effects since such effects would create similarities between Dz twins as
SELF-REGULATION OF STIMULUS INTENSITY
117
well. Epistasis or dominance interactions involving several loci or disassortative mating could be candidates for such effects. Alternatively, common environmental factors might act to increase similarity in Mz twins and decrease similarity in Dz twins (Lykken et al., 1974).
B. Sex and Chromosome Differences We have generally found that women have greater amplitude/ intensity slopes for P100-N140 than men for visual AER (Buchsbaum and Pfefferbaum, 1971). Our normal twin sample also showed significantly greater amplitude intensity slopes in females than in males for P100-N140i no difference was found for N140-P200. Patients with chromatin-negative gonadal dysgenesis (Turner's syndrome, XO) showed even greater amplitude/intensity slopes than XX normal females (Buchsbaum, Henkin, and Christiansen, 1974). Again the sexdifference effects appeared for P100-N140, not N140-P200.
VI.
TOLERANCE FOR HIGH-INTENSITY STIMULATION
Petrie's original conception that reduction was an adaptive, protective mechanism for withstanding sensory overload was only weakly borne out by studies using the KFA test, as we have earlier seen. However, two current AER studies suggest that AER reducers are indeed pain- and noise-tolerant.
A. Pain Tolerance Individual differences in pain response were studied by the use of brief electrical stimuli (constant current 1 ms square waves) administered to the dorsum of the left forearm with the Tursky electrode (Tursky and Watson, 1964). Stimuli of four different intensities (2, 9, 16, and 23 mA) were delivered 1 s apart with 64 randomized presentations of each intensity, and AERs to these stimuli were collected. Subjective pain ratings were also obtained in a separate run in which stimuli (1 to 31 mA in 1-mA steps) were presented three times each in random sequence. Subjects
118
MONTE BUCHSBAUM
rated each stimulus on a 4-point scale (1 = just noticeable, 2 = distinct, 3 = unpleasant and 4 = painful) using a special keyboard. EEG was recorded from vertex (Cz, right ear), and AER amplitude was measured using the "absolute area relative to mean" technique (see Section IV) over two time bands centered on PlOD (76-112 ms) and P200 (168-248 ms). Shock-response rating data were analyzed according to simple nonparametric analogues of signal detection parameters Cx and d'. Since subjects differed widely in the shape, symmetry, and size of the dispersion of their judgments of the four categories, the usual mathematical assumptions underlying signal detection analysis did not seem well met. A measure of response criterion (Cx) was obtained for the distinct/unpleasant categories by location of the stimulus strength for which the least overlap between the categories occurred. A measure of sensitivity (d') was the percentage of total responses in error by use of the above criterion. Mean ratings for the 6-12, 13-19, and 20-26 rnA ranges were also computed, each value being based on 21 judgments. Figure 5 shows the somatosensory evoked-response amplitudes of normal college students divided into pain-tolerant (n = 18) and painintolerant (n = 18) groups on the basis of their being above or below the median on their mean rating for the 20-26 rnA range stimuli. As anticipated, individuals with reducing or relatively low amplitude/ intensity slopes were pain-tolerant, and augmenters were pain-intolerant. These slope differences were confirmed statistically by use of a two-way analysis of variance (p < .05, group by intensity interaction). Individual criterion measures (Cx) for the distinct/unpleasant division were correlated -.48 with the amplitude/intensity slope for PlOD. The d' measure was not significantly correlated. The criterion measure has been previously linked to placebo and suggestion effects (Clark and Goodman, 1974), and our neurophysiological correlates may be related to these phenomena. This linkage between placebo/suggestion effects and PlOD amplitude/intensity slope was further borne out in studies of audioanalgesia (Lavine, Buchsbaum, and Poney, in press). Subjects who heard music with the suggestion that it would reduce pain had lower mean slopes of their somatosensory/intensity function than subjects who were given neither music nor suggestion. This effect was most prominent for PlOD at vertex and over the right postcentral gyrus.
119
SELF-REGULATION OF STIMULUS INTENSITY
~0
> 0
168-248 MS
76-112 MS PAIN INTOLERANT
6
II:
~
::iE
z
4
Z 0
~ :>w
2
Z
0
,-~~...,
.-
~
.- /
,..
PAIN INTOLERANT
PAIN /~ TOLERANT ~
~
~
PAIN
~~~~~---. TOLERANT ~~ ~~
C
« w ::iE
2
9
16
23
2
9
16
23
STIMULUS INTENSITY IN MA
FIGURE 5. Mean somatosensory AER amplitude for four intensities of electric shock for the 76-112 ms and 168-248 ms time bands (generally equivalent to PIOO and P200). Subjects were divided into pain-tolerant and -intolerant groups on the basis of their subjective ratings of shock unpleasantness. Note that the pain-tolerant subjects are relative reducers--have lower rates of increase in AER amplitude with increasing stimulus intensity. Group differences are greatest at the highest intensities.
B. Noise Tolerance As with the pain experiment, we expected reducers to be noisetolerant. We assessed noise tolerance by measuring the rate at which subjects pressed a key to decrease noise while using a teaching machine (Molino, 1974). In a separate session, auditory AERs were collected in a manner similar to that used for somatosensory AERs in the pain tolerance study; i.e., random presentation of four intensities of noise bursts. Individuals showed wide variation in tolerated noise-56-112 dB. Individuals who were relatively tolerant of noise on the key-pressing task were reducers on the auditory AER (Figure 5), whereas noiseintolerant individuals were augmenters. Again, this was statistically confirmed by two-way analysis of variance on the two groups and by correlations between the noise tolerance score (in dB) and the slope of the AER amplitude/intensity function for PlOD. These effects were most
120
MONTE BUCHSBAUM
evident for the highest-octave-frequency band of noise used (8000 Hz), which is of interest in view of the findings of Khechinashvili et al. (1973), who found reducing greatest for high-frequency tones.
VII.
EFFECTS OF AROUSAL, ATTENTION, AND SENSORY OVERLOAD
Since attention and/or arousal is known to affect AER amplitude, could individual differences in augmenting/reducing be explained as differences in these factors? Clearly this would require a differential effect with stimulus intensity; otherwise the AER slope would be unaffected. Difficulties with designing experimental conditions which really distinguish "arousal" and "attention" also make answering this question problematical. Avoiding getting too aroused about these global definitions and attending only to the precise experimental conditions used may be helpful at this point.
A. AER Decrement over Sessions Overall AER amplitude diminution across time or with reduced "arousal" has been widely reported (e.g., Eason, Aiken, White, and Lichtenstein, 1964; Eason, Harter, and White, 1969; Roth, 1973; Hartley, 1970; Landau and Buchsbaum, 1973). Our sample of 128 adult twins had two successive AER sessions in our laboratory about 2 weeks apart. They were initially naIve to EEG recording procedures and received no instructions other than to look forward at the visual stimuli-four intensities of light flashes presented in random order. From the initial to the second session, AER amplitude dropped fairly evenly across all intensities (Figure 6, left). This could reflect diminished "arousal" during the second session in our intimidating laboratory-or diminished attention to the visual stimuli.
B. AER Decrement with Mental Arithmetic Experiments in which attentional factors were manipulated within counterbalanced sessions yielded somewhat different results, however.
2
3
LIGHT INTENSITY
• 4
2.0
4
--0._ _ L'_ - - ' -_ _ _
Tone
><
No tone
FIGURE 6.
Effect of experimental conditions on visual AER amplitudelintensity slopes. (Left) AERs from 128 adult twin subjects tested in two sessions about 2 weeks apart (described in Buchsbaum, 1974). The major effect is a decrease in AER amplitude from the first to the second session, about equal at all intensities. (Middle) AERs from 24 normal subjects (redrawn from Schechter and Buchsbaum, 1973) while making judgments of light intensity and while doing mental arithmetic-the major effect of the mental arithmetic distraction is a decrease in AER at low intensity. (Right) Visual AERs from 40 normal subjects (see text) while exposed to a 95-dB continuous auditory tone and compared with no tone. Major effect is a high-intensity AER-amplitude decrease.
5.0'
2.5
6.0
3.5
3.0
Light intensity judgement
~
7.0
8.0
,.... N ,....
~
Z
~
8
~
til
..,
o."
oz
~
~
~
']1
l"'
til tn
122
MONTE BUCHSBAUM
In another experiment (Schechter and Buchsbaum, 1973) using the same stimuli and recording techniques, subjects either "attended" (counted the appearance of specific-intensity pairs in the sequence) or were "distracted" (performed mental arithmetic while watching the lights). Attending produced AERs of high and equal amplitude across intensities; distraction especially diminished low intensities (Figure 6, middle). Thus what happens across sessions seems to be different from the distraction effect. Consistent with our findings, both Kopell, Wittner, and Warrick (1969) and Chapman and Bragdon (1964) found more enhancement of AER amplitude with selective attention for lowintensity stimuli than for high-intensity stimuli.
C. AER Decrement with Loud Noise In a third experiment (Buchsbaum, Landau, and Morgan, 1972) we presented blocks of 10 flashes of the intensity (intensities 1 and 4 using recording techniques in Buchsbaum and Pfefferbaum, 1971) in a randomized order to 40 normal volunteers. Between the blocks of flashes was a 4-s pause. About 2 s before the onset of half of the blocks of flashes (chosen randomly) we presented a 94-dB 1000-Hz tone which continued throughout the block-a quite noxious sound. For the flashes presented in the blocks without the tone, AER amplitude increased with increasing intensity. For flashes presented during the sounding of the 94-dB tone, however, AER amplitude decreased-with the effect only for the highest-intensity flash (Figure 6, right). A similar effect on the high intensity only was seen in our experiments with audioanalgesia (Lavine, Buchsbaum, and Poncy, in press).
D. Differential Types of AER Decrement Three different effects were seen: (1) an across-session, "arousal" effect which acted on all intensities fairly evenly, affecting slope minimally; (2) an "attention" effect which operated primarily to enhance amplitude for low-intensity stimuli and thus lower AER slope; and (3) a "sensory overload" effect which operated primarily to reduce AER amplitUde at high intensities. Thus reducing appears to be linked to the active phenomena of paying attention and protection from too-intense
SELF-REGULATION OF STIMULUS INTENSITY
123
levels of sensory input. Habitual tendencies to attend to sensory stimuli may be reflected in the AER amplitude/intensity slope as well as habitual tendencies to inhibit sensory input.
VIII.
INDIVIDUAL DIFFERENCES AND INTENSITY JUDGMENTS
A. Psychological Magnitude and Power Functions The responses people make when asked, "How intense is that stimulus?" are variable. Psychological magnitude can be reported with magnitude estimates, rations, or cross-modal matching, and these procedures produce results that roughly fit a power function
where '" is the psychological magnitude and cf> the stimulus intensity (Stevens, 1963, 1971). This relationship holds fairly well when data are averaged across subjects, but the exponent f3 varies widely from subject to subject (e.g., Stevens and Guirao, 1964)-suggesting that the subjective intensity of a stimulus may increase more or less rapidly with increasing stimulus intensity. These individual differences in power-function exponent seem quite reliable across session and modalities (Jones and Marcus, 1963; Rule, 1966; Ekman, Hosman, Lindman, Ljungberg, and Akesson, 1968; Reason, 1972; Wanschura and Dawson, 1974). Luce (1972) cogently reviewed the problems of comparing psychophysical responses and physical units.
B. Power Function Exponents and Augmenting/Reducing Much debate has raged over the source of these differences. Suggested factors include ratio concepts (Stevens, 1971), response bias (Kunnapas, Hallsten, and Soderberg, 1973), regression effects (Wanschura and Dawson, 1974), and memory factors (Engeland and Dawson, 1974). Recently a correlation was found between the amount of augmenting on the Petrie KFA and the slope of the magnitude-estimation function for loudness measured by use of Stevens' procedure (Sales and
124
MONTE BUCHSBAUM
Throop, 1972). This is supported by Cavonius, Hilz, and Chapman (1974), who similarly report KFAlbrightness-estimation slope correlations, although the Kohler and Dinnerstein KFA procedure was used. Reason (1968, 1972) also found correlations between an individual's susceptibility to the spiral aftereffect and the slope of the loudnessestimation function. His explanation of slope differences as individual differences in "reactivity" is analogous to the augmenting/reducing concept. Steeper loudness functions were also found to be correlated with noise-annoyance susceptibility (Moreira and Bryan, 1972), a finding consistent both with the reducer concept and with our own AER study of noise tolerance. Subjects with higher-magnitude judgment slopes also evidenced more anxiety (Stephens, 1970) and more excitability (de Barbenza, Bryan, and Tempest, 1970) on personality tests, further suggesting the importance of complex inhibitory processes.
C. AER and Psychophysical Scaling In view of the vast variety of individual differences in amplitude/ intensity functions for the AER, these individual differences in psychophysical functions are not at all surprising. It would be attractive to believe that a neurophysiological measure might avoid the methodological problems inherent in asking subjects to attach numbers to the loudness of tones. We know, however, that cortical centers where AERs originate are removed several synapses from the sensory transducer and have both inhibitory and excitatory input. Attempts to fit power functions to AER parameters have had only spotty success, especially when AER amplitude is used; the commonly observed pattern of AER saturation or reducing cannot, of course, be fitted with a power function over its whole range (e.g., Butler, Keidel, and Spreng, 1969). Extensive experiments in which ratings of individual stimuli are made and AERs collected simultaneously are needed to relate subjective reports and particular AER components. It seems unlikely, however, that reducing will be paralleled with reports of diminished intensity. AER amplitude for the P100--P200 components seems more linked to the process of making such a report or of suppressing such a process in situations in which it is redundant, unpleasant, or inconvenient to respond.
SELF-REGULATION OF STIMULUS INTENSITY
IX.
125
SENSORY SENSITIVITY AND "STRENGTH OF THE NERVOUS SYSTEM"
A. Response to Low-Intensity Stimuli Buchsbaum and Silverman (1968) hypothesized that reducers were hypersensitive to low-level sensory stimuli and thus required some compensatory process to protect them from sensory inundation at high intensities. Indeed, AER reducers were found to have lower visual thresholds (when the method of limits was used) than AER augmenters (Silverman, Buchsbaum, and Henkin, 1969). And at low to moderate light levels, reducers had larger amplitude AERs than augmenters. This suggests that we might expect reducers to tolerate sensory deprivation better than augmenters-the opposite of the Petrie prediction. The finding that pain-tolerant individuals tolerate sensory isolation better (Peters et al., 1963; Zubek, 1963) is also consistent with the BuchsbaumSilverman stimulus-intensity control model. Petrie viewed reduction as operating evenly on all intensities, whereas we conceptualized it as operating primarily at high intensities in individuals who were unusually responsive at low levels. In this form, the augmenting/reducing dimension and the Pavlov-Teplov "strength of the nervous system" dimension are similar. Pavlov's typology and the subsequent development of this construct are well presented in Gray (1964) and Nebylitsyn and Gray (1972) and are briefly summarized below.
B. "Strength of the Nervous System" and Reducing Soviet psychophysiology since Pavlov has paid close attention to the individual. Subjects are characteristically studied in great detail and attempts are made to explain the variety of responses observed. Pavlov's theories of nervous system types grew out of such an effort. Three continuous dimensions-strength, equilibrium of excitation, and inhibition and mobility-were postulated and animated discussion of dog personality enlivened the Wednesday conferences of Pavlov's group (see Teplov, 1964). Two dogs studied in 1926-1928 were seen as examples of strong and weak nervous systems. Vot-te Chort showed conditioned reflexes which increased with increasing tone intensity; Zhurka
126
MONTE BUCHSBAUM
showed no increase past moderate intensities (reducing, in our terminology) . Pavlov's "law of strength" implied that response should increase evenly with increasing stimulus intensity-and Zhurka broke the law at a lower stimulus intensity than Vot-te Chort. Some animals showed a subsequent decrease in response as intensity was further increased. The level at which the increasing relationship between intensity and response magnitude ceased was termed the "threshold of transmarginal inhibition." Animals with "strong" nervous systems were viewed as having higher thresholds and maintaining linear amplitude/intensity relationships at higher levels than animals with "weak" nervous systems. Transmarginal inhibition was viewed as a special "protective" type of inhibition. Individuals with weak systems were seen as lacking ordinary "internal" inhibition: "the weaker the nervous system the more intense is the excitatory process which is set up by a given physical stimulus" (Gray, 1964). Thus weak nervous systems should have lower sensory thresholds and show decreasing responses with increasing stimulus intensity at the high end of the intensity continuum-like reducers. Strong nervous systems would have high thresholds but show linear amplitude/intensity relationships-like augmenters. Both augmenting/reducing and strength dimensions thus may tend to imply an inverted U function relating response to stimulus intensity with the curves of augmenters/"strongs" shifted toward the high-intensity end.
C. Determination of Strength Strength of the nervous system like augmenting/reducing has been measured in a number of ways, including "absolute thresholds," reaction time or conditioned reflexes to stimuli of varying intensities, and effects of distracting stimuli on psychophysical task performance. Many of these intriguing paradigms have their experimental techniques presented only in sketchy outline and without detailed statistical analysis (Gray, 1964). The one measure which has been used both in Western and in Soviet psychophysiology is the slope of the reaction-time/stimuIus-intensity function. Strong nervous systems (augmenters) should have large increases in reaction rate (decreases in reaction time) with increasing stimulus intensity; weak nervous systems (reducers) should show little change with increasing stimulus intensity. Sales and Throop
SELF-REGULATION OF STIMULUS INTENSITY
127
(1972) found that individuals with low sensory thresholds, indicative of weak nervous systems, indeed had low reaction-rate intensity slopes, and Sales and Throop's subjects were also reducers on the Petrie KFA. Reason (1972) found correlations between the slope of the reaction-rate function and the slope of loudness-magnitude functions. Interestingly, negative reaction/stimulus intensity funct~ons were found by Tizard and Venables (1956) in schizophrenics-who have been reported to be extreme reducers for both the Petrie apparatus (Petrie et al., 1963) and the AER (Landau et al., 1975).
x.
SELF-REGULATION AND SENSORY HOMEOSTASIS
A. Optimum Levels of Stimulation The concept of an optimum level of continuous sensory stimulation to maintain optimal intellectual function or the feeling of well-being has been advanced on a neurological and several psychological levels. This concept was recently reviewed by Zuckerman (1974). Silverman (1967), Ludwig (1971), Sales (1971), and Zuckerman (1974) have suggested that the augmenting/reducing stimulus control mechanism may be one of the methods utilized by individuals to optimize the level of incoming sensory stimulation. Defects in this mechanism could reduce the flexibility of the organism in meeting the challenge of sensory overload (see Miller, 1960; Ludwig, 1972; Lipowski, 1973).
B. Relationships between Pain Tolerance, Sensory Homeostasis, and Distraction The focusing of attention away from the source of painful stimulation appears to be an important source of variation in pain tolerance (e.g., Melzack and Wall, 1970). The ability to reduce, implicated in pain tolerance, is viewed as a dimension of attention by Silverman (1967). The perception-pain-control mechanism of Lykken, Macinoe, and Tellegen (1972) requires paying attention to the warning stimulus. The finding of an interaction between stimulus intensity and attention in three AER studies (see Section VII.B) further links attentional effects to pain and/or homeostatic sensory-control mechanisms. Since the AER technique using stimuli of varying intensities makes it possible to examine correlates of pain response in a subject "attend-
128
MONTE BUCHSBAUM
ing" to other tasks than making pain judgments, it may be especially valuable for research separating these effects.
C. Conclusion As progress is made in the neuroanatomicallocalization of specific AER components, increasingly specific neural models will be possible for the important stimulus-intensity-control mechanisms. The general importance of such processes can be recognized in the diverse sources of such concepts as Petrie's "augmenting/reducing," Pavlov's "strength," Freud's "stimulus barrier," and others. An understanding of such mechanisms may provide important clues about psychiatric dysfunction: AER tools for diagnostic evaluation, drug selection, and identification of genetic vulnerability are all potential applications (Buchsbaum, 1975; Silverman, 1972). ACKNOWLEDGMENTS
The author wishes to thank Dr. Robert Lavine for helpful comments, Sherry Buchsbaum for editorial assistance, and Cathy King and Arlene Ammerman for technical and secretarial aid. REFERENCES ABE, M. Electrical responses of the human brain to acoustic stimuli. The Tohoku Journal of Experimental Medicine, 1953,60, 47-58. ANDERSON, P., ECCLES, J. c., AND SEARS, T. A. Cortically evoked depolarization of primary afferent fibers in the spinal cord. Journal of Neurophysiology, 1964,27,63-77. ARMINGTON, J. C. Adaptational changes in the human electroretinogram and occipital response. Vision Research, 1964a,4, 179--192. ARMINGTON, J. c. Relations between electroretinograms and occipital potentials elicited by flickering stimuli. Documenta Ophthalmogica, 1964b,18, 194-206. BEAGLEY, H. A., AND KNIGHT, J. J. Changes in auditory evoked response with intensity. Journal of Laryngology and Otology, 1967,81, 861--873. BECK, C., AND ROSNER, B. S. Magnitude scales and somatic evoked potentials to percutaneous electrical stimulation. Physiology and Behavior, 1968,3, 947-953. BLfIZ, B., DINNERSTEIN, A. J., AND LOWENTHAL, M. Relationship between pain tolerance and kinesthetic size judgment. Perceptual and Motor Skills, 1966, 22, 463-469. BORGE, G. F. Perceptual modulation and variability in psychiatric patients. Archives of General Psychiatry, 1973,29, 760-763. BROADHURST, A., AND MILLARD, D. W. Augmenters and reducers: A note on a replication failure. Acta Psychologica, 1969,29, 290-296. BUCHSBAUM, M. Average evoked response and stimulus intensity in identical and fraternal twins. Physiological Psychology, 1974,2, 365-370. BUCHSBAUM, M. Average evoked response augmenting/reducing in schizophrenia and
SELF-REGULATION OF STIMULUS INTENSITY
129
affective disorders. In D. X. FREEDMAN (Ed.), The biology of the major psychoses: A comparative analysis. New York: Raven Press, 1975.
BUCHSBAUM, M., GILLIN, J. c., AND PFEFFERBAUM, A. Effect of sleep stage and stimulus intensity on auditory average evoked responses. Psychophysiology, 1975,12(6), 707712. BUCHSBAUM, M., GOODWIN, F., MURPHY, D., AND BORGE, G. AER in affective disorders. American Journal of Psychiatry, 1971,128, 19-25. BUCHSBAUM, M., HENKIN, R. I., AND CHRISTIANSEN, R. L. Age and sex differences in averaged evoked responses in a normal population with observations on patients with gonadal dysgenesis. Electroencephalography and Clinical Neurophysiology, 1974, 37, 137-144. BUCHSBAUM, M., LANDAU, S., AND MORGAN, C. Arousal and attention: Evoked response stimulus intensity modifiers. Presented at the 80th annual convention, American Psychological Association, Honolulu, Hawaii, 1972. BUCHSBAUM, M., LANDAU, S., MuRPHY, D. L., AND GOODWIN, F. Average evoked response in bipolar and unipolar affective disorders: Relationship to sex, age of onset and monoamine oxidase. Biological Psychiatry, 1973,7, 199-212. BUCHSBAUM, M., AND PFEFFERBAUM, A. Individual differences in stimulus intensity response. Psychophysiology, 1971,8, 60()...{j1l. BUCHSBAUM, M., AND SILVERMAN J. Stimulus intensity control and the cortical evoked response. Psychosomatic Medicine, 1968,30, 12-22. BUTLER, R. A., KEIDEL, W. D., AND SPRENG, M. An investigation of the human cortical evoked potential under conditions of monaural and binaural stimulation. Acta Otolaryngologica, 1969, 68, 317-326. CALMA, I. Presynaptic inhibition of the terminals of cutaneous nerve fibers by stimulation of the ventral thalamo-diencephalic region. Journal of Physiology (London), 1966, 185,58-60. CAVONIUS, C. R., Hnz, R., AND CHAPMAN, R. M. A possible basis for individual differences in magnitude-estimation behavior. British Journal of Psychology, 1974,65, 85-9l. CHAPMAN, R. M., AND BRAGDON, H. R. Evoked responses to numerical and non-numerical visual stimuli while problem solving. Nature (London), 1964,203, 1155-1157. CLARK, W. c., AND GOODMAN, J. Effects of suggestion of d' and Cx for pain detection and pain tolerance. Journal of Abnormal Psychology, 1974,83, 364-372. CLYNES, M., KOHN, M., AND LIFSHIlZ, K. Dynamics and spatial behavior of light evoked potentials, their modification under hypnosis, and on-line correlation in relation to rhythmic components. Annals of New York Academy of Science, 1964,112,468-509. CRUIKSHANK, R. M. Human occipital brain potentials as affected by intensity-duration variables of visual stimulation. Journal of Experimental Psychology, 1937,21, 625-64l. DAVIS, H., BOWERS, c., AND HIRSH, S. K. Relations of the human vertex potential to acoustic input: Loudness and masking. Journal of the Acoustical Society of America, 1968,43, 431-438. DAVIS, H., MAST, T., YOSHIE, N., AND ZERLIN, S. The slow response of the human cortex to auditory stimuli: Recovery process. Electroencephalography and Clinical Neurophysiology, 1966,21, 105-113. DAVIS, H., AND ZERLIN, S. Acoustic relations of the human vertex potential. Journal of the Acoustical Society of America, 1966,39, 109-116. DE BARBENZA, C. M., BRYAN, M. E., AND TEMPEST, W. Individual loudness functions. Journal of Sound Vibration, 1970, 11, 399-410. DEBECKER, J., AND DESMEDT, J. E. Les potentiels evoques cerebraux et les potentiels de nerf sensible chez l'homme. Acta Neurologica Belgica, 1964,64, 1212-1248.
130
MONTE BUCHSBAUM
DELL, P., BONVALLET, M., AND HUGELIN, A. Mechanisms of reticular deactivation. In G. E. W. WOLSTENHOLME AND M. O'CONNOR (Eds.), The nature of sleep. London: Churchill, 1961. DEVOE, R. G., RIPPS, H., AND VAUGHAN, H. G. Cortical responses to stimulation of the human fovea. Vision Research, 1968,8, 135-147. DUSTMAN, R. E., AND BECK, E. The visually evoked potential in twins. Electroencephalography and Clinical Neurophysiology, 1965,19, 580-575. EASON, R. G., AIKEN, L. R., WHITE, C. T., AND LICHTENSTEIN, M. Activation and behavior: II. Visually evoked cortical potentials in man as indicants of activation level. Perceptual and Motor Skills, 1964,19, 875-895. EASON, R. G., HARTER, M. R., AND WHITE, C. T. Effects of attention and arousal on visually evoked cortical potentials and reaction time in man. Physiology and Behavior, 1969,4, 283-289. EKMAN, G., HOSMAN, B., LINDMAN, R., L]UNGBERG, L., AND AKESSON, C. A. Interindividual differences in scaling performance. Perceptual and Motor Skills, 1968,26,815-823. ENGELAND, W., AND DAWSON, W. E. Individual differences in power functions for a 1 week intersession interval. Perception and Psychophysics, 1974,15, 349-352. FRANZEN, 0., AND OFFENLOCH, K. Evoked response correlates of psychophysical magnitude estimates for tactile stimulation in man. Experimental Brain Research, 1969,8, 118. GRAY, J. A. (Ed.). Pavlov's typology. New York: Macmillan, 1964. HAGBARTH, K. E., AND KERR, D. I. B. Central influences on spinal afferent conduction. Journal of Neurophysiology, 1954,17, 295. HALL; R. A., RAPPAPORT, M., HOPKINS, H. K., AND GRIFFIN, R. B. Peak identification in visual evoked potentials. Psychophysiology, 1973a, 10, 52--{;0. HALL, R. A., RAPPAPORT, M., HOPKINS, H. K., AND GRIFFIN, R. B. Tobacco and evoked potential. Science, 1973b, 180, 212-214. HALL, R. A., RAPPAPORT, M., HOPKINS, H. K., GRIFFIN, R., AND SILVERMAN, J. Evoked response and behavior in cats. Science, 1970,170, 998--1000. HAMASAKI, D. I., AND WINTERS, R. W. Intensity-response functions of visually deprived LGn neurons of cats. Vision Research, 1973, 13, 925-936. HARTLEY, L. R. The effect of stimulus relevance of the cortical evoked potentials. Journal of Experimental Psychology, 1970,22, 531-546. HENRY, G. B., AND TEAS, D. Averaged evoked response and loudness: Analysis of response estimates. Journal of Speech and Hearing Research, 1968,11, 334-342. HILGARD, E. R., MORGAN, A. H., AND PRYTULAK, S. The psychophysics of the kinesthetic aftereffect in the Petrie block experiment. Perception and Psychophysics, 1968,4, 129132. HILLMAN, P., AND WALL, P. D. Inhibitory and excitatory factors influencing the receptive fields on lamina 5 spinal cord cells. Experimental Brain Research, 1969,9, 284-306. lKUTA, T. Influences of different stimulus intensities on the human somatosensory evoked response (A proposal for the data processing). Folia Psychiatrica et Neurologica Japanica, 1968,22, 121-141. IL'YANOK, V. A. Effect of intensity and depth of pulsation of flickering light on the electrical activity of the human brain. Biofizika, 1961, 6, 72-82. JACOBSON, J. H., KAWASAKI, K., AND HIROSE, T. The human electroretinogram and occipital potential in response to focal illumination of the retina. Investigative Ophthalmology, 1969,8, 545-556. JONES, F. N., AND MARCUS, J. J. The subject effect in judgments of subjective magnitude. Journal of Experimental Psychology, 1963,61, 40-44.
SELF-REGULATION OF STIMULUS INTENSITY
131
KEIDEL, W. D., AND SPRENG, M. Neurophysiological evidence for the Stevens' power function in man. Journal of the Acoustical Society of America, 1965,38, 191-195. KHECHINASHVILl, S. N., KEVANISHVILI, Z., AND KAJAIA, O. A. Amplitude and latency studies of the averaged auditory evoked responses to tones of different intensities. Acta Otolaryngologica, 1973, 76, 395-401. KITAJIMA, H. On the cerebral evoked response in man as a function of the intensity of flicker stimulation. Electroencephalography and Clinical Neurophysiology, 1967, 22, 325-336. KLINGAMAN, R. 1., AND ANCH, A. M. Human auditory evoked response intensity functions from monopolar and bipolar scalp recordings. Psychonomic Science, 1972, 29,17-19. KNISPEL, J. D., AND SmGEL, J. Tegmental stimulation: Aversive effects on behavior and modulation of visual evoked potentials. Brain Research, 1972,37, 317-321. KNISPEL, J. D., AND SmGEL, J. Habituation of aversive reticular stimulation effects on evoked potentials. Brain Research, 1973,56, 340-344. KOELLA, W. P., AND PERRY, A. Cortico-subcortical homeostasis in the cat's brain. Science, 1963, 142, 586--589. KOHLER, W., AND DINNERSTEIN, D. Figural aftereffects in kinesthesis. In A. MICHOTTE (Ed.), Miscellanea Psychologica. Paris: Vrin., 1947. KOLLAR, A. Die Beziehungen zwischen Lautstarke, Latenzzeit und Amplitude in der EEG-Audiometrie. Mschr. Ohrenheilk (Wien), 1971,105, 49-59. KOOI, K. A., AND BAGCHI, B. K. Observations on early components of the visual evoked response and occipital rhythms. Electroencephalography and Clinical Neurophysiology, 1964, 17, 63~3. KOPELL, B., WITTNER, W. AND WARRICK, G. The effects of stimulus differences, light intensity, and selective attention on the amplitude of the average evoked potential in man. Electroencephalography and Clinical Neurophysiology, 1969,26, 619-622. KUNNAPAS, T., HALLSTEN, L., AND SODERBERG, G. Interindividual differences in homomodal and heteromodal scaling. Acta Psychologica, 1973,37, 31--42. LANDAU, S. G., AND BUCHSBAUM, M. Average evoked response and muscle tension. Physiological Psychology, 1973,1, 56--60. LANDAU, S. G., BUCHSBAUM, M., CARPENTER, W., STRAUSS, J., AND SACKS, M. Schizophrenia and stimulus intensity control. Archives of General Psychiatry, 1975,32(8), 12391245. LAVINE, R. A. Phase-locking in response of single neurons in cochlear nuclear complex of the cat to low-frequency tonal stimuli. Journal of Neurophysiology, 1971,34,467--483. LAVINE, R. A., BUCHSBAUM, M. AND PONCY, M. Auditory analgesia: Somatosensory evoked response and subjective pain rating assessment. Psychophysiology, in press. LEWIS, E. G., DUSTMAN, R. E., AND BECK, E. C. Evoked response similarity in monozygotic, dizygotic and unrelated individuals: A comparative study. Electroencephalography and Clinical Neurophysiology, 1972, 32, 309-316. LmBMAN, J., AND GRAHAM, J. T. Frequency and intensity effects of bone conduction signals on averaged evoked auditory potentials. Journal of Auditory Research, 1967,7, 157-161. LIPOWSKI, A. J. Affluence, information inputs and health. Social Science and Medicine, 1973, 7, 517-529. LuCE, R. D. What sort of measurement is psychophysical measurement. American Psychologist, 1972,27, 96--106. LUDWIG, A. M. Self-regulation of the sensory environment. Archives of General Psychiatry, 1971,25,413--418.
132
MONTE BUCHSBAUM
LUDWIG, A. M. "Psychedelic" effects produced by sensory overload. American Journal of
Psychiatry, 1972,10, 114--117. LUNDERBERG, A., AND OSCARSSON, O. Three ascending spinal pathways in the dorsal part of the lateral funiculus. Acta Physiologia Scandinavia, 1961,51, 1-16. LYKKEN, D. T., MACINOE, 1., AND TELLEGEN, A. Preception: Autonomic response to shock as a function of predictability in time and locus. Psychophysiology, 1972,9, 318-333. LYKKEN, D. T., TELLEGEN, A., AND THORKELSON, K. Genetic determination of EEG frequency spectra. Biological Psychology, 1974,1, 245-259. MADELL, J. R., AND GOLDSTEIN, R. Relation between loudness and the amplitude of the early components of the averaged electroencephalic response. Journal of Speech and Hearing Research, 1972,15, 134--141. MARCO, J. Cortical audiometry. Acta Otolaryngologica, 1972,73, 197-202. MARK, R. F., AND STEINER, J. Cortical projection of impulses in myelinated cutaneous afferent nerve fibers of the cat. Journal of Physiology, 1958,142, 544--562. MEIZACK, R. The puzzle of pain. New York: Basic Books, 1973. MEIZACK, R., AND WALL, P. D. Psychophysiology of pain. International Anesthesiology Clinics, 1970,8, 3-34. MILLER, J. C. Information input overload and psychopathology. American Journal of Psychiatry, 1960,116,695-704. MOLINO, J. Equal aversion levels for pure tones and 1/3 octave band of noise. Journal of the Acoustical Society of America, 1974,55, 1285-1289. MONTAGU, J. D. The relationship between the intensity of repetitive photic stimulation and the cerebral response. electroencephalography and Clinical Neurophysiology, 1967, 23, 152-161. MOORE, E. J., AND ROSE, D. E. Variability of latency and amplitude of acoustically evoked responses to pure tones of moderate to high intensity. International Audiology, 1969, 8, 172-181. MOREIRA, N. M., AND BRYAN, M. E. Noise annoyance susceptibility. Journal of Sound and Vibration, 1972,21, 449-462. MORGAN, A. H., LEZARD, F., PRYTULAK, S., AND HILGARD, E. R. Augmenters, reducers, and their reaction to cold pressor pain in waking and suggested hypnotic analgesia. Journal of Personality and Social Psychology, 1970,16, 5-11. MOUNTCASTLE, V. B. (Ed.), Medical physiology (13 ed.), (Vol. 1). Chapter 10: Neural Mechanisms in Somesthesia. St. Louis: C. V. Mosby, 1974. MUSHIN, J., AND LEVY, R. Averaged evoked response in patients with psychogenic pain. Psychological Medicine, 1974,4, 19-27. NEBYLITSYN, V. D., AND GRAY, J. A. Biological bases of individual behavior. New York: Academic Press, 1972. OSBORNE, R. T. Heritability estimates for the visual evoked response. Life Science, 1970,9, 481-490. PETERS, J., BENJAMIN, F. B., HELVEY, W. M., AND ALBRIGHT, C. A. Study of sensory deprivation, pain and personality relationships for space travel. Aerospace Medicine, 1963,34, 830-837. PETRIE, A. Some psychological aspects of pain and the relief of suggering. Annals of the New York Academy of Science, 1960,86, 13-27. PETRIE, A. Individuality in pain and suffering. Chicago: University of Chicago Press, 1967. PETRIE, A. Reduction or augmentation? Why we need two "planks" before deciding. Perceptual and Motor Skills, 1974,39, 460. PETRIE, A., COLLINS, W., AND SOLOMAN, P. Pain sensitivity, sensory deprivation and susceptibility to satiation. Science, 1958,128, 1431-1433.
SELF-REGULATION OF STIMULUS INTENSITY
133
PETRIE, A., HOLLAND, T., AND WOLK, I. Sensory stimulation causing subdued experience: Audio-analgesia and perceptual augmentation and reduction. Journal of Nervous and Mental Disease, 1963, 137, 312-321. PFEFFERBAUM, A., AND BUCHSBAUM, M. Handedness and cortical hemisphere effects in sine wave stimulated evoked responses. Neuropsychologia, 1971,9, 237-240. PICTON, T. W., GOODMAN, W. S., AND BRYCE, D. P. Amplitude of evoked responses to tones of high intensity. Acta Otolaryngologica, 1970,70, 77-82. PLATT, D., HOlZMAN, P. S., AND LARSON, D. Individual consistencies in kinesthetic figural aftereffects. Perceptual and Motor Skills, 1971,32, 787-795. RAPIN, I., SCHIMMEL, H., TOURK, L. M., KRASNEGOR, N. A., AND POLLAK. C. Evoked responses to clicks and tones of varying intensity in waking adults. Electroencephalography and Clinical Neurophysiology, 1966,21, 335-344. REASON, J. T. Relations between motion sickness susceptibility, the spiral aftereffect and loudness estimation. British Journal of Psychology, 1968,59,385-393. REASON, J. T. Some correlates of the loudness function. Journal of Sound and Vibration,
1972,20, 305-309.
REDMOND, D. E., BORGE, G. F., BUCHSBAUM, M., AND MAAS, J. W. Evoked potential studies of brain catecholamine alterations in monkeys. Journal of Psychiatric Research, 1975,12,96-116. REGAN, D., AND BEVERLEY, K. I. Relation between the magnitude of flicker sensation and evoked potential amplitude in man. Perception, 1973,2, 61-65. RICHEY, E. T., KOOI, K. A., AND WAGGONER, R. W. Visually evoked responses in migraine. Electroencephalography and Clinical Neurophysiology, 1966,21, 23-27. RIETVELD, W. J. The occipitocortical response to light flashes in man. Acta Physiologica et Pharmacologia Neerlandica, 1963,12, 373-407. RIETVELD, W. J., AND TORDOIR, W. E. M. The influence of flash intensity upon the visual evoked response in the human cortex. Acta Physiologia et Pharmacologia Neerlandica, 1965, 13, 160-170. ROSE, D. E., AND RUHM, H. B. Some characteristics of the peak latency and amplitude of the acoustically evoked response. Journal of Speech and Hearing Research, 1966,9,412422. ROSNER, B. S., AND GOFF, W. R. Electrical responses of the nervous system and subjective scales of stimulus intensity. In W. D. NEFF (Ed.), Contributions to sensory physiology (Vol. 2). New York: Academic Press, 1967. ROTH, W. T. Auditory evoked responses to unpredictable stimuli. Psychophysiology, 1973,
10, 125-138. ROTHMAN, H. H. Effects of high frequencies and intersubject variability on the auditoryevoked cortical response. The Journal of the Acoustic Society of America, 1970,47,569573. RULE, S. J. Subject differences in exponents of psychophysical power functions. Perceptual and Motor Skills, 1966,23, 1125-1126. RYAN, E. c., AND FOSTER R. Athletic participation and perceptual augmentation and reduction. Journal of Personality and Social Psychology, 1967,6,472-476. SALES, S. M. Need for stimulation as a factor in social behavior. Journal of Personality and Social Psychology, 1971, 19, 124--134. SALES, S. M., AND THROOP, W. F. Relationship between kinesthetic aftereffects and "strength of the nervous system." Psychophysiology, 1972,9, 492-497. SAMSON, H. H., AND YOUNG, M. L. The relation of flash intensity and background illumination to the photic evoked potential in the pigeon's optic tectum. Vision Research, 1973,13, 253-261.
134
MONTE BUCHSBAUM
SCHECHrER, G., AND BUCHSBAUM, M. The effects of attention, stimulus intensity and individual differences on the average evoked response. Psychophysiology, 1973,10. 392-400. SCHOOLER, c., AND SILVERMAN, J. Differences between correlates of perceptual styles and Petrie task performance in chronic and acute schizophrenics. Perceptual and Motor Skills, 1971,32, 595-601. SCHWEI'IZER, P. K., AND TEPAS, D. I. Intensity effects of the auditory evoked brain response to stimulus onset and cessation. Perception and Psychophysics, 1974, 16, 396-400. SHAGASS, C. Evoked potentials in psychiatry. New York: Plenum Press, 1972. SHAGASS, c., OVERTON, D. A., BARTOLUCCI, G., AND STRAUMANIS, J. J. Effect of attention modification by television viewing on somatosensory evoked responses and recovery function. Journal of Nervous and Mental Disease, 1971, 152, 53-62. SHAGASS, c., AND SCHWARTZ, M. Evoked potential studies in psychiatric patients. Annals of the New York Academy of Science, 1964,112, 526-541. SHAGASS, c., SCHWARTZ, M., AND KRISHNAMOORTI, S. Some psychologic correlates of cerebral response evoked by light flash. Journal of Psychosomatic Research, 1965, 9, 223-231. SHIPLEY, T., JONES, R. W., AND FRY, A. Intensity and the evoked occipitogram in man. Vision Research, 1966, 6, 657-667. SILVERMAN, J. Perceptual control of stimulus intensity in paranoid and nonparanoid schizophrenia. Journal of Nervous and Mental Disease, 1964,6, 545-549. SILVERMAN, J. Variations in cognitive control and psychophysiological defense in the schizophrenias. Psychosomatic Medicine, 1967,29, 225-251. SILVERMAN, J. Stimulus intensity modulation and psychological dis-ease. Psychopharmacologia, 1972,24, 42-80. SILVERMAN, J., BUCHSBAUM, M., AND HENKIN, R. Stimulus sensitivity and stimulus intensity control. Perceptual and Motor Skills, 1969,28, 71-78. SOSKIS, D. A., AND SHAGASS, C. Evoked potential tests of augmenting-reducing. Psychophysiology, 1974,11, 175-190. SPILKER, B., AND CALLAWAY, E. "Augmenting" and "reducing" in averaged visual responses to sine wave light. Psychophysiology, 1969, 6, 49-57. STARK, 1. H., AND NORTON, J. c. The relative reliability of average evoked response parameters. Psychophysiology, 1974,11, 600-602. STEPHENS, S. D. G. Personality and the slope of loudness function. Journal of Experimental Psychology, 1970,22, 9-13. STEVENS, S. S. The psychophysics of sensory function. In W. A. ROSENBLITH (Ed.), Sensory communication. Boston: MIT Press, 1963. STEVENS, S. S. Issues in psychophysical measurement. Psychological Review, 1971,78,426450. STEVENS, J. C. AND GUIRAO, M. Individual loudness functions. J. Acoustical Soc. Amer., 1964,36 (11), 221!.>-2213. SUZUKI, T., AND TAGUCHI, K. Cerebral evoked response to auditory stimuli in waking man. Annals of Otology, Rhinology and Laryngology, 1965,74, 128-139. SWEENEY, D. R. Pain reactivity and kinesthetic aftereffect. Perceptual and Motor Skills, 1966,22, 763-769. TEPAS, D. I., AND ARMINGTON, J. c. Properties of evoked visual potentials. Vision Research, 1962,2, 449-461. TEPAS, D. I., ARMINGTON, J. c., AND KROPFL, W. J. Evoked potentials in the human visual system. In E. E. BERNARD AND M. R. KARE (Eds.), Biological prototypes and synthetic systems. New York: Plenum Press, 1962.
SELF-REGULATION OF STIMULUS INTENSITY
135
TEPAS, D. I., BOXERMAN, 1. A., AND ANCH, A. M. Auditory evoked brain response: . Intensity functions from bipolar human scalp recordings. Perception and Psychophysics, 1972,11, 217-22l. TEPLOV, B. M. Problems in the study of general types of higher nervous activity in man and animals. In J. A. GRAY (Ed.), Pavlov's typology. New York: Macmillan, 1964. TIZARD; J., AND VENABLES, P. H. Reaction time responses by schizophrenics, mental defectives and normal adults. American Journal of Psychiatry, 1956, 112, 8OH07. TURSKY, B., AND WATSON, P. D. Controlled physical and subjective intensities of electric shock. Psychophysiology, 1964,1, 151-162. UTIAL, W. R., AND COOK, 1. Systematics of the evoked somatosensory cortical potential: A psychophysical-electrophysiological comparison. Annals of the New York Academy of Science, 1964, 112, 60-80. VAN DER TWEEL, 1. H., AND VERDUYN LUNEL, H. F. E. Human visual responses to sinusoidally modulated light. Electroencephalography and Clinical Neurophysiology, 1965, 18, 587-598. VAUGHAN, H. G. The perceptual and physiologic significance of visual evoked responses recorded from the scalp in man. In Clinical Electroretinography Suppl. Vision Research, 1966, 20~223. VAUGHAN, H. G., AND HULL, R. C. Functional relation between stimulus intensity and photically evoked cerebral response in man. Nature, 1965,206, 720-722. WALL, P. D. The laminar organization of the dorsal hom and effects of descending impulses. Journal of Physiology (London), 1967,188, 403-423. WANSCHURA, R. G., AND DAWSON, W. E. Regression effect and individual power functions over sessions. Journal of Experimental Psychology, 1974,102, 80«'Hl12. WICKE, J. D., DONCHIN, E., AND LINDSLEY, D. B. Visual evoked potentials as a function of flash luminance and duration. Science, 1964, 146, 8~. WOOTEN, B. R. Photopic and scotopic contributions to the human visually evoked cortical potential. Vision Research, 1972, 12, 1647-1660. ZUBEK, J. P. Pain sensitivity as a measure of perceptual deprivation tolerance. Perceptual and Motor Skills, 1963, 17, 641-642. ZUCKERMAN, M. The sensation seeking motive. Progress in Experimental Personality Research, 1974,7, 80-148. ZUCKERMAN, M., MURTAUGH, T., AND SIEGEL, J. Sensation seeking and cortical augmenting-reducing. Psychophysiology, 1974,11, 535-542.
4
Neodissociation Theory of Multiple Cognitive Control Systems ERNEST
R.
HILGARD
Man does more than one thing at a time-all of the time-but the representation of these actions in consciousness is never complete. On occasion he becomes conscious of much that happens within his body and of much that is happening currently in the external world, as well as of remembered or imagined events. His awareness can shift from one to another of these happenings, and there is some question about how much he can comprehend at once. All of this is familiar, but problems arise when a scientific account is attempted. The total problem is usually seen as too large, so that the investigator approaches it topically, that is, by selecting some manageable features that can be appropriately labeled and then studying them. The recent upsurge of interest in locus of control represents an approach taken largely by those committed to the study of personality and social psychology; selective attention in its various manifestations becomes the focus of experimental psychologists, extending into various studies of task interferences. Those with psychophysiological interest become concerned with hemispheric laterality or with state-dependent learning. Psychopathologists are again studying multiple personalities. I am proposing that it is time to take a look at the whole problem and to see if some unifying framework may not bring together the various lines of evidence into a general theory of consciousness and of cognitive control systems. The study of hypnosis moves headlong into this set of problems because hypnotic procedures change the balance between voluntary R. California.
ERNEST
HILGARD
Department of Psychology, Stanford University, Stanford,
137
138
ERNEST
R.
HILGARD
and involuntary processes, interfere with normal information-processing and memory retrieval, and produce a variety of distortions of awareness and splits in consciousness. The phenomena are baffling and are in many ways troublesome to investigate properly. If, however, hypnotic investigators are successful in what they are doing, they should be able to tell us not only about hypnosis but about human functioning more generally and so contribute to the understanding of normal consciousness and the control systems affecting it.
1.
PIERRE JANET'S THEORY OF DISSOCIATION
Pierre Janet (1859-1947) was one of the first to propose a general theory of the organization of consciousness derived in large measure from observations made with the help of hypnosis. He was the first to introduce the term subconscious to refer to a level of cognitive functioning out of awareness that could on occasion be brought to consciousness. The term unconscious was already familiar through the writings of von Hartmann, but Janet preferred to substitute his term to avoid the romantic excesses that were already centered in the term unconscious. Later Morton Prince introduced the term coconscious for essentially the same reason. Janet's explanation for the constellation of ideas not available to the main consciousness was that they were dissociated. The term derived, of course, from the prevalent doctrine of association. If memories are brought to consciousness by way of the association of ideas, then those not available to association must be "dis-associated." Janet offered a simple diagram to show how a system of ideas, coherent in itself, might be separated off from the primary personal consciousness (Figure 1). The diagram was used to illustrate the case of Irene, who in her somnambulistic state repeatedly rehearsed the death of her mother, which she had experienced under traumatic circumstances. In her normal condition she not only forgot what she had dramatized in her somnambulism, but she forgot the events themselves. "I know very well my mother must be dead," she is reported to have said, "since I have been told so several times, since I see her no more, and since I am in mourning; but I really feel astonished at it. When did she die?" The case of Irene represented what Janet called monoideic somnambulism. He had :::ome to his point of view earlier through the study
139
NEODISSOCIATION THEORY OF MULTIPLE COGNITIVE CONTROL SYSTEMS
Janet's diagrammatic representation of the dissociation in the case of Irene. The polygon represents the ideas related to the death of her mother, separated from the main personality (P). S stands for the sight of the face of the dead mother, V for the sound of her voice, M for the feeling of movements in carrying her body, and so on. This isolated, yet integrated, set of ideas and memories is responsible for Irene's strange behavior in the somnambulistic state (Janet, 1907, page 41). FIGURE 1.
s
V ~--~
~~------+-~
M
of alternating personalities manifested by the subjects he studied. In the thesis on Psychological Automatism that he presented to the Faculty of Letters, before he went on to the MD a few years later, he distinguished between the total automatism of the catatonic patient and the partial automatisms of the hysterical patients whom he had studied (Janet, 1889). Two of these, Leonie and Lucie, may serve to illustrate the kinds of phenomena with which he dealt. Leonie was his first-reported and most thoroughly experimentedupon case. She came to Janet's attention early in his career, when he was more interested in experimenting upon her than in attempting to resolve her problems. She had apparently had natural attacks of somnambulism since the age of 3, and had been repeatedly hypnotized by all sorts of people from the age of 16 on; she was 45 when Janet was still studying her. She had spent her childhood as a peasant in country surroundings, but the rest of her life had been spent in "drawing rooms and doctors' offices," as Janet put it. In her normal state she was serious, timid, mild, and a little sad; when hypnotized she became vivacious and noisy, with a tendency to irony and jesting with respect to the strangers who had come to witness her hypnotic behavior. Janet
140
ERNEST
R.
HILGARD
at first performed some dramatic but poorly controlled studies of hypnotic influence at a distance, but he repudiated any parapsychological influences and attempted to give a purely naturalistic account of what he observed. Leonie eventually turned up with three personalities uncovered with the aid of hypnosis, on occasion called Leonie I, II, and III, sometimes given the names by which the first two referred to themselves--Leonie and Leontine-and, for the third, Leonore, a name given by the magnetizer who had first discovered the third personality. Leontine appeared when hypnotized by Janet, as she had for other hypnotists before him. Later on, when Leontine was herself hypnotized, a third personality (Leonore) made her appearance. It was only after studying Leonie for some time that Janet found out that she had been treated hypnotically years before by some of the "magnetizers" of that period and that the "new" personality had already been elicited and christened 20 years earlier. In a much-studied and much-hypnotized case of this kind doubts always arise as to the role of the hypnotist in consolidating personalities out of amnesic material: Janet was not totally unaware of the problem. He recognized the role of defining the secondary personality by naming it: "Once baptized, the unconscious personality is clearer and more definite; it shows its psychological traits more clearly" Oanet, 1889, p. 318). The three personalities of Leonie showed signs of their origins. The first was appropriate to her upbringing as a simple country girl and housewife, now placed in a sophisticated urban setting. She had had her first child while hypnotized and spontaneously fell into the hypnotic state when her other children were born; it is not very surprising that the hypnotized personality (Leonie II) claimed the children as her own, while assigning the husband to Leonie I, who accepted both the husband and the children. Leonore (Leonie III), doubtless a product of the hypnotic manipulations, might have been made use of to reintegrate the personality, for she was aware of the others, although considering Leonie I to be stupid and Leonie II to be disturbed. The amnesic barriers that persisted made Leonie I know only herself; Leonie II knew Leonie I as well as herself; and Leonie III knew them all. At the age of 19 Lucie came to Janet's attention because she was seized with fits of terror without motivation: "I am afraid and I don't know why." Through the use of automatic writing, a favorite 19thcentury method of getting at hidden aspects of personality (deriving in
NEODISSOCIATION THEORY OF MULTIPLE COGNITIVE CONTROL SYSTEMS
141
the first place from mediumistic seances), Janet uncovered the roots of her terror. She had been severely frightened at the age of 7 by two men hiding behind a curtain and trying to playa joke on her. A second personality, Adrienne, relived this experience during her fits of terror. It should be noted that the dissociation was not complete; the emotion broke through to the normal consciousness, although the events that were the occasion of the emotion lay concealed behind a veil of amnesia. Janet relieved her of her symptoms through a combination of hypnosis and automatic writing and the second personality disappeared. Janet's theory was supported for a time in America by Boris Sidis and Morton Prince, but it gradually died out, in part through being superseded by psychoanalysis, in part by failures in experiments to show complete dissociation between events conscious and subconscious. The epitaph was written by White and Shevach (1942), who felt that the concept was no longer useful. Little has been heard of it since, except for occasional purely descriptive references to some of the disjunctions between events and their conscious representation.
II.
WHY A NEODISSOCIATION THEORY?
In proposing a kind of revival of dissociation theory, I am calling attention to the phenomena that gave rise to dissociation theory in the first place. Somnambulisms, amnesias, fugues, and multiple personalities are phenomena of nature, occurring spontaneously and independent of hypnosis. Even if they are rare, they may tell us much about human personality and consciousness. A simple revival of the dissociation concept would, however, be misleading. Such a revival would carry with it Janet's insistence that these and related phenomena are found only among those with hysterical personalities, and a revival would be a defense of many observations of doubtful value. Furthermore, it would reinstate a doctrine of sharp separation between dissociated activities. If the concept of neodissociation is used, many of the controversial questions can be restated in contemporary form, so that while the historical roots of the concept are given full credit, there is no need for adherence to the views of those who propounded the original doctrine of dissociation. Even the associationistic flavor does not have
142
ERNEST
R.
HILGARD
to be preserved, if modem information-processing language proves to be more useful. In summary, neodissociation theory proposes a fresh start in the attack on the kinds of problems that gave rise to the classical theory. The aim is to produce a modem comprehensive theory to account for the multiplicity of processes that control overt behavior and conscious processes, with full recognition that something like parallel processing may occur and that all processed information is not available at anyone time to consciousness. This is a large order, and the thoughts presented in this chapter are therefore incomplete. It is hoped, however, that they may call attention to an enterprise in which others may wish to share.
III.
THE HYPNOTIC MODEL
One of the earliest discovered aspects of hypnosis was posthypnotic amnesia for the events taking place within hypnosis. That this was a true concealment, and not ordinary forgetting, was illustrated by the fact that the memories lost in hypnotic amnesia could be recovered. Amnesia was so consistently a part of hypnotic responsiveness that it was the original defining characteristic of hypnosis, reflected in the term somnambulist assigned to the highly hypnotizable person. While the term somnambulist is still used to describe the hypnotic virtuoso, it has lost its original meaning, which derived from the fact that the sleepwalker is usually amnesic for his sleepwalking activities. Amnesia provides a paradigm for dissociation because it demonstrates that fully registered ideas, ordinarily available to memory, can become temporarily "dissociated" and unavailable. Although amnesia sounds like Freudian repression, the repressed ideas in posthypnotic amnesia need not have any emotional significance for the subject; they are not necessarily "primary-process" thoughts any more than "secondary-process" ones. This difference from repression is worth noting, because psychoanalytic thinking is somewhat more familiar than the concept of dissociation. The distinction is diagramed in simplified form in Figure 2. It may be noted that there is no necessary quarrel between the amnesic dissociation and a dissociation caused by repression, only that they are in some respects different. Under other circumstances they
NEODISSOCIATION THEORY OF MULTIPLE COGNITIVE CONTROL SYSTEMS
f z o Cs ~ and IJ.J
!t
Pcs
Available to Consciousness
Amnesic Barrier
Not Available to Consciousness
IJ.J
a:::
u
~======~======~
~
143
____ Repression Barrier
~
~ ~ a..
Ucs
Available to Consciousness Only Indirectly
DISSOCIATION 2. A distinction between the divisions of consciousness in dissociation and in psychoanalytic theory. In psychoanalytic theory (simplified for this purpose) the available memories lie in the conscious (Cs) and the preconscious (Pes), while the hidden ones are concealed under a repression barrier and lie in the unconscious (Ues). The unavailable ideas are largely those bound up with affect and impulse, and they enter consciousness only indirectly. In a dissociation through amnesia the split is among the usually available memories, and the unavailable memories need have no special affective or impulsive significance. FIGURE
may overlap, as when, in a partial amnesia, it is shown that the targets of the amnesia may very well be those predicted by repression theory (Clemes, 1964). Amnesia lies at the heart of fugues and alternating personalities. If it were not for the amnesia, there would be less problem of integration. We all play different roles under different circumstances, but because our memories are intact we do not thereby suffer a sense of alternating selves. There is a special problem not raised by ordinary posthypnotic amnesia, a problem common also in repression theory. The problem is this: Can an experience be registered, so that information regarding it is being processed, and at the same time that experience be deflected from consciousness before it has ever been conscious? This deflection in advance of awareness is something other than an afterexpulsion and, if it occurs, raises additional theoretical problems.
144
ERNEST
R.
HILGARD
This deflection into amnesia before the event is conscious appears to take place in experiments on hypnotic analgesia. Subjects who are analgesic report no pain or suffering whatever; it is not as though they have forgotten a pain after experiencing it (as is found in some cases of barbiturate sedation), but the pain is simply not felt at all. It can be shown, however, by an automatic talking technique (related to the automatic writing that Janet used) that the subject did indeed feel the pain and processed it at nearly its full intensity (Hilgard, 1973; Knox, Morgan, and Hilgard, 1974; Hilgard, Morgan, and Macdonald, 1975). These results indicate that it is possible to have an experience registered, processed, and stored in memory in a recoverable form, even though it has never been consciously felt or reported. Of course there may be some arguments about the possibility of a progressive amnesia or intermittent amnesia, and these will have to be faced when explanations are considered, but the facts are as stated: recoverable amnesia is evidenced for an experience of which the subject has never been aware. Although this kind of dissociation is dramatic in experiments on pain, it is by no means limited to pain. It has been known informally for a long time that part of the hypnotized person processed information more accurately than the hypnotized consciousness knew. William James devoted several pages to an account of gaps in consciousness, with evidence that the mind is active even when the person afterwards ignores the fact (James, 1890, Vol. 1, pp. 201-213). He pointed to experiments of Janet and of Binet, showing that hysterics with anesthesia could be shown under another condition to be sensitive. An anesthetic person, ordinarily unable to distinguish between the two points of a compass as in experiments on the two-point threshold, is able to discriminate as accurately as anyone else by way of automatic writing. James recorded his own experiment: "In a perfectly healthy young man who can write with a planchette, I lately found the hand to be entirely anesthetic during the writing act; I could prick it severely without the Subject knowing the fact. The writing on the planchette, however, accused me in strong terms of hurting the hand" (p. 208). We have been able to demonstrate in our own laboratory that hypnotic blindness and hypnotic deafness, as well as positive hallucinations, can all be penetrated by automatic talking. That is, the "hidden observer," as we metaphorically describe the part out of normal awareness, tells us the actual physical situation, the numbers that were not seen, the
NEODISSOCIATION THEORY OF MULTIPLE COGNITIVE CONTROL SYSTEMS
145
sentences that were not heard, the nothingness that was seen as a playful dog. The hypnotic experience does not stop with this double consciousness, one of distorted reality as perceived within hypnosis, the other of undistorted reality as described in automatic talking. There is a third split that can best be described as an observing consciousness, so far as its conscious role is concerned. By inference it is also an active monitoring and controlling agent. In its phenomenal role the observing consciousness is present during hypnosis, knows whether the person feels hypnotized or not, is sometimes surprised by how effective the hypnosis is, is disappointed when the hypnotic suggestions do not work. Subjects often describe the situation as if the hypnotized self were taking the center of the stage, performing the acts called for by the hypnotist, while the observing part is in the wings, watching what is going on. This observing part does not have access directly to the hidden observer and hence is separated from it by an amnesic block. At the same time it may be inferred that this observing, monitoring, and controlling part is not powerless; when the hypnotist desires access to the hidden observer, there must be some central controlling part that allows this. In self-hypnosis, it must be the observing-monitoring part that gives the self-suggestions that the hypnotized part then carries out. There are puzzling metaphors in this kind of commentary on what goes on, but the puzzles cannot be resolved satisfactorily by shortcuts such as labeling everything as role behavior or by assuming that everything is confabulated. The hypnotic model, as I have described it, is not an explanatory one. It serves to dramatize the phenomenal or inferred splits that a theory has to account for in more general terms.
IV.
NEODISSOCIATION MODEL OF MULTIPLE COGNITIVE CONTROL STRUCTURES
With the aid of certain basic assumptions we can construct a generalized model of cognitive control systems, holding the evidence from hypnosis in view but moving beyond that evidence. The first assumption is that there are many subordinate cognitive structures, each with a degree of unity, persistence, and autonomy of functioning.
146
ERNEST
R.
HILGARD
The concept of unity of the total consciousness is an attractive one, but it does not hold up under examination; there are too many shifts, as, for example, between the waking consciousness and the dream consciousness. There are also degrees of automatization achieved with practice, so that well-learned habits-such as playing a musical instrument, driving an automobile, or saying the alphabet-can go on with a minimum of conscious control once the activity is begun. It is important to note, however, that all these available activities do not go on all of the time or all at once; hence there must be some method of inhibiting them, on the one hand, and of facilitating them, on the other. The second assumption is that there is some sort of hierarchical control that manages the competition between these structures; if not, there would be a veritable deluge of thoughts and actions going on all of the time. This hierarchy must have a controlling or regulatory mechanism, however, or the hierarchies could not shift appropriately to the demands being made upon the person. Hence in addition to multiple structures in some sort of hierarchy, it is convenient to make a third assumption of some sort of central monitoring and controlling structure. In my first attempt to call attention to these essential features I presented the diagram of Figure 3. This highly generalized diagram was designed to convey the idea of multiple structures (of which only three are shown), arranged in hierarchical order, as suggested by their positions in the chart, each with appropriate inputs and outputs and with multiple feedbacks among them. At the top is an executive ego or central control structure that has the planning, monitoring, and managing functions that are required for appropriate thought and action. So that the appearance of favoring excessive freedom on the part of this executive could be avoided, another box was added to indicate that the executive system has constraints upon it. The assumption was made that suggestions from the hypnotist might influence the executive function and perhaps change the hierarchical arrangements within the subsystems. Before criticizing and amending this diagram, I wish to call attention to the many other psychologists who have found it desirable to consider subordinate cognitive structures of one kind or another. Without being exhaustive, I shall mention six kinds of concepts closely related to the cognitive control structures of Fig. 3. My purpose is to show that this part of my analysis is closely allied to other developments within general psychology.
NEODISSOCIATION THEORY OF MULTIPLE COGNITIVE CONTROL SYSTEMS
Constro i nts on Ego Autonomy (Including Hypnosis)
147
Executive Ego; Central Control Structure
FIGURE 3. Subordinate cognitive control systems, in a hierarchical order subject to change, under a dominant executive ego or central control structure. (From Hilgard, 1973, p. 405.)
1. The term cognitive structure was made familiar by Kurt Lewin (1935) and Edward Tolman (1932). Such a structure has much in common with my substructures; it may be pervasive, but a single structure is never all-embracing, and there are problems of communication between them. 2. One of Clark Hull's central concepts was that of habit-family hierarchy, in which a number of habits (each of which may be considered a small substructure) permit the organism to achieve its goals in a given situation, and these small substructures are organized in a preferential or hierarchical system, so that if one is blocked the next is activated (Hull, 1934). Berlyne (1965) later translated Piaget's theory into these terms. 3. Donald Hebb's conception of cell assemblies serves to provide a physiological substratum for cognitive structures (Hebb, 1949), and in a talk before the Canadian Psychological Association in 1974 he noted a possible coordination between his proposals and the existence of the "hidden observer" as I have described it (Hebb, 1975). 4. The images and plans of Miller, Galanter, and Pribram (1960) are
148
ERNEST
R.
HILGARD
their selection of levels of analysis that would provide for control of thought and action; these, too, may be considered to be substructures. 5. The subordinate ego-structures of Gill and Brenman (1959), with a dominant ego, represent an interpretation similar to mine, at least within hypnosis. In a larger setting, the various ego-apparatuses of Hartmann (1958), especially those in his conflict-free ego sphere, are readily assimilated. 6. Sarbin's roles can also be considered to be cognitive substructures (e.g., Sarbin and Coe, 1972). The important problem of the present theory is, however, to provide for the hierarchical management of roles and for their inhibitions and facilitations. The nature of the central control structure is an important but troublesome problem. The extreme possibilities are that there is a very powerful central control (replacing the older concept of a strong will), or that there is none at all, that the hierarchy is determined by a competition among the parts for the control of the final common path. For many years psychologists evaded the problem of a planning self, so that, in essence, the second of these alternatives was accepted. To the extent that man is controlled by external stimuli, and the habits conditioned to them, what he does will be the compromise behavior that adapts to the totality of these forces upon him. It was not only the stimulus-response psychologists, however, who evaded the problems of a central mechanism. William James, whose pluralism led him to make a number of forays into problems of control, not always self-consistent, at one point assigned the role of the thinker to the passing thought. He was given to some excesses, especially when demolishing the views of those he opposed, in this case the transcendental ego of the post-Kantians: "Our 'thought'-a cognitive phenomenal event in time--is, if it exists at all, itself the only thinker that the facts require" (James, 1890, Vol. 1, p. 369). He went on to say a little later: "We may sum up by saying that the personality implies the incessant presence of two elements, an objective person, known by a passing subjective Thought and recognized as continuing in time. Hereafter let us use the words Me and I for the empirical person and the judging Thought" (Vol. 1, p. 371). All of this is in the chapter on the consciousness of self, in which he also gave favorable accounts of Janet's case of Leonie and of a fugue state with which he himself had dealt, that of the Reverend Ansel Bourne. It is hard to see how these larger unities could be accepted solely because of
NEODISSOCIATION THEORY OF MULTIPLE COGNITIVE CONTROL SYSTEMS
149
persistence in time and the passing thought as the thinker, but this extreme possibility cannot be ignored. In different words, this view asserts that all behavior is merely a modification of the entering behavior, and that way of putting it sounds more modem than James. The idea of a control system, however, also finds contemporary expression. If one examines the concept of planning, it appears that a sort of planner must be inferred. Even a simple matter such as making an appointment for luncheon next week is negotiated with those involved, is written down, and is then acted upon at a later time. This planfulness controls the possible behavior on that future date quite effectively. The many competing thoughts that occur the morning of the luncheon do not appear to be as determinative of what is going to happen in the midday as the plans that were made by some responsible part of the cognitive apparatus during the prior week. Appointments of this kind are kept with a very high probability-perhaps as high as 90% of the time--so that the planning function must be taken seriously. The event may be trivial, but its implications are not. The support for a special executive function has come into the open from an unlikely source--the computer. Heuristic computer programs commonly have an executive program that monitors the computer's attempts to solve problems (e.g., Newell and Simon, 1972). If one direction goes on too long without reaching a solution, the executive calls a halt and a new direction is taken. This close analogy to what a thinker does makes the idea of an executive ego a plausible one. The diagram of Figure 3 is therefore all right as a suggestion of some of the features of cognitive control systems, but it has to be developed in far greater detail in order to lead to decisive experimentation. For one thing, the indication of separate inputs and outputs for each system does not take into account that the organism has a limited set of receptors and effectors and that the subsystems will be in competition for information sources through these receptors. There is competition also for the final common path leading to action and for cognitive capacities that are limited. To make this more concrete, we may examine the kind of thing that happens when a person is immersed in one activity, controlled by one structure, and a strong invitation to another activity intrudes. To what extent will the conflict be resolved at the level of these structures themselves, and to what extent will a higher-order control process intervene?
150
ERNEST
R.
HILGARD
Suppose that someone is playing a well-practiced piece on the piano, without an audience (so that the social pressure to continue is reduced), when the selection is interrupted by a ringing telephone. He can hear the phone ringing without interrupting his play and can begin a conversation with himself, along the lines: "If I weren't home, I wouldn't answer; if it's important they will call again. I shall go on playing." The phone keeps on ringing intermittently, and he decides to interrupt his playing and to answer. How do we analyze this behavior, in terms of the cognitive systems involved and the central control mechanism? For convenience, we shall consider Cognitive Control Structure 1 to be the behavior related to playing the particular learned selection on the piano and Cognitive Control Structure 2 to be the habit of answering ringing telephones. These come into conflict, and the question arises whether the conflict can be resolved according to the relative strengths of these two systems, or whether some additional decision-mechanism is needed. A diagram of these relationships, which may be considered to be a further specification of Figure 3, is given in Figure 4. Playing the piano has now come so high in the hierarchy that it dominates the behavior of the piano player. Whose decision was it to play the piano at this time? Presumably there was some tendency to expression inherent in the background of piano playing, so that the performance was ready to be emitted; still, it had to be "permitted" to play against other competing possibilities. Here a central mechanism, involved in managing the daily activities, may have exerted an influence. Once the piano playing gets under way the central monitor steps aside and is a relatively passive observer, except for vigilance regarding other plans, for example, that would call for terminating the piano playing at a set time, even though the "piano playing part" might favor continuing. Now the ringing telephone intrudes. There is an arrow pointing from this new input to the piano playing input because this is an uninvited interruption. The new input arouses the second cognitive system that has habits regarding telephone answering, and a seIfconversation may ensue with little interruption of the piano playing if that is well automatized. The alternatives diagramed are either continuing to play or interrupting in order to answer the phone. The outputs are incompatible; there is a competition for the gross musculature that will either keep the person at the piano or move him to the telephone. The issue now is whether the conflict can be settled at the level of these
NEODISSOCIATION THEORY OF MULTIPLE COGNITIVE CONTROL SYSTEMS
151
Central Monitor and Control
Input VisualMusica I Score; Keyboard Auditory Kinesthetic
Cogn. Control
*
I
Planned Output A Play Piano
II
Playing Piano
Planned Output B Stop Ploying
Planned Output A Auditory Ringing Phone·, Repetition of Sound-Silence Alternation
Do not Answer Answering Telephone
Planned Output B Answer Phone
FIGURE 4. Resolution of a conflict between two cognitive systems: piano playing and telephone answering. At issue is the question whether the conflict between the two systems can be resolved by their relative strengths, or whether some intervention is needed by the central monitor and control.
two competing cognitive control systems, or whether the central control mechanism becomes involved. 1 The planning boxes have been added merely to call attention to the idea that some decision time is involved during which reflection upon the alternatives is possible. In the case of the telephone this is fairly short, because the caller may hang up and the phone stop ringing. In other cases the monologue may be more protracted. In a related case that came to my attention last summer, an organist became physically tired after having played most of the day in which a college baccalaureate and a commencement program occurred on the same day, with an organ recital between them. Part of the fatigue was in his legs, which he had to hold up much of the time in order to play on the foot levers of the organ. In the college hymn he foresaw some difficulty in moving his right leg and foot at a climactic part in 1
More complex motivations may enter, such as a preference for self-initiated activities as against those externally imposed, perhaps along with some unconscious ones based on the individual life history. For the purposes of illustration, the conflict is treated at a simplified level.
152
ERNEST
R.
HILGARD
which a move from the left extreme to the right had to be made. He told me that he talked to himself a lonb time about this, while playing right along, trying to decide whether or not he could do it accurately and what would happen if he omitted it. He decided to omit it, lest he make an error, and so far as I know nobody detected his omission. Does the central control have to be involved? This is somewhat problematical when everything is in the open, as in these illustrations, but many aspects of memory and of plans unrelated specifically to telephones and piano playing enter the picture. The ultimate weight of factors influencing the decision may well go beyond the two structures primarily involved. One way of putting it is that if there were an identifiable central agency, the chance of intrusion would be great. Here again the hypnotic model is of some pertinence: in a practiced hypnotic subject the monitoring part could simply assure not listening to the telephone by a self-suggestion not to hear it. This favors the piano player, but it does not seem to be something on which the piano-playing system alone would take the initiative. I have deliberately chosen an illustration in which the question of consciousness is secondary to the question of control. Once the notion of multiple controls is familiar, the possibility of inhibited cognitive structures becomes more plausible, and then one is ready to consider splits in consciousness. Figure 4 supplements Figure 3 by making more explicit the manner in which cognitive control structures compete, both in attention to stimulus inputs and for the final common path in outputs, and it shows circumstances in which the central mechanism might conceivably step in to influence a decision by throwing its weight (presumably via information from other sources) to resolve the conflict. This begins to get close enough to concrete situations that one senses the empirical problems that may be amenable to study, but it is not yet specific enough to define a program of research.
V.
EMPIRICAL ApPROACHES TO MULTIPLE CONTROL STRUCTURES AND DIVISIONS OF CONSCIOUSNESS
As noted earlier it is somewhat easier to approach these large problems by breaking them down into topics that lend themselves to
NEODISSOCIATION THEORY OF MULTIPLE COGNITIVE CONTROL SYSTEMS
153
experimentation, even though some of the issues may be bypassed temporarily. For example, the study of divided attention is very relevant to the issues being discussed here. One theoretical issue is whether or not when one of two channels of information is suppressed the suppressed channel is filtered out before registration or perhaps processed but the information not perceived. The former point of view has been made familiar in the filter theory of Broadbent (1958). Although Treisman (1969) permitted the information to move somewhat deeper into the processing mechanism, her views are not very different from Broadbent's on the issue of recognition and conscious perception. The extreme position was taken by Deutsch and Deutsch (1963), who said essentially that the information is processed in some kind of recognition system but diverted prior to conscious perception. For my purpose it is sufficient to call attention to the large literature bearing on these problems (e.g., Norman, 1969; Moray, 1970; Kahneman, 1973) and to note that at least one of the theories, that of the Deutsches, hypothesizes the very kinds of processes that our experiments on neodissociation have revealed: information that is incorporated within a cognitive system but does not reach the conscious level until special procedures are used to bring it to subsequent awareness. Investigations of attention characteristically study the division of attention between tasks, because a central problem is that of the parallel versus the serial processing of simultaneous inputs. Kahneman (1973) has pointed out that there has been some neglect of capacity models, which should explain how, when there is enough capacity, more than one cognitive activity can be engaged in, although if some task requires too much cognitive effort it excludes others. Such considerations are obviously relevant to the hypnotic dissociation experiments in which simultaneous tasks have to meet the special requirements of hypnotic suggestions, e.g., performing one task out of awareness while the other task is performed with full awareness. Kahneman's discussion of task interference does not consider hypnotic behavior, but it is relevant to the demands made upon subjects in hypnotic experiments because these often strain their cognitive capacities as they attempt to comply with the demands made upon them. Studies that bear on recoverable amnesia, whatever the source of the amnesia, are pertinent to neodissociation theory. The literature of
154
ERNEST
R.
HILGARD
psychopathology furnishes many clinical illustrations, and these must be incorporated into psychological science. There are also experimentally induced amnesias, outside hypnosis, that deserve consideration. Some of these have been studied in the context of state-dependent learning. If something learned in one state, as while under the influence of a drug, is forgotten in the nondrugged state but recalled again in the drug state, that is an experimental illustration of a recoverable amnesia. This arrangement is, of course, the paradigm of state-dependent learning. The literature has been reviewed by Overton, who is also one of the leading investigators in the field (Overton, 1972, 1973). An alternative method of studying the differential effects of the drugs has been widely used, especially with animal subjects. For example, a rat can be taught to tum right in a T-maze under the influence of a drug and to the left in the normal, nondrugged state. According to Overton (1973) drug discrimination can be established more rapidly than most sensory discriminations, and it is doubtful that the discriminations involve the sensory consequences of the drug action. Instead he believes that the differential responding may be based "on the dissociative barrier which impairs a transfer of training between the drug and no-drug conditions." The concept of dissociation employed here is consonant with that which I am using, that is, two types of behavior control through information isolated from each other. There are a number of experiments with human subjects which indicate the possibilities. Several anesthetics are "dissociative" in that they permit the patient to be analgesic but conscious and perhaps to remain amnesic for the operation after the experiment is over. Ketamine is such an anesthetic in common clinical use; patients are not unconscious, their cardiac functions and respiration are not depressed, and they can reply to the surgeon's questions. They may have some delirium, possibly as an accompaniment of the memory disorder; after the drug wears off the patient is amnesic for what has happened. Unfortunately there are few studies in which there has been any serious effort made to see if the memories are recoverable. In one study in which I participated, in which the drug thiopental (trade name, Sodium Pentothal) was used with volunteer normal subjects not undergoing surgery, the subject while intoxicated with the drug was able to learn simple paired associates and to demonstrate the capacity for picture recognition, but after a few hours' rest during
NEODISSOCIATION THEORY OF MULTIPLE COGNITIVE CONTROL SYSTEMS
155
which the drug wore off he was commonly amnesic for all that had happened and showed no retention of the material learned. However, when the subject was then hypnotized and asked to recall, there was some slight recovery of memory, especially in the more hypnotizable subjects (Osborn, Bunker, Cooper, Frank, and Hilgard, 1967). Statedependent learning was not demonstrated, in that under the drug there was no forgetting of material learned in the prior waking state, and when the drug was readministered there was no recall of the material learned in the earlier session and later forgotten. The experiment should be repeated, especially with selected highly hypnotizable subjects, because some pretest subjects, who were more hypnotizable than those used in the experiment, did show more dramatic recovery of their memories for things learned in the drugged state. There have indeed been some reports of the recall of things said by the surgeon in ordinary general chemoanesthesia, when later tested by hypnotic methods (Cheek, 1966; Levinson, 1965). In these experiments the action of the drugs was the primary consideration; hypnosis was used merely as a method of inquiry to determine whether or not the amnesia produced was recoverable. It makes a difference whether the absent memories have simply not made their way into the permanent memory store or whether they are present and in some sense dissociated and unavailable to recall by ordinary methods. One would like to know the physiology behind these states, but the physiology of consciousness is elusive in any case. One lead that has proved interesting is the differential function of the two hemispheres, which, at least in the split-brain studies, represents a massive dissociation of two brains in one head (Sperry, 1968; Gazzaniga and Sperry, 1966; Gazzaniga, 1970). As these studies are carried over into the study of normal laterality of function, it is quite possible that additional aspects related to dissociation will emerge. For example, it has been shown that in well-Iateralized right-handed males, a preference for right-hemisphere function is associated with hypnotizability (Gur and Gur, 1974). The distinction has to be made between specialization of function and preference, for those who consistently do "right-hemisphere tasks" with evidence that they are using the right hemisphere, and "left-hemisphere tasks" with evidence that they are using the left hemisphere, do not by these signs indicate hypnotizability. Given both kinds of tasks, the hypnotizable right-hander tends to use his right
156
ERNEST
R.
HILGARD
hemisphere more frequently than the nonhypnotizable right-hander, that is, he shows a preference for using that hemisphere. This is all somewhat puzzling, and there may be some intermediaries to provide the explanation, such as a preference for the use of visual imagery when presented with a problem. The dissociations within sleep provide another source of information that does not rely on extreme pathology. For example, the studies on sleepwalking and sleep talking (somniloquy) illustrate states in which considerable control may be manifested, including interactions with the geographical and social environment, yet forgotten when the person has returned to normal sleep (Arkin, Toth, Baker, and Hastey, 1970). The dreams recalled from the night are not usually connected with these episodes. Of course night dreams themselves remain interesting as dissociated activities. The evidence from psychopathology ought to be included in any account of dissociative phenomena. Although fugues and multiple personalities are rather infrequent, now that clinicians are again on the lookout for them, more are being found. After a dearth of cases of multiple personality, several have been reported since the case reported as the Three Faces of Eve several years ago (Thigpen and Cleckley, 1957; the new ones include Ludwig, Brandsma, Wilbur, Bendfeldt, and Jameson, 1972; Schreiber, 1973; Stoller, 1973). Others, unpublished, have come to my attention. Long ago Hart (1929), who had himself done much to popularize psychoanalytic concepts, felt that psychoanalysts had failed to take note of dissociation and double personality and that they should overcome this neglect. The dearth of cases during the height of psychoanalytic popularity could have resulted, if Hart was correct, because a neurotic who might have demonstrated these symptoms was most likely to be treated by an analyst. If at the first sign of a split the analyst began to seek an integration of the parts (perhaps quite properly from a therapeutic standpoint), the extent of the dissociation might not have become fully revealed. The most interesting and puzzling feature of multiple personalities is nearly always the distribution of memories among the components. In Janet's case of Leonie, for example, Leonie I knew only herself, Leonie II knew Leonie I as well as herself, and Leonie III knew them all. In the more recent case of Jonah, the primary personality (Jonah) was not aware of the three others, though they were all aware of Jonah and dimly aware of each other (Ludwig et al., 1972).
NEODISSOCIATION THEORY OF MULTIPLE COGNITIVE CONTROL SYSTEMS
VI.
157
THE DUALITY OF RESPONSIVENESS TO PAIN AS RELATED TO NEODISSOCIATION THEORY
The consciousness of pain can serve as an illustration of specific problems that a neodissociation theory is designed to deal with. An advantage of this area of investigation is that the neuroanatomical problems, if not all resolved, are at least well identified because of the many efforts to relieve pain by surgery and drugs and by non drug procedures, of which hypnosis is one. The basic facts of supra threshold pain can be summarized as follows: 1. Felt pain has two components related to a source of stimulation, of which one is the sensory-information aspect, which makes pain very much like other sense-perception experiences. The sensory pain can usually be localized, quite precisely, as in cutaneous pains, or at least regionally and often laterally, as in toothache, headache, or abdominal pains. Sensory pain increases with the intensity of stimulation and can be studied by standard psychophysical methods. The second component is variously described as the reactive component (e.g., Beecher, 1959) or the motivational-emotional component (Melzack and Casey, 1968). The second component is also referred to as anguish or suffering, to distinguish between this aspect and sensory pain. We may temporarily neglect a third component, the central or cognitive control of pain. There are some issues here reminiscent of the James-Lange theory of emotion. The reactive interpretation assumes that the sensory pain comes first in time and the suffering follows upon it, just as in the James-Lange theory the visceral or motor responses were said to come first and the emotion to come as a repercussion from them ("1 am afraid because I run"). The Cannon-Bard criticism was that both the responses and the felt emotion might be released at once by way of different systems, so that the two aspects could be viewed as concurrent rather than as successive. The same issue arises in whether pain and suffering are concurrent or successive; before we decide that issue, it is preferable to note the distinction between the two aspects and to await the evidence from experiments. 2. The two components have an anatomical basis in separate nerve pathways in the spinal cord and higher regions, with differing central connections. There are still some ambiguities here, but in general the sensory conduction system is by way of neospinothalamic fibers to the
158
ERNEST
R.
HILGARD
thalamus and the somatosensory cortex, and the motivational-affective conduction system is by way of medially coursing fibers comprising a paramedial ascending system reaching the reticular formation, the medial intralaminar thalamus, and the limbic system (e.g., Melzack and Casey, 1968; Casey, 1973). If these two systems are recognized, then the possibility of the simultaneous occurrence of pain and suffering has to be taken into consideration before commitment to a totally reactive position with respect to suffering. Some difficulties are encountered in the clearing up of all of the issues because the sensory conduction system is indeed the more rapid one, and even though the motivational-emotional system has been activated concurrently with the sensory system, it is quite likely that the suffering is in some way moderated by feedback from the prior sensory information. Hence the motivational-emotional system may be partly aroused in parallel while also being reactive to the sensory component. 3. The interactions between the two components may be either integrated or dissociated. The possibilities of integration have been well explored in the gate-control theory of pain (Melzack and Wall, 1965). This theory proposes that an interaction between the two systems occurs at the spinal cord level, in the substantia gelatinosa of the dorsal hom. The anatomical basis is provided by collaterals from the large fibers of the sensory-information system and the small fibers of the motivational-affective system entering through the dorsal roots of the cord. A further control is exercised by central processes that operate centrifugally, sending impulses down to the gate from above. If large fibers are stimulated simultaneously with small ones, the theory proposes that the gate should be closed and suffering reduced; this implication found immediate practical application in the control of deep pain by superficial stimulation (e.g., Wall and Sweet, 1967). It may be that this is one of the routes by which acupuncture is effective; it also offers some support for the folk practices of relieving deep pain by liniments and mustard plasters that produce surface stimulation of the large fibers in the regions in which small-fiber stimulation is occurring. The two systems may also be separately affected. One of the observations made when lobectomy was used in the attempted relief of intractable pain was that patients often reported that they still felt the pain but that it no longer bothered them (Freeman and Watts, 1950). The cutting of the paths in the prefrontal area was probably responsible
NEODlSSOCIATION THEORY OF MULTIPLE COGNITIVE CONTROL SYSTEMS
159
for destroying some of the responsiveness of the motivational-affective pain system while leaving the innocuous sensory-information processing intact. The separateness of the two systems allows for psychological dissociati~n because nonsurgical procedures may modify one system more than the other. 4. Pain mechanisms are incompletely described by any explanation that relies exclusively on a segmental interpretation of impulse transmission and local signs. The notion of dermatones related to specific segments of the spinal cord is so firmly ingrained that wellestablished alternative distributions of pain are treated as anomalous, if not ignored altogether. The first of these anomalies is referred pain, that is, pain that is felt in some part of the body not the source of the irritation. A referred pain violates the principle of local signs that usually locate the source of stimulation. Referred pains are well known, and diagrams can be found in most neurology texts. Often they are referred within the same segment, as when the pain of a diseased kidney is felt as a pain in the normal one or anginal pain is felt in the arm. But the pain may be in a different segment, as when a lower-back injury may be felt in the eye region or a knee injury in the wall of the chest. An interesting set of observations concerns so-called trigger points, that is, points which when stimulated give rise to referred pains at a distance. If these points are anesthetized-or in some cases hyperstimulated-the referred pain may be reduced (Travell and Rinzler, 1952). Rather than being viewed as anomalies, these anatomical complexities should be incorporated into our understanding of the total neurophysiology of pain perception. Some of these nonsegmental indications are very dramatic, as in the case of a patient with a transected cord whose right leg is totally analgesic, yet who feels a pain on the left side in the abdomen when a particular area of the analgesic right leg is stimulated (Nathan, 1956). The autonomic system may be involved when the referred pains do not follow usual segmental somatic patterns. The nonsegmental aspect of hysterical anesthesias, such as glove anesthesias, or corresponding nonsegmental changes induced by hypnosis become more acceptable when we recognize the many anomalies of pain in varied contexts. 5. Another fact of the pain experience is the persistence of pains, whether referred or not, after the source of irritation no longer exists.
160
ERNEST
R.
HILGARD
Phantom limb pains fall into this category, that is, the pains that are felt in an amputated limb that no longer exists. The possibility that this is merely an illustration of the "law of specific energies," with pain produced through stimulation of the cut ends of nerves formerly serving the limb, is not a sufficient explanation, for the pain may persist even when the stump is anesthetized. Melzack (1973) points out that the pain is seldom felt unless there was pain in the bodily member prior to its loss. Hence there is perhaps a memory component that assigns the pain to the phantom. The persistence of pain reduction, as in the continuing effects of superficial stimulation of large fibers after the stimulation ceases, also has to be accounted for by explanations that go beyond any immediate sensory process of stimulation and response. The explanation again must be in terms of some sort of persisting process, whether peripheral or central. There is much more that might be said about pain; for example, many social and psychological factors are known to influence it. There is the athlete who completes the game without being aware of his injury until the game is over, or the wounded soldier who does not express pain until stuck by a hypodermic needle (Beecher, 1956). The modem "pain clinic" takes advantage of these facts and has numerous methods by which to ease the pain of patients without the use of drugs or surgery. Sternbach (1970, 1974) has listed a number of these such as desensitization, progressive relaxation, biofeedback and conditioning methods, and strategems of distracting attention, as well as hypnosis. Our earlier experiments on laboratory pain did not distinguish between pain and suffering, confusing the issue slightly by using an aversive component in defining the pain scale as one on which 10 would be a pain so severe that the subject would wish to terminate it. This did not matter for the original purposes, but it became important when the experiments were extended to the study of the distinction between pain and suffering. The first published experiment from our laboratory which clearly used the distinction was that of Knox, Morgan, and Hilgard (1974). The experiment employed two reporting scales, one for sensory pain, one for suffering. In the ordinary nonanalgesic hypnotic condition, the subjects were asked for separate reports of pain and suffering at 2minute intervals. They found it easy to make the distinction; both pain and suffering rose with time of ischemia, the suffering mounting more slowly than the pain, as shown in Figure 5. To determine whether there
161
NEODISSOCIATION THEORY OF MULTIPLE COGNITIVE CONTROL SYSTEMS
12~~----------r---------.----------r---------'--'
N=8 Critical Level
IO~---------------------------
8 ~ 0
u
(f)
C7'
c
6
0Co a:: 4 Sensory Pain
0--0 Hypnotic Verbal (Open)
...-- Automatic Talking (Hidden)
2
Suffering
0- -{) Hypnotic Verbal (Open)
... -e
Automatic Talking (Hidden)
OL-~----------L---------~--------~--------~--~
o
2
4 Minutes of Ischemia
6
8
FIGURE 5. Reports of pain and suffering in ischemia in the hypnotic nonanalgesic state. Reports of pain and suffering were obtained every 2 minutes in the usual way and plotted as Open reports. Immediately following each open report, a report was obtained by the method of automatic talking and plotted as Hidden reports. In the hypnotic nonanalgesia condition the two sets of reports are essentially alike, indicating no amnesic effect on the open reports. (From Knox, Morgan, and Hilgard, 1974, p. 844.)
might be some amnesiac component even in the nonanalgesic report by these highly hypnotizable subjects, a report by the "automatic talking" method used for studying experiences concealed under amnesia was used; the reports were practically identical, as shown by the overlapping curves in Figure 5, so that there was no evidence for an amnesic component under the nonanalgesic condition. When the corresponding reports were called for in hypnotic analgesia, the verbal report obtained in the usual way in hypnosis was of no pain and no suffering throughout; the hidden report, of the experience
162
ERNEST
R.
HILGARD
concealed under amnesia, was of almost full pain and suffering. This was puzzling because there were no overt signs of suffering within hypnotic analgesia, and retrospective accounts following hypnotic analgesia with other subjects had indicated that the concealed pain was felt normally in hypnotic analgesia but that it was not aversive. Some doubts were expressed in the report about the findings. One possibility noted there was that the hypnotic nonanalgesia may have itself resulted in a quiescent subject, so that although the pain was normal, the suffering was not very severe to begin with. Perhaps the report of suffering was inferred to be the amount of suffering that ought to be felt for that amount of pain, in view of the definition of both scales as converging at a pain state of 10, for at that point the subject would find the pain (hence the suffering) sufficiently aversive as to want the experiment terminated. Owing to the repeated call for concealed reports while the stimulation was occurring, the subject might have adopted a practice of "sampling" the sensory pain by momentarily turning off the analgesia and then turning it back on, inferring what suffering would accompany that amount of pain if the analgesia were broken for a longer time. In any case, it seemed desirable to repeat the experiment in some other manner. A few pilot subjects have shown what the results are likely to be; the new experiments are still in progress and a full account is not yet available, but the direction of the results differs enough from the reported experiment to justify some comments. The pilot subjects of the new experiment have been run first in a normal waking condition, rather than a nonanalgesic hypnotic one, with reports of pain and suffering at 30-second intervals. They have responded as the earlier subjects did, by reporting pain and suffering both increasing with time in ischemia, with the suffering less than the concurrent pain. Then, in another session, the ischemia was continued with hypnotic analgesia for a sufficient time after exercising for the pain and suffering to have become quite aversive, according to the reports obtained without analgesia. The earlier findings were repeated: no pain and no suffering were reported. After the ischemic arm was restored to normal, the subject, while still hypnotized, was questioned in the manner that provides the hidden report, breaking the amnesia for the experience. The consistent finding with two pilot subjects has been that the hidden sensory pain was essentially normal, reaching the level of the non analgesic day, but the suffering was entirely absent. This contradicts the results of Knox et al. (1974) but is more coherent with other
NEODISSOCIATION THEORY OF MULTIPLE COGNITIVE CONTROL SYSTEMS
163
observations made in the laboratory. The discrepancies point to the hazards of interpretation in this kind of experiment, but presumably further experiments, with the same subjects serving under various aITangements, will eventually resolve the discrepancies. At least for these pilot subjects, we have some complex results according to which to test the theory of duality of pain and the role of hypnosis in relation to it. The situation is characterized in Figure 6. The normal waking nonanalgesia pattern of consciousness is shown on the left, with the central control permitting the full experience of pain, and the central monitor then noting the pain and suffering felt and reporting it verbally. "Permitting" pain is said advisedly, because these subjects need not have felt pain and suffering unless they had consented to. The diagram on the right shows what is happening in a concealed way while the central monitor is reporting no pain and no suffering. The central control system has been responsible for accepting the hypnotic suggestions for pain reduction, which includes setting up the amnesia barrier against reporting what the hidden consciousness is experiencing. Only when the amnesia barrier is broken by the "automatictalking" procedure does the report of the hidden material find its verbal expression. Separating the central control from the central monitor may seem arbitrary, but phenomenally this is what takes place. Some aspect of the person (the central control) permits the ascendance of the hypnotic responsiveness, and once it is operative, the monitor rather passively reports the experiences that are being felt. The lower boxes indicate that the involuntary physiological components, such as the cardiovascular changes, reflect the physical stresses much as in the waking state, but the voluntary responses of fist clenching, squirming, and grimacing are absent in hypnotic analgesia. The absence of the voluntary components may be due to less felt suffering, or the reduction in responses may be responsible for reducing the suffering; present data do not permit a choice between these interpretations. The bearing on the pain-suffering dichotomy may be formulated in three assertions: 1. Hypnotic analgesia, on the part of highly responsive hypnotic subjects, eliminated both sensory pain and suffering and hence did not discriminate between these two components. The findings are clear evidence, however, for central control of felt pain and suffering. This is
164
ERNEST
R.
HILGARD
NORMAL WAKING
CENTRAL ASPECTS
SENSORY PAIN
SUFFERING
PHYSIOLOGICAL CONCOMITANTS
r------i Centro I Control
Central Monitor
Pain Fell
Suffering Felt
Voluntary Components Present Involuntary Components Present NONANALGESIA
6. The dissociation between the usual verbal reports in hypnotic analgesia and those revealed through automatic key-pressing or automatic talking, interpreted according to an amnesia barrier. The diagram on the left represents the situation in normally experienced pain, with the central control permitting the pain and suffering; the central monitor is aware of all that is going on. The diagram on the right represents the situation FIGURE
in itself a dissociation between stressful stimulation and consciousness of what is happening. 2. The hidden report, made after the amnesia barrier is broken, shows that sensory pain is less reduced at this level than suffering. This differentiation represents a pain-suffering dichotomy and is in that
NEODISSOCIATION THEORY OF MULTIPLE COGNITIVE CONTROL SYSTEMS
165
HYPNOSIS Verbal Report
Central Control
Central Monitor
------, L- ---'1
II II
JI
Pain Felt
I I I J
Suffering Reduced or not Felt o
Voluntary Components Absent
.,
.iii c
E
Involuntary Components Present HYPNOTIC ANALGESIA -HIDDEN ASPECT in hypnotic analgesia, the central control having set up the amnesia barrier, making the experiences of pain and suffering unavailable to the central monitor. Only when the central control releases the amnesia barrier (through the automatic methods) will the experience of pain and reduced suffering be available to the central monitor.
way coherent with the physiological and anatomical evidence and with one aspect of the gate-control theory. There is a dissociation here between the normally simultaneous sensory pain and suffering. 3. Because the reduction in suffering is accompanied by a reduction in the voluntary components of the total reaction, those who favor
166
ERNEST
R.
HILGARD
a reaction theory of suffering can assimilate this finding to their position. That is, when sensory pain is registered but there are no voluntary responses to its aversive character, the sensory pain may be assumed not to lead to reported suffering. This interpretation could account for a separation between pain and suffering even if there were only a single afferent system involved. However, the data can also be interpreted on the assumption that the afferent suffering system, relatively independent of the sensory pain system, may itself be responsible for evoking the voluntary indications of suffering. Now to summarize the bearing of these assertions on neodissociation theory. With respect to the sensory pain aspect, the bearing on dissociation is clear, for here is a sensory experience registered within some cognitive system, yet denied by another. The essence appears to lie in an amnesialike process that is responsible for making unavailable the information about either pain or suffering in the usual hypnotic analgesia report. If suffering is reported in neither the usual report nor the report of the concealed experience, then the amnesic aspect is not relevant for suffering; now, however, the dissociation between pain and suffering has to be accounted for on other grounds. The dissociation is still important because pain and suffering are usually found together, and the fact of different physiological systems does not account for the hypnotic dissociation; the separate systems hint only at a possible physiological locus for the effect of hypnotic analgesia. To illustrate how the phenomena in this kind of experiment and this kind of theory may be used to test specific hypotheses about pain and pain controC I wish to mention a hypothesis suggested by Dr. Avram Goldstein, which he and I are now engaged in testing. 2 As indicated above, the dissociation of the pain and suffering systems that shows in the hypnosis experiments may have a basis in the separate conduction systems. What would be a possible mechanism by which this might come about? It could come about if hypnotic analgesia were to modify the central neural activity occurring in the medial portions of the thalamus that are related to the suffering system but not to the sensory pain system. It is known that destroying the medial thalamus surgically may reduce the affective component of intractable pain with2
I am grateful to Dr. Goldstein for proposing the hypothesis, for participating in carrying out the relevant collaborative experiment, and for permitting this reference to it prior to its publication. (Reference added in proof: Goldstein and Hilgard, 1975.)
NEODISSOCIATION THEORY OF MULTIPLE COGNITIVE CONTROL SYSTEMS
167
out destroying the sensory discriminative capacity (Mark, Ervin, and Yakolev, 1963). Morphine injected into this area may have the same effect. Interestingly enough, electrical stimulation in the same region may also reduce pain. This was demonstrated by Reynolds (1969), who operated on rats without their showing signs of pain while the electrical stimulation continued; the rats then showed signs of pain when the stimulation ceased. Corresponding findings in the cat have been reported by Liebeskind, Guilbaud, Besson, and Oliveras (1973). Is there any connection between morphine and electrical stimulation? From all we know about neurohumoral transmission, there is a strong possibility that the electrical inhibition of pain depends upon a chemical intermediary; perhaps this chemical intermediary has some pharmacological similarities to morphine. We now have a background for Dr. Goldstein's hypothesis. Suppose that the pain-reducing intermediary for pain reduction in medial thalamic stimulation is morphine-related. How can this be found out? One method is to use a known morphine antagonist to counteract whatever agent is responsible for the pain reduction. When electrical stimulation produces the analgesia, naloxone, a morphine antagonist that is alone neutral with respect to pain, does in fact counteract the pain reduction and hence gives support to the hypothesis (Mayer and Hayes, 1973). Suppose, now, that hypnotic analgesia were to work in the same way. Then it might be responsible for the release of a morphinelike substance somewhere in the medial system; sensory pain would not be reduced but suffering would, which is the result reported in hypnosis for the hidden reactions to normally painful stress. The hypothesis can then be advanced that naloxone might counteract the hypnotic analgesia, as it counteracts the analgesias of morphine and electrical stimulation. The hypothesis offered by Dr. Goldstein was tested in a doubleblind experiment in which pilot subjects known to be responsive to hypnotic analgesia were sometimes injected intramuscularly with normal saline solution as a placebo and sometimes with naloxone. If the hypothesis were confirmed, no change in the amount of reduction of pain and suffering under hypnotic analgesia as usually reported would be required, because the underlying changes are masked by some sort of amnesic process. After the release of the amnesia, by the methods of exploring the hidden effects, a change with naloxone would be ex-
168
ERNEST
R.
HILGARD
pected. The suffering, normally absent, would be expected to return if naloxone neutralized a morphinelike substance responsible for the reduction of suffering in the first place. The preliminary results have failed to confirm the hypothesis; that is, there was no increase in reported suffering after the amnesia was lifted, whether naloxone or the placebo had been administered. Of course negative results are somewhat uncertain at this stage of the investigation, but at least no positive effects of naloxone in reversing the suffering have thus far been found. An important point to be made is that the ordinary procedures of hypnotic analgesia would not have been critical for this experiment, because even if the experiment had in other ways shown positive results for the effect of naloxone, there would have been no effect on the reduction of pain and suffering in the usual reports of hypnotic analgesia. The theoretical reason for the persistence of the analgesic effect is that this usual report has an amnesic component in it, so that no pain or suffering would be reported regardless of what was happening at some other level of information processing. It was only under circumstances in which the dissociation of pain and suffering could be shown that the experimental test became critical. As a result, a series of experiments related to neodissociation theory opened up the possibility of testing a neuropharmacological hypothesis. The experimental design has permitted a crucial test; that the results thus far have been negative is a matter of empirical fact. After all, the purpose of an experiment is to find out what is so.
VII.
CONCLUSION
A new theory is valuable when its formulation calls attention to problems that have been neglected and shows ways in which, with the aid of the theory, the problems can be converted to experimentable form in a manner that would be unlikely without the theory. In this introduction to neodissociation theory I have sketched its historical background and the kinds of general observations on which it is based. The example of hypnosis in the investigation of pain and suffering was elaborated in order to show in concrete detail the empirical consequences of exploring phenomena in the context of neodissociation theory. While hypnosis dramatizes the existence of dissociative phenomena, neodissociation theory serves to call attention to many empiri-
NEODISSOCIATION THEORY OF MULTIPLE COGNITIVE CONTROL SYSTEMS
169
cal problems on nonhypnotic topics in which dissociative aspects are also prominent. The ultimate aim of neodissociation theory is to achieve a coherent understanding of these diverse phenomena. 3 ACKNOWLEDGMENT
The preparation of this chapter and the conducting of the experiments reported here were aided by Grant #MH-03859 from the National Institute of Mental Health. REFERENCES
ARKIN, A. M., TOTH, M., BAKER, J., AND HASTEY, J. M. The degree of concordance between the content of sleep-talking and mentation recalled in wakefulness. Journal of Nervous and Mental Disease, 1970,151, 375--393. BEECHER, H. K. Significance of wound to pain experienced. Journal of the American Medical Association, 1956,161, 1609-1613. BEECHER, H. K. Measurement of subjective responses. New York: Oxford University Press, 1959. BERLYNE, D. E. Structure and direction in thinking. New York: Wiley, 1965. BROADBENT, D. Perception and communication. London: Pergamon, 1958. CASEY, K. L. Pain: A current view of neural mechanisms. American Scientist, 1973,61, 194200. CHEEK, D. The meaning of continued hearing sense under general chemo-anesthesia: A progress report and report of a case. American Journal of Clinical Hypnosis, 1966,8, 275--280. CLEMES, S. R. Repression and hypnotic amnesia. Journal of Abnormal and Social Psychology, 1964,69, 62-69. DEUTSCH, J. A., AND DEUTSCH, D. Attention: Some theoretical considerations. Psychological Review, 1963,70, 80-90. FREEMAN, W., AND WATTS, J. W. Psychosurgery in the treatment of mental disorders and intractable pain. Springfield, Ill.: Thomas, 1950. GAZZANIGA, M. S. The bisected brain. New York: Appleton-Century-Crofts, 1970. GAZZANIGA, M., AND SPERRY, R. Simultaneous double discrimination response following brain bisection. Psychonomic Science, 1966,4, 261-262. GILL, M. M., AND BRENMAN, M. Hypnosis and related states: Psychoanalytic studies in regression. New York: International Universities Press, 1959. GOLDSTEIN, A. AND HILGARD, E. R. Failure of the opiate antagonist naloxone to modify hypnotic analgesia. Proc. Nat. Acad. Sci. USA, 1975, 72, 2041-2043. GUR, R. E., AND GUR, R. C. Handedness, sex, and eyedness as moderating variables in the relationship between hypnotic susceptibility and functional brain asymmetry. Journal of Abnormal Psychology, 1974,83,635--643. HART, B. Psychopathology (2nd ed.). Cambridge, England: Cambridge University Press, 1929. 3
The sketch of neodissociation theory in this chapter is elaborated in Hilgard (in preparation) .
170
ERNEST R. HILGARD
HARTMANN, H. Ego psychology and the problem of adaptation. New York: International Universities Press, 1958. HEBB, D. O. The organization of behavior. New York: Wiley, 1949. HEBB, D.O., Science and the world of imagination. Canadian Psychological Review, 1975, 16, 4-11. HILGARD, E. R. A neodissociation interpretation of pain reduction in hypnosis. Psychological Review, 1973,80, 396-411. HILGARD, E. R. Divided consciousness. New York: Wiley-Interscience (in preparation). HILGARD, E. R., MORGAN, A. H., AND MACDONALD, H. Pain and dissociation in the cold pressor test: A study of hypnotic analgesia with "hidden reports" through automatic key-pressing and automatic talking. Journal of Abnomzal Psychology, 1975,84, 280-289. HULL, C. 1. The concept of the habit-family hierarchy and maze learning. Psychological Review, 1934,41, 33-52, 134-152. JAMES, W. Principles of psychology 2 vols. New York: Holt, 1890. JANET, P. L'Automatisme psychologique. Paris: A1can, 1889. JANET, P. The major symptoms of hysteria. New York: Macmillan, 1907. KAHNEMAN, D. Attention and effort. Englewood Cliffs, N. J.: Prentice-Hall, 1973. KNOX, V. J., MORGAN, A. H., AND HILGARD, E. R. Pain and suffering in ischemia: The paradox of hypnotically suggested anesthesia as contradicted by reports from "the hidden observer." Archives of General Psychiatry, 1974,30, 840-847. LEVINSON, B. W. States of awareness during general anesthesia-preliminary communication. British Journal of Anesthesiology, 1965,37, 544-546. LEWIN, K. A dynamic theory of personality. New York: McGraw-Hill, 1935. LIEBESKIND, J. c., GUILBAUD, G., BESSON, J. M., AND OLIVERAS, J. L. Analgesia from electrical stimulation of the periaqueductal gray matter in the cat: Behavioral observations and inhibitory effects on spinal cord interneurons. Brain Research, 1973,50, 441-446. LUDWIG, A. M., BRANDSMA, J. M., WILBUR, C. B., BENDFELDT., F., AND JAMESON, D. H. The objective study of a multiple personality. Archives of General Psychiatry, 1972, 26, 29&-310. MARK, V. H., ERVIN, F. R., AND YAKOVLEV, P. I. Stereotactic thalamotomy. Archives of Neurology, 1963,8, 52&-538. MAYER, D. J. AND HAYES, R. L. Stimulation-produced analgesia: Development of toleran"" and cross-tolerance to morphine. Science, 1973, 188, 941-942. MEIZACK, R. The puzzle of pain. New York: Basic Books, 1973. MEIZACK, R., AND CASEY, K. 1. Sensory, motivational, and central control determinants of pain: A new conceptual model. In D. Kenshalo (Ed.), The skin senses. Springfield, Ill.: Thomas, 1968. MEIZACK, R., AND WALL, P. D. Pain mechanisms: A new theory. Science, 1965,150,971979. MILLER, G. A., GALANTER, E., AND PRIBRAM, K. H. Plans and the structure of behavior. New York: Holt, 1960. MORAY, N. Attention: Selective processes in vision and hearing. New York: Academic Press, 1970. NATHAN, P. W. Reference of sensation at the spinal level. Journal of Neurology, Neurosurgery, and Psychiatry, 1956,19,8&-100. NEWELL, A., AND SIMON, H. A. Human problem solving. Englewood Cliffs, N.J.: PrenticeHall, 1972. NORMAN, D. A. Memory and attention: An introduction to human information processing. New York: Wiley, 1969.
NEODISSOCIATION THEORY OF MULTIPLE COGNITIVE CONTROL SYSTEMS
171
OSBORN, A. G., BUNKER, J. P., COOPER, L. M., FRANK, G. S., AND HILGARD, E. R. Effects of thiopental sedation on learning and memory. Science, 1967,157, 574--576. OVERTON, D. A. State-dependent learning produced by alcohol and its relevance to alcoholism. In B. Kissin and H. Begleiter (Eds.), The biology of alcoholism Vol. 2. New York: Plenum Press, 1972. OVERTON, D: A. State-dependent learning produced by addictive drugs. In S. Fisher and A. M. Freedman (Eds.), Opiate addiction: Origins and treatment. Washington, D. c.: Winston, 1973. REYNOLDS, D. V. Surgery in the rat during electrical analgesia induced by focal brain stimulation. Science, 1969,164, 444-445. SARBIN, T. R., AND COE, W. C. Hypnosis: A social psychological analysis of influence communication. New York: Holt, Rinehart and Winston, 1972. SCHREIBER, F. R. Sybil. Chicago: Regnery, 1973. SPERRY, R. W. Hemisphere disconnection and unity in conscious awareness. American Psychologist, 1968,23, 723-733. STERNBACH, R. A. Strategies and tactics in the treatment of patients with pain. In B. L. Crue (Ed.), Pain and suffering: Selected aspects. Springfield, III.: Thomas, 1970, pp. 17~185.
STERNBACH, R. A. Pain patients: Traits and treatment. New York: Academic Press, 1974. STOLLER, R. J. Splitting: A case of female masculinity. New York: Quadrangle Books, 1973. THIGPEN, C. H., AND CLECKLEY, H. M. The three faces of Eve. New York: McGraw-HilI, 1957. TOLMAN, E. C. Purposive behavior in animals and men. New York: Century, 1932. TRAVELL, J., AND RINZLER, S. H. The myofascial genesis of pain. Postgraduate Medicine, 1952, 11, 425-434. TREISMAN, A. M. Strategies and models of selective attention. Psychological Review, 1969, 76, 282-299. WALL, P. D., AND SWEET, W. H. Temporary abolition of pain in man. Science, 1967,155, 108-109. WHITE, R. W., AND SHEVACH, B. J. Hypnosis and the concept of dissociation. Journal of Abnormal and Social Psychology, 1942,37, 309-328.
5
Hypnotic Susceptibility, EEG-Alpha, and SelfRegulation DAVID
I.
R.
ENGSTROM
INTRODUCTION
Hypnosis is probably more popularly known, revered, feared, and mystified than any other informational control known to man. Its methods, motives, and outcomes have been bent out of shape by hundreds of old movies and old wives' tales. The mass media have formed popular knowledge of hypnosis largely from improbable clinical applications and naively effective coercive control by bad guys. Because of this kind of popularity, and despite its responsible professional use in the treatment of certain behavior disorders, hypnosis has never gained much respectability outside of its own practioners and fans. This is almost surely because until the last several decades its application has not been systematically investigated. And it has not been widely used because it has not been systematically understood, especially in terms of its physiological and psychological correlates. Despite its long history, it is difficult to be very specific about hypnosis. There are several good reasons for this. First, the hypnotic state has been hard to define by any rigorous behavioral, subjective, or physiological boundaries. The same is true for most other psychological states, which are similarly influenced by the subjectivity of observer and participant. Most internal states are too unstable within the same person and inconsistent between people for much generality at all. DAVID R. ENGSTROM . Department of Psychiatry & Human Behavior and Student Health Service, University of California, Irvine, California.
173
174
DAVID
R.
ENGSTROM
Secondly, hypnosis can mean two very different things. It is popularly known as a distinct psychological state, with unique attributes that separate it from other states. But it is also an operation, or a label for a set of operations that constitute the hypnotic situation, or Uwhat hypnotists do." So the word can mean either a state or an instructional set. And, in fact, the most basic theoretical difference among scientific investigators of hypnosis centers around whether hypnosis is best defined as the private, internal responses of the hypnotized subject or by external characteristics implicit in the hypnotic situation. The definition of hypnosis has always been impure, and it is by virtue of this impurity that hypnosis is an appealing model for eliciting the effects of words on behavior, of instructions on action, and of cognitive control on performance. As an operation, it serves to magnify the subtleties of human informational control, thereby exposing them for systematic manipulation as variables. In the process it becomes a caricature of all communication, which makes it seem more funny than scientific. At times, it is certainly both. Hypnosis is most often characterized as a unique state of consciousness, either concurrent with or antecedent to specific behaviors. It can be arrived at either arbitrarily in life situations of through formal induction techniques. Certainly the "trance state" has been the central explanatory concept in most analyses of hypnotic phenomena that cannot be attributed to observable input, or to more objective role-theory or motivational accounts. But this position has often been challenged, both in theoretical argument and in objective fact. The number of behaviors that can be called unique to the hypnotized subject have been diminished by these challenges. Many behaviors considered "hypnotic" have been exactly replicated by unhypnotized subjects who are either encouraged to pretend to be hypnotized or are simply given a pep talk about what is expected of them in careful laboratory studies. Experienced hypnotists cannot always tell the difference. And the phenomenal reports of subjects who have been hypnotized are often misleading because of their demonstrated close relationship to the demands upon and expectations of the subject in the hypnotic situation. To date, no one has adequately defined the conditions that are both necessary and sufficient for hypnosis to occur. Under some circumstances, some subjects will enter hypnosis spontaneously, with-
HYPNOTIC SUSCEPTIBILITY, EEG-ALPHA, AND SELF-REGULATION
175
out formal induction procedures, while others never become hypnotized despite exhaustive formal induction techniques. The most important advance in the scientific study of hypnosis has been the development of valid, reliable scales to measure individual differences in response to standardized hypnotic procedures. The question "What is hypnosis?" is replaced with the more answerable "What responses are brought about among subjects exposed to hypnotic operations?" Defined as the degree to which an individual is able to enter into hypnosis and become involved in its characteristic behaviors and experiences under standardized conditions, hypnotic susceptibility is an ability, in which people vary. Most measures to susceptibility are direct and standardized. They are designed to assess the behavioral and subjective limits of hypnotic involvement of different individuals during and following the same hypnotic induction and instruction procedures.
II. THE
ASSESSMENT OF HYPNOTIC SUSCEPTIBILITY
A. Early Objectification The oldest problem for systematic investigators of hypnosis is response specification, that is, the identification of a set of criterion behaviors elicited following hypnotic instructions. Many clinical practicioners of hypnosis have developed observational systems for the assessment of responsiveness to or of depth of hypnosis. Braid (1843) used the criterion of spontaneous amnesia for events occurring during trance to differentia~e highly susceptible subjects. From this early specification of a single criterion as a measure of response or state, there developed a large number of more complex systems, specifying discrete kinds of hypnotic states. Charcot (1882) and his co-workers described three discrete types of hypnotic state: catalepsy, lethargy, and somnambulism. Each was characterized by unique behavioral attributes including waxy flexibility of the limbs, passive unresponsiveness, and automatic responsiveness to verbal suggestion, respectively. From the time of Charcot through the end of the 19th century, methods of scaling hypnotic susceptibility were qualitative and evolved in two ways. First, there was the addition of many more
176
DAVID
R.
ENGSTROM
discrete states to earlier criteria, still with the assumption that each state was unique and separate from the others. And second, there is evidence of beginnings of the idea that hypnosis can be measured unidimensionally and continually, that subjects who respond to a given degree of depth would show all of the symptoms of lesser degrees on the same scale (Bernheim, 1888; Liebeault, 1889). Hilgard (1965) has suggested that, in a sense, these continuous categories have the properties of a Guttman-type scale, in that response at one level indicates that a subject will probably be able to perform all of the behaviors at lower levels. Another important characteristic of this scale is the emergence of a continuous ability factor that varies from person to person. Fortunately records of subjects were frequently kept by investigators in the late 19th century. Hilgard, Weitzenhoffer, Landes, and Moore (1961) have combined these records and standardized subject classification on a four-point scale of responsiveness. The distribution they obtained for 14 early investigators, including 19,534 subjects in these categories, was (1) refractory or nonsusceptible, 9%; (2) drowsylight, 29%; (3) hypotaxy-moderate, 36%; and (4) somnambulisticdeep, 26%. Hilgard (1965) cautioned that there are several uncontrollable variables in such a study that would be crucial for definitive results, including individual variations in criteria for specifying different states, differences in induction procedures, and semantic-operational differences among hypnotists.
B. Modern Hypnotic Susceptibility Scales Since the beginning of the 20th century, two important predecessors of modem assessment instruments have been developed. Davis and Husband (1931) constructed a scale in which hypnotic suggestions were scored in clustered groups, rather than on an individual basis. Subjects were classified according to a five-point scale, with each point associated with a group of items. The scale did not have a standardized induction technique, however, and norms and standards for scoring have never been made adequate. In 1938, Friedlander and Sarbin introduced a four-part scale with a standardized verbatim
HYPNOTIC SUSCEPTIBILITY, EEG-ALPHA, AND SELF-REGULATION
177
induction procedure, objective criteria for scoring each item, and selfscoring by the experimenter. Because of these features, the scale has higher interrater reliability than its predecessors. The development and refinement of the modem hypnotic susceptibility scales used in current research has built upon the progressively more systematic methods of earlier scales. This effort has resulted in a number of scales for different purposes which are thoroughly standardized and easily administered and have good psychometric properties. The bulk of this work has been done at the Laboratory of Hypnosis Research at Stanford University under the direction of Ernest R. Hilgard. The Stanford Hypnotic Susceptibility Scales, Forms A and B (SHSS:A and B) were the original scales, developed by Weitzenhoffer and Hilgard (1959) and are parallel alternate forms for repeated measurements, when necessary. They are partly derived from earlier scales, but with the addition of easier items to normalize the distribution of obtained scores and to avoid the piling up of scores at the lower end of the range. The scales have 12 items, all scored on a passfail basis and all given equal weights. The items included in Forms A and B are described in Table 1. They include a "waking suggestion" for postural sway, before eye closure, in which the hypnotist stands behind the subject and repeatedly suggests that he will sway backward into the hypnotist's arms. This item is described by the experimenter as a nonhypnotic suggestion, to prepare the subject for future suggestions. The hypnotist then gives eye-closure suggestions, followed by "deepening suggestions," intended to increase hypnotic responsiveness once the eyes have closed. The items are presented in a mixed sequence of difficulty, so that the subject does not reach a plateau and meet with repeated failures beyond that point. The scale, which has served as a prototype for others, is constructed like a good aptitude test. In describing the factorial composition of the SHSS:A items, Hilgard (1965) reported that three unrotated factors account for 69% of its total variance. These are (1) the "challenge" or "negative suggestion" items, involving loss of voluntary control over bodily musculature, which accounts for 53% of the variance; (2) the direct suggestion motor items, in which there is a muscular response to direct suggestion (10% of the variance); and (3) the cognitive items, represented
From Weitzenhoffer and Hilgard, 1959.
Arm rigidity Moving hands Verbal inhibition Hallucination Eye catalepsy Posthypnotic suggestion Amnesia
6. 7. 8. 9. 10. 11. 12.
a
Eye closure Hand lowering Arm immobilization Finger lock
2. 3. 4. 5.
1. Postural sway
Item
TABLE
1
Left arm Together Name Fly Both eyes closed Change chairs Recall items 3-11
Backward Form A induction Left Right arm Before chest
FormA
Right arm Apart Home town Mosquito Both eyes closed Rise, stretch Recall items 3-11
Backward Form B induction Right Left arm Over head
Form B
Falls without forcing Eyes close without forcing Lowers at least 6 inches Arm rises < 1 inch in 10 seconds Incomplete separation of fingers after 10 seconds < 2 inches of arm bending in 10 seconds Hands move 6 inches Name unspoken in 10 seconds Behavioral acknowledgement Eyes remain closed 10 seconds Any movement response Recall of 3 or fewer items
Criterion for passing
Items on the Stanford Hypnotic Susceptibility Scales, Forms A and B"
......
~
=::
~
C"l
tn Z
('i
El
t:l
~
"I
HYPNOTIC SUSCEPTffiILITY, EEG-ALPHA, AND SELF-REGULATION
179
by hallucination, amnesia, and posthypnotic suggestion (6% of the variance). The SHSS:A and B have test-retest reliabilities of r = .83 for the original standardization sample (N = 124), and a later sample (N = 96) yielded a retest reliability of r = .90. Reliabilities of the individual items on retest are high, and the scales show a high internal consistency. A third form of the same scales, the Stanford Hypnotic Susceptibility Scale, Form C (Weitzenhoffer and Hilgard, 1962) was developed within the same format to sample a wider variety of hypnotic experiences that SHSS:A and B, particularly of the cognitive type (hallucination, dream, age-regression). The scale is administered in increasing item difficulty and has score distributions, reliabilities, and internal consistency similar to SHSS:A and B. The correlation between SHSS:C and SHSS:A and B is r = .72. Because of their excellent psychometric properties, the format and content of the Stanford scales have been a basis for the development of other useful susceptibility scales. This parallel development has kept data closely comparable. The Harvard Group Scale of Hypnotic Susceptibility (Shor and Orne, 1962) was derived from SHSS:A to measure hypnotic responsiveness in groups. Items are somewhat modified to permit selfscoring in booklets given to subjects so that they can rate their behavior and experiences as an observer would. Group self-scoring norms are very similar to those of individual-observer scoring, although average scores tend to be slightly higher when the subject rates himself (Shor and Orne, 1963). The Childrens Hypnotic Susceptibility Scale (London, 1962) is another scale which closely parallels the content and style of the Stanford scales but was developed to test susceptibility in children. Both qualitative and quantitative observer scoring-systems are used. Two scales that use different formats are the "diagnostic scale" developed by Orne and O'Connell (1967) and the Barber Suggestibility Scale (Barber and Glass, 1962). The "diagnostic scale" utilizes both the subject's observed responses and the experienced hypnotist's qualitative impressions, and instead of a standardized-item format, techniques and items are freely selected which are best tailored to the individual subject. This scale gives the experienced hypnotist freedom to choose the items that might best assess maximum susceptibility in
180
DAVID
R.
ENGSTROM
subjects. The Barber Suggestibility Scale is a scale which presents items without the usual hypnotic induction. The items are presented as tests of imagination, and "task-motivating instructions" are used instead of hypnotic instructions. Norms indicate that average scores for subjects who are administered the Barber scale without prior induction are slightly lower than those for subjects who received prior hypnotic induction (Barber and Calverley, 1963). The development of standardized, face-valid, and reliable hypnotic susceptibility scales has led to the accumulation of distributions of susceptibility, as it exists as an ability in all people. From these basic data, considerable research has been done on the various parameters of hypnotic susceptibility, including situational influences, personality variables, and physiological responses.
III.
STABILITY OF HYPNOTIC SUSCEPTIBILITY
An individual's ability to enter into hypnosis and become involved in its characteristic objective behaviors and subjective experiences under standardized conditions has long been considered a very stable trait or subject variable, incapable of significant modification. An enormous amount of data reveals that the susceptibility scores of individual subjects obtained at two different times under standard conditions are usually highly correlated. Hilgard (1965) has reported correlations, based on retest reliability, in the .80-.90 range, when subjects are retested after several days or weeks. More recently, Morgan, Johnson, and Hilgard (1974) retested 85 subjects on the SHSS:A, following a retest interval of between 8 and 12 years (mean interval = 10.5 years). There were no overall changes in susceptibility scores, and the correlation between total scores of the two testings was r = .60. Interestingly, intratest item changes over time were much more apparent, with motor "challenge" items, such as arm immobilization, arm ridigity, and verbal inhibition, all improving even in the adult years and cognitive-type items, such as amnesia, posthypnotic suggestion, and hallucinations, all decreasing with age. These changes in opposite directions tend to balance out and mask themselves in the stability of the whole test score. Normally, without special intervention, susceptibility scores reflect little change over time.
HYPNOTIC SUSCEPTIBILITY, EEG-ALPHA, AND SELF-REGULATION
IV.
181
MODIFICATION OF HYPNOTIC SUSCEPTIBILITY
Although it is widely believed that hypnotizability is easily modified with practice, this idea probably came about because of the increased speed in which the hypnotic state can be brought on with practice or rehearsal. Especially in the clinical setting, as the subjeCt undergoes repeated induction with a hypnotist, the cuing process for bringing about hypnosis can be reduced to a preselected signal, greatly increasing induction speed, while having no effect on ultimate depth of hypnosis or ability to respond to hypnotic suggestion. Some investigators have attempted to modify it within subjects, typically through simple practice or repeated hypnotic induction. The majority of these attempts have met with little or no success (As, Hilgard, and Weitzenhoffer, 1963; Barber and Calverly, 1966; Cooper, Banford, Schubot, and Tart, 1967; Gill and Brenman, 1959; Sachs and Anderson, 1967; Shor, Orne, and O'Connell, 1962). In one such experiment (As, Hilgard, and Weitzenhoffer, 1963), a wide variety of manipulations were used flexibly, including practice induction techniques and psychotherapeutic attitude discussions, resulting in "only very slight gains" on before-and-after susceptibility measures. Repeated similar results led London (1967) to conclude that "scale results are so consistent from session to session as to suggest that susceptibility is a very stable personality trait, extremely resistant to change" (p. 72). Although there were early reports of success in increasing susceptibility by intensive practice and deepening procedures (Blum, 1963; Wiseman and Reyher, 1962), these results are both sparse and restrictive in the number of subjects used, making them hard to generalize. From this evidence, a conservative conclusion would be that without special intervention susceptibility is fairly stable, and even with intervention it is not very easy to change. Despite these early findings and conclusions, later investigations have demonstrated significant increases in susceptibility-scale scores following diverse training procedures within the experimental situation. Both operant-conditioning techniques (Sachs and Anderson, 1967) and observational-learning procedures (Diamond, 1972) have been used successfully to raise susceptibility scores. In addition to traditional learning procedures, a number of other techniques have
182
DAVID
R.
ENGSTROM
recently been used to increase susceptibility, including sensory deprivation (Sanders and Reyher, 1969; Wickramasekera, 1970), EEG alpha biofeedback training (Engstrom, London, and Hart, 1970), EMG biofeedback training (Wickram, 1972; Engstrom, 1972), and LSD-25 and dextroamphetamine (Ulett, Akpinar, and Hil, 1972). Tart (1970) and Shapiro and Diamond (1972) found that nonlaboratory related "personal growth experience" such as residential programs and encounter groups tend to increase the hypnotizability of participants. Of course, there are large differences among these studies in the selection of subjects, the criterion instruments used to measure hypnotic susceptibility, and the initial level of susceptibility of the subjects. And, most importantly, many of these results have not yet been replicated. When combined, however, these studies do indicate that susceptibility can be modified for at least some more susceptible subjects by assorted experimental techniques, including behavioral shaping, biofeedback, sensory restriction, and pharmacological and other nonspecific methods. The very fact that this ability is not completely fluid and changeable on the one hand and not totally stable on the other supports the assumption that there may be two interacting parts of susceptibility: a state-specific skill component, reflected in the subject's ability to alter his internal state, and a less stable situationalattitudinal component, dictated by the subject's assessment of and response to the hypnotic situation. Trying to tease out responses that are attributable to internal changes is very difficult. The hypnotizable subject's capacity to resist hypnosis is always present, as is the unhypnotizable subject's ability to simulate hypnotic performance. Both conditions can alter obtained scores on susceptibility scales, since all scales rely on objective behavior and subjective report for data.
V.
HYPNOTIC SUSCEPTIBILITY AND PERSONALITY
Currently there is little understanding of individual differences in responsiveness to hypnosis as they relate to measured personality traits. If susceptibility to hypnosis is more than a situational response and is, as many experimental findings imply, a structural trait, then it
HYPNOTIC SUSCEYI'ffiILlTY, EEG-ALPHA, AND SELF-REGULATION
183
should have ongoing correlates in physiological, cognitive, and social response patterns. But if they are there, they are hard to find, and efforts to relate susceptibility to other stable aspects of personality have been extremely unsuccessful. According to Hilgard's review (1967), the search for personality correlates of hypnotizability has produced rather sparse results in light of the considerable amount of effort expanded. Studies relating personality variables to hypnotic susceptibility have generally yielded few and often contradictory results. The variables studied have included sex (Cooper and London, 1966; Hilgard, Weitzenhoffer, and Gough, 1958; Weitzenhoffer and Weitzenhoffer, 1958); intelligence (Hilgard, 1967; Weitzenhoffer, 1953); projective measures of personality (Sarbin and Madow, 1942; Schafer, 1947); objective personality inventories (Faw and Wilcox, 1958; Schulman and London, 1963); attitudes toward hypnosis (Melei and Hilgard, 1964; Rosenhan and Tomkins, 1964); and psychiatric diagnosis (Gill and Brenman, 1959; Kramer and Brennan, 1964; London, Cooper, and Johnson, 1962). These studies have all assessed hypnotizability in a standardized, reliable manner and have all yielded results that are highly equivocal. Either the variable measured was uncorrelated with hypnotic susceptibility at all, or it yielded a significant correlation of low predictive power and with poor results on replication. The relationship between most measured personality variables and susceptibility is, at best, a very subtle one. Still, more is known about the trait dimensions of susceptibility than its state properties.
A. Age and Development On the positive side, a number of studies (Stukat, 1958; Barber and Calverley, 1963; Cooper and London, 1966; London and Cooper, 1969; Morgan and Hilgard, 1973) have found a generally higher level of susceptibility in children than in adults. The average maximum level of susceptibility is reached between ages 9 and 14, with a slow decline to stable adult levels between ages 14 and 16. In a detailed analysis of age differences in susceptibility, Moore and Lauer (1963) found that children tend to pass hallucination and amnesia items much more frequently than motoric items.
184
DAVID
R.
ENGSTROM
Recently Josephine Hilgard (1970) has shown, through very careful and extensive interview methods, that high susceptibility is correlated with certain aspects of childhood development and present personality. The correlation was especially high (in decreasing order of magnitude) for imaginative involvements, severity of punishment in childhood, ease of communication, motivation for hypnosis, and temperament similarity to the opposite-sex parent, as well as normal, outgoing personality attributes. From these results it is clear that hypnotic susceptibility does have longitudinal, ongoing correlates not easily measured by personality tests. Some people, through certain developmental and life experiences, are predisposed to be more hypnotizable than others. Apparently part of the ability to respond to hypnotic operations is age· and experience-related and changes developmentally.
B. Motivation Motivation seems to be another potentially important determinant of hypnotic susceptibility. Shor (1959) observed that low-susceptible subjects volunteered to tolerate higher-intensity levels of shock than did high-susceptibles in the same experiment, and Orne (1963) found that highly susceptible subjects tended not to return to the laboratory for subsequent appointments, while un susceptible subjects were more reliable and punctual. London and his co-workers, using hypnotic susceptibility as the primary criterion of subject selection, unexpectedly found that lowsusceptible subjects tried harder and performed better in a laboratory situation on tasks involving strength, endurance, and motoric stability. Since these were base-rate tasks, administered before the subjects were hypnotized, the results seem to have more to do with differences in motivational set that high- and low-susceptible subjects brought into the experimental situation than with any set elicited by experimental manipulation. Further, a number of different experiments have attempted to examine these differences between high- and lowsusceptible subjects across different parameters of instruction, learning and performance tasks, and experimental conditions (hypnosis,
HYPNOTIC SUSCEPTIBILITY, EEG-ALPHA, AND SELF-REGULATION
185
base rate, exhortative, emotionally involving, and relaxing} on physical performance measures of maximum exertion and motor coordination (London and Fuhrer, 1961; Rosenhan and London, 1963). Lowsusceptible subjects also perform better on rote verbal-learning tasks under all instruction conditions, including relaxation, muscle tension, exhortation, and hypnosis, and in two of four experiments did better under base-rate instructions as well (London and Rochman, 1967; 5chaefler and London, 1968). These same differences between groups hold up under experimental conditions of severe environmental stress (London and McDevitt, 1967) and even when the subjects have no idea of the nature of the experiment or that it is to involve hypnosis (Rosenhan and London, 1963; London and Rochman, 1967). From this series of experiments, it is apparent that low-susceptible subjects "try harder" than high-susceptibles under almost all experimental conditions and that this difference in motivational set is, to some extent, brought into the experimental situation and not just stimulated by it. In fact, Rosenhan and London (1963) have concluded that low susceptibility is a positively motivating trait.
VI.
HYPNOSIS AND THE
EEG
Another set of variables whose relationship to hypnosis has been superficially investigated are the physiological ones, with particular emphasis on events in the central nervous system and specifically the EEG. For many decades behavioral scientists have attempted to find a criterion for hypnosis that is more convincing than the subject's selfreport. In hypnosis the search for a physiological indicator of trance has continued unfailingly. There is inherent complexity in attempts to define any state using the EEG as a criterion measure. For one thing, the brain-wave recordings of subjects who are in light sleep, deep hypnosis, relaxed, or fully aroused look more like than different from each other on examination. These records all typically consist of mixtures of slower, high-amplitude alpha waves (8-13 Hz, 30-90 fLV) and low-voltage fast activity, in apparently random patterns. So there is a problem in the
186
DAVID
R.
ENGSTROM
clear discrimination of state variables based solely upon EEG patterns. Also, as Evans (1972) has pointed out, the alpha rhythm is paradoxical in its appearance. It desynchronizes and disappears in the relaxed, waking subject as he becomes either drowsy and increasingly sleepy on the one hand, or increasingly aroused and preparing for more complex cognition on the other. Evans suggests that alpha is related to attention in a V-shaped manner, in the sense that it disappears at either extreme of arousal and attention. Most of the early investigations relating EEG and hypnosis have not used the sophisticated computerized techniques now available but have relied on visual scanning of analogue records, as well as hand counting and scoring. They have been of a purely correlational nature and have failed to demonstrate EEG differences between waking and hypnotic states (Barber, 1970; True and Stephenson, 1963; Weitzenhoffer, 1953). Early studies in this area strongly suggested that waking EEG patterns persist during passage into and throughout the hypnotic state (Chertok and Kramarz, 1959; Dynes, 1947; Loomis, Harvey, and Hobart, 1936). Similarly Edmonston (1967) reported failure to reliably demonstrate changes in autonomic functioning as a result of hypnosis alone. These results, and others like them, have led Barber (1970) to conclude that "physiological variables vary in hypnotic subjects in the same way as in normal individuals, that is, in accordance with whatever activity they are engaged in" (p. 159). Several studies have, however, shown evidence of a relationship between hypnosis and specific EEG output. Barker and Burgwin (1949) reported the induction of EEG patterns resembling those of sleep by utilization of hypnotic suggestion. Through hypnotic suggestion to relax, Ford and Yeager (1948) found that persistent EEG alpha rhythms could be established in subjects who had previously exhibited little or no alpha. The vast majority of these early studies have tried to identify hypnotic state differences, but little or no attention was paid to more basic individual differences in susceptibility between subjects. Either the investigators were random and unselective in their choice of subjects by susceptibility, or they chose only extremely hypnotizable subjects by nonstandardized measures. Although reliable instruments for the assessment of this trait have existed for many years, most early investigators have failed to score subjects on susceptibility, allowing
HYPNOTIC SUSCEPTIBILITY, EEG-ALPHA, AND SELF-REGULATION
187
no adequate means for evaluating results in terms of individual differences in responsiveness to hypnotic procedures.
VII. EEG
AND HYPNOTIC SUSCEPTIBILITY: INDIRECT
RELATIONSHIPS
Reliable changes in both EEG activity and hypnotic susceptibility have been noted by different investigators, concurrent with both developmental patterns and environmental manipulation. One area includes changes in EEG and susceptibility as a function of age, the other as a result of perceptual or sensory deprivation.
A. Age The relationship between age and hypnotic susceptibility, described earlier in this chapter, has been explored by a number of different investigators (Barber and Calverley, 1963; London, 1965; Moore and Lauer, 1963; Stukat, 1958). Among 240 children, London found marked changes in responsiveness to hypnosis which were agerelated, with the maximum susceptibility scores attained between ages 9 and 14, falling slightly to stable adult levels between 14 and 16 years of age. A similar age-related pattern has been reported for changes in slow-wave EEG activity. Stevens, Sachdev, and Milstein (1968) found that the EEG in infancy and early childhood is slower and higher in amplitude and shows less regional differentiation of wave forms than the EEG of older children and adults. Maturation of the EEG is gradual, and 4-7 Hz theta predominates early from all regions. By mid-childhood (6-8 years old) a strong 8--12 Hz alpha component appears posteriorly, and average frequency increases through the alpha range throughout early adolescence. Figure 1 shows fitted curves representing the developmental patterns of both susceptibility and brain-wave changes for ages 4 through 16. Hypnotic susceptibility scores are combined means reported in two studies using a standardized scale with hypnotic induction procedures (London, 1965; Stukat, 1958). The scores from
188
DAVID
~
til
R.
ENGSTROM
12
12
11
11
--------
~ 10
0
~
10
U
til
9
9
8
8
7
7
e
6
6
~
5
Susceptibility - - - -
5
4
Alpha Rate
4
&l N
I-<
~ ~H
I-< ,..J I-<
'"
H
H
'"'" U
til
=>
til
3
3
2
2
1
1
0
0 4
5
6
8
9
10
11
12
13
14
15
e '"~ :a '"~ ~'" '" ~ N
16
AGE (YEARS) FIGURE 1. Changes in hypnotic susceptibility and average alpha rate with age. (Susceptibility data: London, 1962; Stukat, 1958. Age data: Lindsley and Wicke, 1974)
the different scales used are made comparable by the fitting of a standard 1-12 scale to both. The closeness of fit of these two variables suggests that they may undergo parallel developmental influences which change them in a similar way.
B. Perceptual or Sensory Deprivation The experimental restriction of sensory or perceptual experience has also produced similar effects on brain waves and susceptibility, measured independently, in different situations. The slowing of EEG activity during exposure to deprivation has been reported in numerous experiments (see Zubek, 1973). A decrease in the mean frequency of the occipital alpha rhythm is usually greater with perceptual deprivation than with sensory deprivation (Zubek and Welch, 1963). Perceptual deprivation is a condition characterized by greater percep-
HYPNOTIC SUSCEPTIBILITY, EEG-ALPHA, AND SELF-REGULATION
189
tual-motor and cognitive impairment. Also, the recovery period after deprivation is slow and gradual, with the mean EEG frequency slowly returning to preexperimental levels. The magnitude of this alphafrequency slowing is dependent upon deprivation conditions, but the result is very reliable (Zubek, 1973). Taken together, these findings are not definitive but barely suggestive of a relationship between hypnotic susceptibility and EEG changes. They do suggest a variable that is basic to hypnosis: the restriction of sensory experience. If lower levels of cortical arousal represent a greater predisposition in certain subjects to restrict sensory experience, this should relate to the skills involved in becoming hypnotized. These interrelationships and the speculation generated by them converge on the direct evidence of a relationship between hypnotic susceptibility and the EEG.
VIII.
EEG AND HYPNOTIC SUSCEPTIBILITY: DIRECT EVIDENCE
A. Base-Rate Alpha Density The EEG alpha rhythm is a recurring 8-13 Hz brain-wave pattern associated with a relaxed, waking condition with the eyes closed. The alpha pattern disappears in response to most visual, auditory, and tactile stimuli and is replaced with the higher-frequency fast EEG activity associated with alertness and vigilance. Most of the existing correlational studies of EEG and hypnotic susceptibility have measured the alpha rhythm recorded from a single occipital site. Since Kamiya's (1969) claims of operant control of these brain rhythms, there has been a renewed interest in alpha brain waves and their meaning, and the alpha rhythm has been studied more extensively as a welldefined part of consciousness. London, Hart, and Leibovitz (1968) originally found evidence that highly susceptible subjects produced more waking alpha than nonsusceptibles, in a sample of 125 subjects. Subjects were grouped by score on the HGSHS:A, and subsequently, the number of seconds per minute of alpha were recorded during waking and eyes-closed rest periods. Those subjects scoring between 0 and 4 on susceptibility had a mean alpha density of 37%; those from 4 to 7 in susceptibility, 56%; from 7 to 11, 42%; and 12, 70%. At the extremes, all subjects scoring
190
DAVID
R.
ENGSTROM
12 on the HGSHS:A produced base-rate alpha for an average of 42.3 seconds per minute, while 25 subjects scoring 4 or less produced alpha for 24.0 seconds per minute (p < .005). From an evaluation of his own unpublished data, Evans (1972) compared the resting 2-minute alpha base-rate measures of 139 volunteer subjects with their scores on the HGSHS:A, the SHSS:C, and clinical diagnostic ratings of hypnotizability. Although the three subjects who initially scored 12 on the HGSHS:A generated an average of 90% alpha, 40 low-susceptible subjects (scores 0--4) produced 65%, considerably more than the 37% obtained by London, Hart, and Leibovitz (1968). The addition of the individual Stanford Scale and diagnostic rating also reduced the correlation. Using the same susceptibility categories as London, Hart, and Leibovitz, Evans obtained insignificant correlations (p > 1.0) between all susceptibility measures and frequency, amplitude and density of alpha output. A study by Engstrom, London, and Hart (1970) partially supported the correlation, using a sample of 30 subjects preselected for medium and low hypnotic susceptibility (average score of 7 or less on both HGSHS:A and SHSS:A) and less than 50% alpha recorded during a subsequent 4-minute base-rate period (2 minutes eyes open, 2 minutes eyes closed). For these 30 subjects with low to medium susceptibility and low alpha output, a correlation of .56 (p < .01) between alpha density and susceptibility was obtained. Using similar selection criteria, Evans (1972) obtained a correlation between HGSHS:A and alpha of .26 (p < .05) for a group of 48 subjects low in susceptibility and alpha output, drawn from data of prior experiments, although the correlation between HGSHS:A and alpha for unselected subjects was still nonsignificant (- .02). Nowlis and Rhead (1968) found a correlation of .70 (p < .001) between the sum of HGSHS:A and SHSS:C and resting, eyes-closed alpha for a group of 21 unselected subjects. Evans (1972) has calculated separate correlations according to susceptibility and obtained a correlation of .16 (nonsignificant) for the 12 high-susceptible subjects and .79 (p < .05) for the 9 low-susceptibles. Hartnett, Nowlis, and Svorad (1969) obtained a rank-order correlation of .69 between alpha and SHSS:C for a subgroup of 14 subjects selected from 28 subjects by unreported criteria. For all 28, the correlation was nonsignificant (- .27). Two groups of investigators (Galbraith, London, Leibovitz,
HYPNOTIC SUSCEPTmILITY, EEG-ALPHA, AND SELF-REGULATION
191
Cooper, and Hart, 1970; Ulett, Akpinar, and Hil, 1972) have related hypnotic susceptibility to quantitative digital computer methods of cross-spectral EEG-frequency-analysis data collection. Galbraith et al. gave the HGSHS:A to 80 subjects and 2 weeks later selected 59 of these to participate further in an ostensibly independent "study of brainwaves." Quantitative EEG data were analyzed through stepwise multiple linear regression to isolate the EEG frequencies most closely related to susceptibility, but alpha variables did not significantly predict susceptibility scores. Conversely, Ulett et al. found decreased slow-wave activity (4-7 Hz) and increased alpha and very-highfrequency superimposed beta activity (40-50 Hz) correlated with hypnotizability, defined by the Barber Suggestibility Scale. These results are confusing. They include significant positive correlations and nonsignificant correlations, as well as no correlation at all between alpha and susceptibility. Additionally there are considerable data suggesting a correlation limited to especially selected subgroups of subjects (Hartnett, Nowlis, and Svorad, 1969), particularly to low susceptibles (Engstrom et al., 1970; Evans, 1972; and Nowlis and Rhead, 1968). Evans (1972) has correctly noted that in the only two studies in which subjects were not told of the connection between susceptibility and EEG measures (Evans, 1972; Galbraith et al., 1970), no correlation was obtained.
B. Base-Rate Alpha Amplitude There are some data which relate EEG alpha-amplitude differences to susceptibility. Engstrom (1971) and Evans (1972) have both reported nonsignificant correlations between mean alpha amplitude and score on HGSHS:A (r = .12; r = .04). Engstrom found a similar nonsignificant correlation between amplitude and SHSS:A (.09) administered to the same subjects while Evans further reported no correlation between amplitude and SHSS:C (.03) and a diagnostic rating of hypnotizability (- .02) among the same subjects. In contrast to these findings, Morgan and MacDonald (1973) found that among 26 subjects, high-susceptibles produced higher amplitudes of alpha. The relationship between alpha amplitude and susceptibility is unclear from the present data.
192
DAVID
R.
ENGSTROM
C. EEG Asymmetry The proposition that the two hemispheres of the brain operate somewhat autonomously and perform different functions has resulted in several relevant studies involving EEG differences between hemispheres. Lateral asymmetry is the measured difference in electrical activity between the same part of the two hemispheres of the cerebral cortex, and hemispheric dominance has been shown to be related to subjective reports of thought process and type of task. Morgan, McDonald, and MacDonald (1971) recorded EEG alpha activity bilaterally in a sample of 10 high-susceptible and 10 lowsusceptible subjects under task conditions designed to activate first the left and then the right hemisphere. More alpha was recorded from the right than the left in both conditions, and there was significantly less alpha in the right hemisphere when it was engaged in a task. Morgan and MacDonald (1973) found significant asymmetry in lateral alpha between analytic and spatial tasks, and between eyesopen baseline and eyes-open measurement during hypnotic amnesia, in 26 right-handed subjects, but no differences were found between low- and high-hypnotizables in the percentage difference (laterality) measure. Results suggest a relationship between EEG asymmetry and type of task, but not susceptibility.
D. Evoked Potentials The sudden, high-amplitude spike responses in the EEG tracing that typically follow presentation of an unexpected stimulus, called evoked potentials, have been reported to be alterable by hypnotic suggestion (Clynes, Kohn, and Lifshitz, 1963). These almost reflexive patterns, which apparently reflect the attentional process at some level, have enormous appeal to hypnosis researchers. Unfortunately an abundance of recent evidence (Beck and Barolin, 1965; Beck, Dustman, and Beier, 1966; Halliday and Mason, 1964) has failed to find a significant relation between hypnosis and evoked potentials.
E. Conclusion In reviewing the experimental findings on hypnotic susceptibility and alpha, Evans (1972) has stated that "in spite of conflicting results
HYPNOTIC SusCEPTmILITY, EEG-ALPHA, AND SELF-REGULATION
193
it is concluded that waking alpha frequency, density and amplitude are probably not correlated with susceptibility to hypnosis" (p. 59). The evidence is equivocal and it is not possible to conclude that there is a relationship between hypnotic susceptibility and waking EEG activity among un selected subjects. Nor is it possible to conclude that there is not a relationship. Before a conclusion is possible, the search for a systematic relationship between EEG and either the ability to be hypnotized or the result of hypnotic induction must be extended and evaluated in more detail. First it is imperative to determine the stability of an EEG base-rate measure over time. Physiological base rates can change, sometimes drastically, as a result of a subject's accommodation to the laboratory setting. Evans (1972) found that base-rate differences in skin potential between susceptible and non susceptible subjects which exist in initial sessions disappear with subsequent sessions. This suggests that differences in base rates are an artifact of relaxation, since repeated measures in subsequent sessions weakened the initial group differences.
IX. THE
STABILITY OF
EEG
BASE RATES
To try to relate a fairly stable variable like hypnotic susceptibility to a potentially unstable one may be foolish to begin with. EEG measures in humans, which change from moment to moment, might change so much from session to session that any correlation between them is spurious. Whether base-rate EEG changes as a result of special interventions or life experiences is only a meaningful question after it is established that it does not change on repeated measurement in the same setting, without special interventions. Since repeated measurement in the same setting requires an ongoing baseline measure without experimental intervention, 18 subjects with no prior experience in experiments were brought into the laboratory for "a series of physiological measurements." The base-rate EEG was recorded during eight sessions, within a period of 35 days. All recordings were taken while the subjects were instructed to remain in a relaxed, waking state for a total of 6 minutes, 3 minutes eyes-
194
DAVID
i' ....
::<: ...... u
e J>l
~ ....
'"~ c:l
i:;]
R. ENGSTROM
Eyes Closed - - - -
60f
50 40
30 : 20
10
-----
Eyes Open
--------, "
............
__ _
-------
...----...... "
"
0 1
2
3
4
5
6
7
8
SESSION 2. Mean eyes-closed and eyes-open alpha density over eight base-rate recordings. FIGURE
closed and 3 minutes eyes-open. EEG recordings were taken from a monopolar occipital (02 ) to right mastoid site and recorded on a Grass Model 7 polygraph with output to filter cutoffs at 8 and 13 Hz. Output to an EPUT meter permitted all alpha activity between 8- and 13-Hz frequency and 30- and 70-p,v amplitude to be counted in seconds per minute. The subjects were given instructions to relax and close their eyes for 3 minutes and then remain relaxed with eyes open for 3 minutes. Following these periods, all subjects were instructed to perform a variety of routine card-sorting tasks. Eight repeated baserate measures of alpha output were taken for all subjects and the results are shown in Figure 2. When both eyes-closed and eyes-open averages were examined for all eight sessions and 18 subjects, there were increasing trends in early sessions, which suggest relaxation and accommodation, and a slight decrease in both conditions toward the last session, perhaps caused by drowsiness or boredom. But the deviations from original baseline for both eyes-closed condition (+ 1, - 3) and eyes-open (+ 3, -6) are still small. With an analysis of variance repeated-measures design, none of the differences are significant for eyes closed (F = 0.13, P < .20) or for eyes open (F = 0.11, P < .20). The difference in eyes-closed alpha density from Session 1 to Session 8 was not significant (t = 0.31, P < .10) and was moderately Significant for eyesopen (t = 1.93, P < .05). The overall correlations for all subjects between sessions were R = .61 (eyes-closed) and R = .37 (eyes-open).
HYPNOTIC SUSCEPTffiILITY, EEG-ALPHA, AND SELF-REGULATION
195
These results indicate that base-rate alpha is at least fairly stable, and eyes-closed base-line measures are slightly more stable over sessions than those taken with eyes open. Relaxation, boredom, and other situational consequences of repeated laboratory visits would appear to have a minimal effect on alpha output. This finding is especially important since the waking EEG is such a variable measure when examined from moment to moment. In the same setting, over time, average alpha density appears to be fairly stable.
X. INCREASING SUSCEPTIBILITY BY EEG FEEDBACK In independent investigations, Hart (1967) and Kamiya (1969) found that EEG alpha can be brought under voluntary control by the placement of the subject in an electronic feedback loop in which EEG alpha activity produces an audible tone. The subjects who produced high alpha levels described a state of "passive alertness" and "selective or focused attention," very similar to reports about the hypnotic experience by highly susceptible subjects (Engstrom, 1970). Since high susceptibility may be reflected in long durations of base-rate alpha output and since it is known that alpha output can be increased by feedback methods, it follows that a manipulation which increases alpha production might also serve to raise susceptibility. This was demonstrated experimentally by Engstrom, London, and Hart (1970). In a screening of 180 volunteers subjects of low to moderate susceptibility and low alpha base rate were selected. The HGSHS was administered to them by tape recording, and those subjects who scored 7 or less were asked to return for EEG recordings and individual testing on the SHSS:A. Subjects with an average score on the HGSHS and the SHSS:S of 7 or less and an alpha density of 25% or less in a total of 4 minutes of eyes-open and eyes-closed recordings were retained to participate further. Of the 30 subjects who met these requirements and were selected, 20 were assigned to an experimental condition (contingent feedback training) and 10 to a control condition (pseudofeedback training). An attempt was made to assign subjects to the conditions on a random basis. Subsequent to the publication of Engstrom et al. (1970), it was
196
DAVID
R.
ENGSTROM
found that 4 of the subjects in the experiment had undergone previous EEG biofeedback training and were included in the results reported. 1 For this revision, 1 experimental subject and 3 control subjects who had had previous training were eliminated, leaving 19 experimental and 7 control subjects. The groups were very similar with regard to age and sex. The EEG recording and training apparatus consisted of a Grass Model 7 polygraph; Krone-Hite Model 330-A bandpass filters set at 8 and 13 Hz; an amplitude-sensitive trigger set to close for unattenuated alpha signals; a Berkeley Model 554 EPUT meter for digital timing of alpha density; a Mighty-Light photoelectric strobe activated by unadjusted-amplitude EEG signals; a Heath EUW27 audio oscillator activated by adjusted-amplitude EEG signals; an Ampex SP300 tape recorder; and a Roberts 1670 internally modified tape recorder. Outputs from the trigger activated the EPUT meter, and a signal marker on the polygraph permitted visual checking of the trigger accuracy. Outputs from the trigger and three channels of the polygraph were connected to the Ampex tape recorder. Output from the Roberts tape recorder connected to a trigger input. This system provided tracking feedback via the strobe light and the on-off target feedback via the speaker when the trigger was operated in the feedback mode. The apparatus was located in a polygraph room, which was separate from the subject room except for the feedback light and speaker. A two-way intercom and a two-way tone signal device connected the two rooms. For both screening and training, Grass silver disk electrodes were placed according to the international 10-20 system. Channell (FP2 Lear) was selected for on-line monitoring and training. The frontal electrode site was chosen to set a stringent criterion for the alteration of alpha levels. Once having qualified for the study, each subject received six training sessions spread over 1-2 weeks. Each training session consisted of a 4-minute base-rate EEG measurement period and six blocks of 5-minute feedback periods in which all frontal brain waves produced synchronous flashes, and alpha waves produced the tone as well. The six blocks consisted of alternating 5-minute "free" sequences, in which the subjects were told to experiment with different mental states to see what effects they had, and "test" sequences, in which the subjects were told to try to keep the tone on as much as 1
The recalculation of data was done by Leslie M. Cooper.
197
HYPNOTIC SUSCEPTIBILITY, EEG-ALPHA, AND SELF-REGULATION
possible. During each 30-minute training session, a tape recording was made of the feedback signals. The feedback tone for the control subjects was not their own but was activated for all of them by a prerecorded EEG tape of a single experimental subject, who had previously demonstrated progressive increases in alpha density during feedback. After each control subject finished the experiment, he was informed that he had been in a control group and was told about the nature of the feedback he had received. He was then given the opportunity to receive real feedback in two postexperimental sessions. At the end of each session, the subjects were asked to describe their experiences during alpha production or hypnosis by rating each of 48 adjectives describing different mental states. At the end of the sixth session, a final EEG base-rate measurement was recorded, and the SHSS: B was administered by an experimenter who had not previously tested the subjects. The two groups did not differ significantly on their pretraining susceptibility scores (SHSS:A) nor on their pretraining base-rate alpha density. The experimental group had a mean susceptibility score of 3.16 (SD = 2.27) and the control group, a mean of 4.00 (SD = 2.00). The experimental group had a mean of 27.26 seconds of alpha per 240 seconds (SD = 16.78), and the control group, a mean of 33.00 (SD = 35.33). Neither of these differences in means was statistically significant. Table 2 presents the mean number of seconds of base-rate alpha duration per 240 seconds of both groups for the base-rate measure, the six training sessions, and the administration of the SHSS: B. For the
TABLE 2 Mean Seconds of Alpha per 240 Seconds of EEG: Raw Scores Training sessions Group
Base rate
1
2
27.26 16.78
46.47 35.50
62.26 45.96
33.00 35.53
45.14 26.59
41.14 51.00 48.14 20.76 22.77 23.48
3
4
5
6
Posttraining: SHSS:B
Experimental
X
SD
74.16 71.79 44.73 43.37
Control
X
SD
87.32 104.79 35.92 35.09 57.14 30.69
74.00 46.52
109.95 39.99 79.28 54.71
198
DAVID
R.
ENGSTROM
six training sessions these figures represent the mean alpha output during the initial 4-minute base-rate phase of each training session prior to the feedback training phase. Thus they reflect the effect of the learning from the previous session. It had been anticipated that there would be a positive relation between base-rate alpha output and initial hypnotic susceptibility. The findings clearly supported this expectation. For the experimental group, the correlation between pretraining susceptibility and pretraining base-rate alpha was .80 (p < .001); and for the control group, the correlation was .88 (p < .05). These correlations did not differ significantly from each other. Both the experimental (contingent-feedback) group and the control (pseudofeedback) group increased their base-rate alpha production over the six sessions; however, the experimental group increased significantly more. The experimental mean was 104.79 seconds (SD = 35.92) at the beginning of the sixth training session, while the control mean was only 74.00 seconds (SD = 46.52) at that point (Table 2). This indicates that the training procedure was successful in increasing alpha output. A cursory examination of the means and variances of the two groups over all sessions suggested that the two are related. A more formal analysis revealed that the means and variances correlated .55 (p < .05). Consequently, a square-root transformation (Winer, 1971) was performed to obtain more independent sample variances. The transformed means and standard deviations are presented in Table 3 (from London, Cooper, and Engstrom, 1974). It will be noted that the control-group means for Sessions 4 and 5 are slightly larger than the corresponding means of the experimental group. As would be expected, however, the experimental means for Session 6 are higher than the control means. In all training sessions, moreover, the raw experimental means exceeded the raw control means. An analysis of covariance was performed between the alphaduration measures of the control and experimental groups for Session 6 using the base-rate measure as the covariate. The means for the two groups differed significantly from one another (F = 4.88, dt = 1/23, p < .05). These findings are further supported by the alpha production measured during the administration of the SHSS: B, which occurred
SD
X
Control
SD
X
Experimental
Group
10.3268 5.6471
9.9976 3.4202
Base rate
16.7366 6.0159
16.0839 3.7621
1
22.07?3 4.0437
17.6393 5.0845
2
18.7929 4.1985
19.3613 4.3688
3
20.5162 3.9075
18.1485 5.0984
4
Training sessions
21.1919 1.7465
19.1286 4.4831
5
6
19.5416 3.0782
21.6227 3.8810
TABLE 3 Mean Seconds of Alpha per 240 Seconds of EEG: Transformed Scores
17.0718 5.6838
20.6672 3.9326
SHSS:B
Posttraining:
\0 \0
,.....
Z
(
~
~
~
j
C')
gJ
~
;)
~~
U'l
rl
~
:x
200
DAVID
R.
ENGSTROM
TABLE 4 Means and Standard Deviations of Number of Seconds of Alpha Production per 240 Seconds during Training Phase and Correlation with Base-Rate Alpha Production
Session Statistic
2
3
4
5
6
Experimental group
X SD r with base-rate alpha
X SD r with base-rate alpha .p
1
69.11 29.96 .66"
82.26 39.91 .83"
100.00 39.94 .53"
90.21 46.59 .82"
98.05 42.35 .72"
119.89 40.63 .75"
Control group 79.29 53.89 -.14
128.00 45.21 .59
98.00 47.19 .58
11D.43 40.82 .52
115.71 18.69 .48
97.00 32.07 .33
< .05.
after the sixth session. The raw mean number of seconds of alpha per 240 seconds was 109.95 (SD = 39.99) for the experimental group, while for the control group it was only 79.28 (SD = 54.71). An analysis of covariance was performed between the corresponding transformed means of the alpha measures (obtained during the administration of the SHSS: B) by use of the base-rate measure as the covariate. These differed significantly from one another (F = 4.45, df = 1/23, P < .05). A more striking perspective of the differential effects of the training procedures on the two groups can be noted in the alpha production during the actual training itself, rather than during the 4minute base rate of each session. The latter was thought to be the more conservative measure of the effects of feedback and, as indicated, did yield different results for the experimental and control groups. Table 4 presents the mean alpha production per 240 seconds of both groups during the actual feedback and the correlation for each group between its feedback alpha production and its base-rate alpha production in each session. There is a fairly steady monotonic increase in the mean alpha production of the experimental group over the six sessions, but there are large and erratic variations for the control group. The variations of the control group are certainly not typical of
HYPNOTIC SUSCEPTffiILITY, EEG-ALPHA, AND SELF-REGULATION
201
what we would expect in a learning situation. Furthermore, large correlations between training and base-rate alpha production over the six sessions are all significant for the experimental group, but not one of the correlations for the control group differs significantly from zero. A significant increase in hypnotic susceptibility as measured by the SHSS:B was found for the experimental group only. Its mean score increased from the pretraining score of 3.16 (SO = 2.27) to 7.42 (SO = 2.69) (t = 6.67, df = 17, P < .01). For the control group, the score increased from 4.00 (SO = 2.00) to 5.14 (SO = 3.10), an insignificant difference. An analysis of covariance of the posttraining susceptibility scores between the two groups, with the pretraining score used as the covariate, yielded a significant difference (F = 4.97, df = 1123, P < .05). Thus, the gain for the experimental group was significantly greater than the gain for the control group. The most important finding of the original study by Engstrom et al. (1970) was that the experimental group rose in hypnotic susceptibility significantly more than the control group did as a result of alpha feedback training. The present analysis reconfirms that finding, even after the contaminating influence of four subjects who had had previous experience with alpha training is removed. In comparison with their base-rate performances, the subjects in the experimental group improved significantly more, in both alpha production and susceptibility, that did the subjects in the control group, though the latter also improved significantly in alpha production. The present analysis produced two other important findings. First, the higher rate of alpha production obtained at the beginning of the sixth training session persisted through the administration of the SHSS: B at the end of that session for both contingent and pseudofeedback groups. Second, alpha production during the training phase of each session correlated significantly with pretraining base-rate alpha for the experimental (contingent) group but not for the control (Pseudofeedback) group. The fact that alpha output during the feedback part of each training session correlates significantly with pretraining base-rate alpha production for the experimental group but not for the controls is important because it indicates that learning occurred more consistently in the former than in the latter group. Some learning did take place among the pseudofeedback subjects as well, though, because their overall alpha production did increase significantly from the
202
DAVID
R.
ENGSTROM
beginning to the end of the experiment. The lack of correlation between alpha production during training and in the pretraining base rate indicates that this learning must have been far more sporadic than in the contingent feedback group. Although something changed systematically in the experimental group and unsystematically in the control, these results may not be direct effects of alpha training. Perhaps relaxation or accommodation to the experimental procedures was indirectly facilitated by real feedback and/or inhibited by false feedback. The significant increase in alpha production for the control group is of interest and requires comment. This increase may have been due to the fact that some portion of the feedback for this group may have occurred when alpha was actually being produced. The feedback mechanism for this group consisted of the actual alpha record of a subject in the experimental group. Whenever he received feedback, as timed from the beginning of the session, the control subject received feedback. Some portion of the feedback tone or flashing light probably coincided with the production of alpha by the control subject, although the exact amount is unknown. To assure that the control subjects were receiving false feedback at all times would require crossfiltering of alpha production of the control subjects in such a way that the tone would have sounded only when there was no alpha being produced. This was not done in the present study. A future study might actually measure different degrees of pseudofeedback and test for differential effects on alpha production. Paskewitz and Orne (1973) have presented data that support the idea that alpha production is largely under situational control. They argue that attempts to increase alpha by feedback often mask the impact on the subject of the experimental situation and do not take into account the initial suppression of alpha to levels below base rate. Instead of learning to increase alpha beyond base rate during subsequent feedback training, the subject is simply returning to baseline from an initially suppressed level. So, according to Paskewitz and Orne, observed alpha increases are more attributable to the subject's acquisition of the ability to disregard the situational stimuli which depress alpha below base rate than they are to significant increases over a relaxed base-rate measure. This argument is one of perspective, and it makes the increases observed in the control group more understandable. In any case, the experimental group increased in
HYPNOTIC SUSCEPTIBILITY, EEG-ALPHA, AND SELF-REGULATION
203
base-rate alpha production to a significantly greater degree than the control group in the present study.
XI.
CHANGES IN
EEG
DURING HYPNOSIS
In a recent review of this area, Evans (1972) noted that most studies trying to find EEG changes concurrent with hypnosis have been anecdotal. He concluded that a definitive study should demonstrate relevant EEG alpha changes in hypnotized subjects, without similar changes in unhypnotizable subjects appropriately motivated to experience hypnosis as well as they can (simulators). Evans described preliminary data on separate groups of 12 hypnotized and 12 simulating subjects tested by an experimenter who was "blind" as to which subjects were "real" and which were simulators. Alpha frequency, amplitude, and density (percentage) were recorded for all subjects, with eyes closed, during normal waking and deeply hypnotized (or simulating) conditions, administered identically to each subject by tape recording. These were scored blind. Simulating subjects produced significantly more alpha than hypnotized subjects in both waking (52% versus 27%) and hypnosis (or simulating) (45% versus 17%) conditions (p < .001). All subjects produced a smaller percentage of alpha density during the hypnosis (or simulating) condition than in the waking condition (p < .01). These are both striking results. The first finding is puzzling, especially in view of repeated results from different laboratories indicating an increase in alpha density during hypnosis, especially among subjects who are highly susceptible (Engstrom, 1970; Morgan and MacDonald, 1973; Ulett, Akpinar, and ltiI, 1972). One investigation (Ulett et al.) further reported an increase in amplitude of alpha during trance induction and hypnosis. But in all cases, no comparison was made with physiological changes in either simulating subjects or highly motivated unsusceptible subjects. Evans's second finding, that Simulating subjects produced more alpha than those in the hypnosis group, is difficult to interpret as it stands. Hypnotic susceptibility scores are not reported for either group, nor is there an indication if these data were collected. As a result, the effect of the real versus the simulator conditions is unclear, since base-line differences between the two groups are not fully
204
DAVID
R.
ENGSTROM
known. The importance of this preliminary finding depends largely on the differences in hypnotic susceptibility between groups. Evans's study does raise a fundamental question implicit in any physiological investigation of hypnosis. It has been shown by Orne (1959) that unhypnotized subjects who are urged to simulate the role of a hypnotized subject in an experiment exhibit many behaviors previously attributable to hypnosis. These behaviors are often a result of the exigencies of the experimental situation. The same question may be asked about the EEG of hypnotized subjects. As an alternative method, extremely high- and low-susceptible subjects might be compared after the unsusceptibles learn to play the role of hypnotized subjects. Diamond (1972) has recently demonstrated that susceptibility can be significantly increased in unsusceptible subjects through exposure to observationally presented information about hypnosis. Furthermore, there appear to be certain kinds of verbal information that are more effective than others, and in fact the presentation of some kinds of behavioral information about hypnosis leads to slight decreases in susceptibility. The investigation suggests an informational, primarily verbal technique for reliably enhancing the susceptibility scores of previously unsusceptible subjects. If susceptibility is increased by observationally presented verbal information, is this increase accompanied by physiological changes such as increased alpha density or reduced muscle tension? Or are the effects of attitudinal and informational manipulation physiologically invisible? In order to maximize the disparity in hypnotic susceptibility and to expose the effects of experimental conditions on the EEG of highand low-susceptible subjects, a study was devised (Engstrom, 1973a) to teach low-susceptible subjects to respond to hypnosis as highsusceptible subjects do by specialized practice and to compare their subsequent EEG alpha output during hypnotic induction with that of initially high-susceptible subjects. From a group of 67 students 30 volunteer subjects were selected who were administered a tape-recorded version of the HGSHS. The subjects represented the extremes of the distribution of obtained susceptibility scores with one group of 10 scoring 9 and above, and two groups of 10 scoring 3 and below on the Harvard scale. All subjects were later administered the SHSS:A individually by a blind
HYPNOTIC SUSCEPTffiILITY, EEG-ALPHA, AND SELF-REGULATION
205
experimenter. The means of the combined scale scores were 10.7, 2.6, and 2.3 for each group of 10 subjects. Each of the 30 subjects was scheduled for two laboratory sessions, spaced 1-4 days apart. At the first session, on the day after individual susceptibility screening, base-rate EEG and EMG measures were taken for each subject. Both EEG and EMG signals were input to a Grass Model 7 polygraph. Amplified output was displayed on a strip-chart recorder for on-line monitoring, and amplified EEG and EMG signals were put into two EPUT meters which were set to indicate the percentage of time of brain-wave activity in the alpha range and the EMG activity in microvolts, peak-to-peak, for each 5-minute trial. The subjects were told to relax, sit still, and close their eyes, while 5 minutes of occipital (02 ) EEG and frontalis EMG records were taken simultaneously. On the second visit each subject was told that he would be given approximately 45 minutes to become acquainted with the laboratory surroundings. During this time 10 unsusceptible subjects were each exposed to 45 minutes of observational information and coaching in hypnosis. Two identical 45-minute presentations were videotaped with a male and a female model presented to same-sex subjects. The information presented was divided into the following general categories: (1) verbal modeling cues, consisting of both disinhibitory information and facilitative information, 18 minutes (adapted from Diamond, 1972); (2) motivational information, 7 minutes; (3) coaching in the role skills involved in hypnosis and the behavior expected from a hypnotized subject, 20 minutes. The other 20 subjects (10 susceptible and 10 un susceptible) were left alone to read magazines or study in the laboratory during the time. Although a monitor was present in the room during the time, the subjects were instructed not to interact with him. At the end of this time EEG and EMG electrodes were again attached to each subject and a blind experimenter administered a standardized hypnotic trance induction adapted from the SHSS:A, with additional deepening instructions. During this 20-minute induction, 5-minute EEG and EMG records were taken at 5- and IS-minute intervals. After trance induction and a 5-minute rest period, all subjects were given the SHSS: B. The results are presented in Table 5, Base-rate alpha was significantly higher for the susceptible group than for either group of unsusceptibles (p < .01), duplicating earlier findings that base-rate
TABLE 5
Susceptible (No information) (N = 10) Unsusceptible (No information) (N = 10) Unsusceptible (Information) (N = 10)
Group 9.8 2.8 5.7
2.6
2.3
SHSS, B
10.7
Mean pretreatment susceptibility (HGSHS+ SHSS:A)
+3.4
+0.2
-0.9
Change
.01
ns
ns
Significance of change (twotailed)
24
28
39
Base-rate alpha (%)
28
23
47
(%)
Alpha during hypnosis
ns
ns
-5 4
.05 8
Significance of change (twotailed) Change (%)
Changes in Mean Susceptibility Scores before and after Treatment and in Alpha Density prior to and during Hypnotic Induction
~
~
ttl
~
6 ?O
~
N
HYPNOTIC SUSCEPTIBILITY, EEG-ALPHA, AND SELF-REGULATION
207
alpha and hypnotic susceptibility are related. The unsusceptible group exposed to observational information was the only group to increase hypnotic susceptibility scores significantly (p < .01), supporting the results obtained by Diamond. This group did not significantly increase alpha output during this time (p < 1), however, and only the susceptible group increased significantly in alpha density during trance induction testing (p < .05). This finding supports the observations of Ulett et al. (1972) that, at least among susceptible subjects, there is a significant increase in alpha density during hypnotic trance induction. Base-rate EMG was not found to be significantly related to hypnotic susceptibility or alpha. Both the no-treatment un susceptible and the susceptible control groups maintained fairly stable susceptibility scores, but the susceptible group showed significant increases in alpha during trance testing. The unsusceptible controls showed no change in either alpha or EMG. These results suggest that while observationally presented information can significantly increase susceptibility, there are at least some physiological differences during trance induction between subjects who are initially susceptible and those whose susceptibility is increased by the learning of the overt behaviors associated with hypnosis. As a result of situational demands and external activities, physiological functions may vary in subjects who are simulating hypnosis. Playing the role of a hypnotized subject may simply produce changes as a result of relaxation or other intervening variables. These results must be cautiously interpreted, but they at least suggest the possibility that there can be physiological differences among highly hypnotizable subjects.
XII.
TASK-SPECIFIC
EEG
CHANGES
Almost all of the previous studies have measured the EEG of subjects either in a relaxed waking state or while performing hypnotic behaviors. An alternative way of exploring susceptibility, frequently used in studies of the relationship of susceptibility to performance and personality measures, is to examine EEG differences among high- and
208
DAVID
R.
ENGSTROM
low-susceptible subjects in nonhypnotic tasks as well as those involving hypnosis. Morgan and MacDonald (1973) recorded occipital alpha bilaterally in 26 subjects during analytic (verbal and numerical), spatial (imagery), and music tasks, as well as during SHSS:A administration. Although no lateral differences were found, highly hypnotizable subjects produced more alpha activity, both outside of hypnosis and within it, except during eyes-open base-line and eyes-open amnesia conditions. The authors concluded that the fact that high-susceptible subjects showed more alpha density and amplitude suggests that the overall production of alpha may be positively related to the particular cognitive mode that characterizes the subject who is able to experience hypnotic phenomena, since few task-related EEG differences were found. The high-susceptible subjects produced more alpha on most tasks. This finding fits the conceptualization of hypnotic susceptibility as a trait and suggests, as have other observed base-rate correlations, that alpha output covaries with the trait of susceptibility. In a recent study in our own laboratory (Engstrom, 1973b), alpha output changes as a function of task, operation, and hypnotizability were observed. Specifically, a comparison was made between alpha output during biofeedback and hypnotic alteration of peripheral skin temperature. Body-temperature regulation is one of the most stable, predictable, and automatic physiological functions in humans. Normally the only circumstances under which it is not are when external environmental conditions are extreme or when pathological internal conditions are present, and in both cases the changes are involuntary. Recently, however, several studies have demonstrated that this basic regulatory body function can be brought under voluntary control by two techniques that are vastly different from an operational point of view: hypnosis and biofeedback. The hypnosis literature includes several informal accounts of skintemperature alterations in one or a few subjects either assumed or shown to be highly susceptible to hypnosis, but these are reported anecdotally and lack any control subjects. One recent exception is a study by Maslach, Marshall, and Zimbardo (1972), in which a group of three subjects trained extensively in hypnosis were successful in changing the temperature of their two hands in opposite directions,
HYPNOTIC SUSCEPTmILITY, EEG-ALPHA, AND SELF-REGULATION
209
while a group of six waking control subjects with no hypnosis training were not, even when all the subjects received the same motivating instructions and verbal suggestions. In spite of the small number of subjects in the hypnosis group for definitive statistical purposes, the difference between bilateral temperature changes of hypnotic and control subjects is highly significant (p < .001). No measure of hypnotic susceptibility was reported, and it would seem crucial to know whether the subjects in each group were equivalent in their ability to be hypnotized. Similar results were obtained by Roberts, Kewman, and MacDonald (1972) using 5 and 9 hours of both hypnosis and skintemperature feedback to train six subjects to raise the temperature of one hand relative to the other. All the subjects attained significant skin-temperature changes in the appropriate direction. Although the authors of both of these studies parenthetically noted that the feedback seemed either to inhibit or to have no effect on performance when given alone, there are several examples of the successful application of skin-temperature feedback in the literature. Studies by Green, Green, and Walters (1970), as well as Taub and Emurian (1972), describe the use of feedback to establish rapid operant control over hand temperature, typically of the same differential magnitude as the hypnosis studies. Feedback was provided by an electronic differential thermometer that represented the temperature differential between any two sites on the surface of the skin. A dial, centered at zero, moved in one direction when one site was warmer than the other and reversed direction when it was cooler. None of these hypnosis or biofeedback studies present data on the long-term retention of skin-temperature control, and nothing is reported about the subsequent performance of subjects when the external supports of trance induction and/or feedback information are removed. In each of these studies the subjective strategies of subjects were idiosyncratic and diverse. Subjects used unilateral warming, unilateral cooling, holding one hand constant while changing the other, or changing the temperature of both hands. To do this, they reported using realistic imagery, symbolic imagery, and imageless commands. Clearly there was no meaningful pattern of cognitive strategy used by successful subjects. The same is not true for the context of successful
210
DAVID
R.
ENGSTROM
control. The only similarity in set among the reports of successful subjects was the concept of "passive volition." Subjects frequently found that "trying too hard" proved futile in learning the task. So peripheral skin temperature is an objective, easily measured variable that it is possible to learn by several superficially different operations. Its relevance to the EEG and susceptibility is derived from the similarity of reports of subjects learning to control it. The phenomenon of "passive volition" described by Green et al. (1970) bears a striking subjective similarity to the observations of Kamiya (1969) and Engstrom (1970) of subjects trying to increase alpha output. In both cases, when the subjects "tried too hard" they failed to increase the desired response. And it was only after the subject reported increases in subjective states called "passive," "relaxed," and "not trying" that any progress was made. Since alpha and skin-temperature changes apparantly must be brought under voluntary control within the same context, or learning set, perhaps they are related in some basic way. Our study sought to compare the separate effects of the acquisition of bilateral handtemperature control by biofeedback and hypnosis training among high- and low-susceptible subjects on EEG alpha output. Of 102 student volunteers who were initially given the HGSHS via tape recording, 42 who scored 10 or more or 3 or less were individually administered the SHSS: C; 12 subjects who scored 10 or more on both scales were randomly split into two groups of 6, and 12 subjects scoring 3 or less on both scales were similarly divided into two groups. One group of 6 high-susceptibles and one group of 6 lowsusceptibles then received 11/2 hours of hypnosis training in a group format. The training consisted of standardized deepening instructions, verbal suggestions stressing the subjects' ability to achieve selfhypnosis independently, and several depth-criterion tests. The other two groups of 6 subjects spent the same amount of time in a group in which they received general information about biofeedback, individualized demonstration on a device irrelevant to the experimental task (EMG biofeedback), and verbal information about the potential ability to transfer skills learned via biofeedback to real-life situations. Every effort was made to present parallel motivational sets and information about hypnosis and biofeedback to the four groups, without specific reference to the task to be subsequently performed.
HYPNOTIC SUSCEPTmILITY, EEG-ALPHA, AND SELF-REGULATION
211
All the subjects were individually tested in a temperature-controlled room, relaxed in a comfortable chair with feet elevated. A thermistor was taped to the center palm of each hand, and bilateral occipital EEG leads were connected at 0 1 and O2 , Temperature information was input to a digital feedback thermometer, and EEG leads were input to a modified filtered dual trigger which provided information on average alpha density from both leads to an EPUT counter. All subjects were told to relax with their eyes closed while a 2minute base-rate recording of peripheral skin temperature from both hands and averaged bilateral occipital EEG measures were taken. The task was explained to all the subjects in the same way. They were told to try to increase the temperature of the right hand and decrease the temperature of the left hand at the same time. They were further told to remain still and relaxed and to keep their eyes colsed, with palms facing upward. The test procedure consisted of three lO-minute trials, each ending with a 2-minute test. All the subjects were told to try to maximize the temperature differential between the hands as much as possible during the 2-minute tests. The subjects in the hypnosis groups were given a standardized 8minute hypnotic induction procedure before the trials, while the subjects in the biofeedback groups were not. The subjects in the biofeedback groups got two types of feedback: continual auditory feedback from a speaker that changed pitch with changes in temperature, and verbal feedback every 15 seconds about the disparity of the temperature of their hands in degrees Fahrenheit. The hypnosis subjects got no feedback during the test sequence. During the 2minute test at the end of each trial, temperature and EEG data were recorded. All 24 subjects were rescheduled for return visits at 2- and 8week intervals, and on retest all the subjects were given the same task and test sequence. The only difference was that no hypnotic induction was administered to the hypnotic subjects and no feedback given to the other groups-the task was presented in a straightforward manner. Mean temperature and EEG changes are summarized in Table 6. Both the high-susceptible biofeedback and hypnosis groups and the low-susceptible biofeedback group learned the task, and in fact all of the subjects in both biofeedback groups and the high-susceptible
.01 .01 .005
31 30 32
60
76 71
29 46
39
4.2
6.3
6.0
.05
16
47
31
Unsusceptible (Hypnosis) Unsusceptible (Biofeedback) Susceptible (Hypnosis) Susceptible (Biofeedback)
1.4
Significance of change (one-tailed)
Change (%)
Alpha during test period (%)
(F')
Group
Base-rate alpha (%)
Temperature difference during test
TABLE 6 Mean Maximum Bilateral Temperature Differences and Mean Changes in EEG Alpha Density during Test Periods for All Subjects and Conditions
~
a::
~
~
m
ro
s
o
tv
tv ......
HYPNOTIC SUSCEPTIBILITY, EEG-ALPHA, AND SELF-REGULATION
213
subjects in the hypnosis group demonstrated the ability to produce bilateral changes in skin temperature. Mean temperature differences between base rate and test scores were significant for three groups (susceptible, hypnosis t = 6.17, P < .005; susceptible, biofeedback t = 5.91, P < .01; unsusceptible, biofeedback t = 4.79, P < .05) but nonsignificant for the unsusceptible hypnosis group t = 1.99, P > .10). Within this last group, 3 of the 6 subjects were able to control temperature significantly. Alpha output increased from base line during task performance for all four groups (susceptible, hypnosis t = 7.06,p < .01; susceptible, biofeedback t = 7.19, P < .005; unsusceptible, biofeedback t = 6.93, P < .01; unsusceptible, hypnosis t = 3.87, P < .05), but the unsusceptible hypnosis group was significantly lower in alpha change than the other three groups (p < .01). The mean algebraic sum of temperature differences and alpha density changes for all groups are shown in Figure 3. The pattern of learning the temperature task was somewhat different for hypnotic subjects, particularly high-susceptibles, whose performance across the test sequences was characterized by rapid increases followed by stable levels of response. The subjects in the biofeedback groups, by contrast, learned the response gradually, with slow, steady progress. The only group to show no appreciable change was the unsusceptible hypnosis group. The mean change in alpha production for the four groups was striking. Alpha output showed an overall correlation of .72 (p < .005) with temperature changes for all test periods and groups. The unsusceptible hypnosis group did not significantly change temperature and showed lower overall changes in alpha density. Even the three unsusceptible hypnosis subjects who were able to do the task to a limited extent did not significantly change alpha output. After 2 weeks, retention of temperature control was just above 70% for the hypnosis groups combined but only 61 % for the biofeedback groups. And after 8 weeks, the hypnotic subjects could still produce nearly 70% of the original rate attained after hypnotic induction, but the biofeedback groups dropped to 56%. In all, the hypnotic subjects retained the original performance levels better than the biofeedback subjects. In this study, highly susceptible subjects produced more alpha when performing the same kind of task, regardless of whether they
214
DAVID
7
R.
100
ENGSTROM
100
"
75
,,
50
I
I
I
1\,
,
,
,
50
~
H Ul
ffi
""
--------
25
25
1
1
a -+--------10 a 1
2
TEST PERIOD
b
3
7
a
a 1
100
2
3
TEST PERIOD
7
100
75
,/
,I
I
~
I
50
...... ...
I I
25
1
I
H
50
~
""
25
1
0-11--------------10
O~----------------~ c
I
I
I
1
2
3
TEST PERIOD
d
1
2
3
TEST PERIOD FIGURE 3. Mean bilateral skin temperature and alpha density scores for all subjects over three test periods: (a) unsusceptible/hypnosis group, (b) Susceptible/hypnosis group, (c) unsusceptible/biofeedback group, (d) susceptible/biofeedback group. - - Temperature, - - - - alpha density.
HYPNOTIC SUSCEPTIBILITY, EEG-ALPHA, AND SELF-REGULATION
215
were hypnotized or not, from an operational point of view. One explanation for these results is that what the subjects were doing in both hypnosis and biofeedback conditions was identical. In other words, although they were operationally different means to the same end, the end was still the same. All the successful subjects learned to assume a particular cognitive set-one that focused attention on the specific body response and eliminated distracting stimuli. Now if the nature of the task is a more important variable than susceptibility, hypnotic susceptibility should not be critical. The subjects should all be able to perform it and should produce more alpha when performing it, regardless of hypnotizability. These data suggest, first, that among high-susceptible subjects, biofeedback and hypnosis can be used to alter skin temperature with about the same efficiency and that hypnotic training seems to lead to a more permanent skill in this regard. But the more important implication is that there appears to be a similar cognitive set, reflected in alpha output, that accompanies successful performance of this task, which seems to transcend the apparent operational differences between hypnosis and biofeedback. This set seems much more directly related to hypnotizability and the nature of the task and may suggest a more general state related to it.
XIII.
CONCLUSIONS
The origins of the EEG in terms of its neural mechanisms or behavioral determinants are still unknown. And whether increased EEG alpha output represents a controlled increase over base rate or, conversely, a return to base rate following its initial suppression is an equivocal issue. What is clear, however, is that alpha output in a standardized situation is a variable ability among subjects, and whatever the attentional skills required to produce high alpha rates, some subjects have them and some do not. Enough evidence has already accumulated to support a moderate relationship between alpha output and hypnotic susceptibility. From the present evidence it is further apparent that alpha density is fairly stable over time and that its sustained increase, by different methods, can result in increased susceptibility scores. Also some means of increasing measured hypnotic susceptibility, such as role-skills learning, can increase behavioral responsiveness to hypnosis without
216
DAVID
R.
ENGSTROM
increasing alpha density. If this finding is confirmed, it implies that there are several variables within susceptibility, that the EEG is critical in their separation, and that role theory alone cannot explain hypnosis. This finding magnifies some of the definitional differences referred to earlier. Hypnosis is both a state and an operation, and the state cha:racteristics are partially define by the EEG. Researchers and reviewers in the area are often too quick to ignore the physiological study of hypnosis, citing shaky and often contradictory experimental results. But the results are really very decent, in light of the sparse and crude effort expended. With more specificity the EEG could become a refined criterion variable, especially valuable to the study of hypnosis and other self-regulatory procedures. There are many fruitful areas still largely unexplored. For normative purposes, alpha base-rates might be compared between susceptible and un susceptible subjects during specific task performance, instead of in the relaxed, waking state, which often makes them too nosey anyway. Also very few data are now available for item analysis regarding EEG changes during the administration of hypnotic susceptibility scales. More specificity is also possible through the laterality measure of the EEG. More attention should be paid to the relationship between laterality and susceptibility and other personality variables. Although many recent studies demonstrate that hypnotic susceptibility can be increased, the areas of sensory deprivation and drug effects on both susceptibility and the EEG are still largely unexplored. In these cases physiological changes may accompany psychological ones, leading to greater understanding of how the two interrelate. And much needs to be added to knowledge of the relationships between different operations, such as hypnotic induction, verbal motivational procedures, and biofeedback, with regard to task performance and EEG criteria. Only by these increasingly specific tests will the effects of hypnosis become clear. Hypnosis is one operation that permits muscular relaxation, enhancement of concentration of attention on the relevant task, and the complementary removal of distracting stimuli from attention. Highly susceptible subjects are better at performing some tasks that
HYPNOTIC SUSCEPTffiILITY, EEG-ALPHA, AND SELF-REGULATION
217
require this ability following different operations including but not limited to hypnosis. In this sense, biofeedback, hypnosis, meditation, and other training operations which enhance these abilities should have a similar effect on highly susceptible subjects, reflected in the EEG. They all permit control of behavioral and physiological processes by altering consciousness in such a way that verbal and symbolic modalities can be transformed into specific responses. Hypnosis is just one operation by which an individual can change his perception of his own control of internal and external responses, or his attributional system. And susceptibility to hypnosis is an important variable in determining the extent of this change.
REFERENCES As, A., HILGARD, E. R., AND WEITZENHOFFER, A. M. An attempt at experimental modification of hypnotizability through repeated individualized hypnotic experience. Scandinavian Journal of Psychology, 1963,4, 81-89. BARBER, T. X. Hypnosis: A scientific approach. New York: Van Nostrand, 1969. BARBER, T. X. LSD, marihuana, yoga, and hypnosis. Chicago: Aldine, 1970. BARBER, T. X., AND CALVERLEY, D. S. Hypnotic-like suggestibility in children and adults. Journal of Abnormal and Social Psychology, 1963,66, 589-97. BARBER, T. X., AND CALVERLEY, D. S. Toward a theory of hypnotic behavior: Experimental evaluation of Hull's postulate that hypnotic susceptibility is a habit phenomenon. Journal of Personality, 1966,34, 416--433. BARBER, T. X. AND GLASS, L. B. Significant factors in hypnotic behavior. Journal of Abnormal and Social Psychology, 1962,64, 222-228. BARKER, W., AND BURGWIN, S. Brain wave patterns during hypnosis, hypnotic sleep and normal sleep. Archives of Neurology and Psychiatry, 1949,62, 412-420. BECK, E. c., AND BAROLIN, G. S. Effect of hypnotic suggestions on evoked potentials. Journal of Nervous and Mental Disease, 1965,140, 154-16l. BECK, E. c., DUSTMAN, R. E., AND BEIER, E. G. Hypnotic suggestions and visually evoked potentials. Electroencephalography and Clinical Neurophysiology, 1966, 20, 397-400. BERNHEIM, H. Hypnosis and suggestion in psychotherapy. Reprinted, New Hyde Park, N.Y.: University Books (1888), 1964. BLUM, G. S. Programming people to simulate machines. In S. S. TOMKINS AND S. MESSICK (Eds.), Computer simulation of pt!rsonality. New York: Wiley, 1963. BRAID, J. Neurypnology: Or the rationale of nervous sleep considered in relation to animal magnetism. London: Churchill, 1843. CHARCOT, J. M. Physiologie pathologique: Sur les divers etats nerveux determines pour l'hypnotization chez les hysteriques. CR Academy of Science, 1882, 94, 403-405. CHERTOK, L., AND KRAMARZ, P. Hypnosis, sleep and electroencephalography. Journal of Nervous and Mental Disease, 1959, 128, 227-238.
218
DAVID
R.
ENGSTROM
CLYNES, M., KOHN, M., AND LIFSHnz, K. Dynamics and spatial behavior of light evoked potentials, their modification under hypnosis, and on-line correlation in relation to rhythmic components. Annals of the New York Academy of Science, 1963, 112, 468509. COOPER, L. M., BANDORD, S. A., SCHUBOT, E., AND TART, C. T. A further attempt to modify hypnotic susceptibility through repeated individualized experience. Journal of Clinical and Experimental Hypnosis, 1967,15, 118-124. COOPER, L. M., AND LONDON, P. Sex and hypnotic susceptibility in children. Journal of Clinical and Experimental Hypnosis, 1966, 14, 55-60. DAVIS, L. W., AND HUSBAND, R W. A study of hypnotic susceptibility in relation to personality traits. Journal of Abnormal and Social Psychology, 1931,26, 175-182. DIAMOND, M. J. The use of observationally presented information to modify hypnotic susceptibility. Journal of Abnormal Psychology, 1972,79, 174-180. DYNES, J. B. Objective method for distinguishing sleep from the head hypnotic trance. Archives of Psychiatry and Neurology, 1947,57, 84-93. EDMONSTON, W. E. Stimulus-response theory of hypnosis. In J. E. GORDON (Ed.), Handbook of clinical and experimental hypnosis. New York: Macmillan, 1967. ENGSTROM, D. R. The enhancement of EEG alpha production and its effects on hypnotic susceptibility. Unpublished doctoral dissertation, University of Southern California, 1970. ENGSTROM, D. R. Previous exposure to feedback: A subject variable in psychophysiological hypnosis research. Paper presented at meeting of American Psychological Association, Washington, D.C., September, 1971. ENGSTROM, D. R. Interactional effects of muscle tension and EEG alpha production on hypnotic susceptibility. Paper presented at meeting of American Psychological Association, Honolulu, September, 1972. ENGSTROM, D. R. Effects of observationally presented information on hypnotizability: Physiology of enhanced susceptibility. Proceedings, Annual Convention, American Psychological Association, 1973a. ENGSTROM, D. R. Task-specific EEG output among highly hypnotizable subjects. Paper presented at meeting of Society Clinical and Experimental Hypnosis, Newport Beach, California, December, 1973b. ENGSTROM, D. R, LONDON, P., AND HART, J. T. Hypnotic susceptibility increased by EEG alpha training. Nature, 1970,227, 1261-1262. EVANS, F. J. Hypnosis and sleep: Techniques for exploring cognitive activity during sleep. In E. FROMM AND R. E. SHOR (Eds.), Hypnosis: Research developments and perspectives. Chicago: Aldine-Atherton, 1972. FAW, V., AND WILCOX, W. W. Personality characteristics of susceptible and unsusceptible hypnotic subjects. Journal of Clinical and Experimental Hypnosis, 1958,6, 83-94. FORD, W. L., AND YEAGER, C. L. Changes in the electroencephalogram in subjects under hypnosis. Diseases of the Nervous System, 1948,9, 190-192. FRIEDLANDER, J. W., AND SARBIN, T. R The depth of hypnosis. Journal of Abnormal and Social Psychology, 1938,33, 453-475. GALBRAITH, G. c., LONDON, P., LEIBOVnz, M. P., COOPER, L. M., AND HART, J. T. EEG and hypnotic susceptibility. Journal of Comparative and Physiological Psychology, 1970, 72, 125-13l. GILL, M. M., AND BRENMAN, M. Hypnosis and related states: Psychoanalytic studies in regression. New York: International Universities Press, 1959. GREEN, E. E., GREEN, A. M., AND WALTERS, E. D. Self-regulation of internal states. In J. ROSE (Ed.), Progress of cybernetics: Proceedings of the International Congress of Cybernetics, London, 1969. London: Gordon and Breach, 1970.
HYPNOTIC SUSCEPTIBILITY, EEG-ALPHA, AND SELF-REGULATION
219
HALLIDAY, A. M., AND MASON, A. A. Cortical evoked potentials during hypnotic anesthesia. Electroencephalography and Clinical Neurophysiology, 1964,16, 314. HART, J. T. Autocontrol of EEG alpha. Paper presented at meeting of the Society for Psychophysiological Research, San Diego, October, 1967. HARTNETT, J., NowLIs, D., AND SVORAD, D. Hypnotic susceptibility and EEG alpha: Three correlations. Hawthorne House Research Memorandum, No. 97, 1969. HILGARD, E. R. Hypnotic susceptibility. New York: Harcourt, Brace and World, 1965. HILGARD, E. R. Individual differences in hypnotizability. In J. E. GORDON (Ed.), Handbook of clinical and experimental hypnosis. New York: Macmillan, 1967. HILGARD, E. R., WEfIZENHOFFER, A. M., AND GOUGH, P. Individual differences in susceptibility to hypnosis. Proceedings of the National Academy of Science, 1958, 44, 1255-1259. HILGARD, E. R., WEIlZENHOFFER, A. M., LANDES, J., AND MOORE, R. K. The distribution of susceptibility to hypnosis in a student population: A study using the Stanford Hypnotic Susceptibility Scale. Psychological Monographs, 1961, 75, 1-22. HILGARD, J. R. Personality and hypnosis: A study of imaginative involvement. Chicago: University of Chicago Press, 1970. KAMIYA, J. Operant control of the EEG alpha rhythm and some of its reported effects on consciousness. In C. T. TART (Ed.), Altered states of consciousness. New York: Wiley, 1969. KRAMER, E., AND BRENNAN, E. P. Hypnotic susceptibility of schizophrenic patients. Journal of Abnormal and Social Psychology, 1964,69, 657-659. LnlBEAULT, A. A. Le sommeil provoque et les etats analogues. Paris: Doin, 1889. LINDSLEY, D. B., AND WICKE, J. D. The electroencephalogram: Autonomous electrical activity in man and animals. In R. F. THOMPSON AND M. M. PATTERSON (Eds.), Bioelectric Recording Techniques-Part B. New York: Academic Press, 1974. LONDON, P. Childrens Hypnotic Susceptibility Scale. Palo Alto, California: Consulting Psycholgists Press, 1962. LONDON, P. Developmental experiments in hypnosis. Journal of Projective Techniques and Personality Assessment, 1965,29, 189-199. LONDON, P. The induction of hypnosis. In J. E. GORDON (Ed.), Handbook of clinical and experimental hypnosis. New York: Macmillan, 1967. LONDON, P., AND COOPER, L. M. Norms of hypnotic susceptibility in children. Developmental Psychology, 1969,1, lU-124. LONDON, P., COOPER, L. M., AND ENGSTROM, D. R. Increasing hypnotic susceptibility by brain wave feedback. Journal of Abnormal Psychology, 1974, 83, 554-560. LONDON, P., COOPER, L. M., AND JOHNSON, H. Subject characteristics in hypnosis research: II. Attitudes toward hypnosis, volunteer status and personality measures; III. Some correlates of hypnotic susceptibility. Journal of Clinical and Experimental
Hypnosis, 1962,10, 13--21. LONDON, P., AND FUHRER, M. Hypnosis, motivation and performance. Journal of Personality, 1961, 29, 321-333. LONDON, P., HART, J. T., AND LEIBOVIlZ, M. P. EEG alpha rhythms and susceptibility to hypnosis. Nature, 1968,219, 71-72. LONDON, P., AND McDEVITT, R. A. AMRL-TR-67-142 (W-P AF Base, Ohio: Aerospace Medical Research Laboratories, 1967). LONDON, P., AND ROCHMAN, G. Untitled mimeographed paper, Department of Psychology, University of Southern California, 1967. LOOMIS, A. L., HARVEY, E. N., AND HOBART, G. Brain potentials during hypnosis. Science, 1936,83, 239-241.
220
DAVID R. ENGSTROM
MASLACH, c., MARSHALL, G., AND ZIMBARDO, P. Hypnotic control of peripheral skin temperature. Psychophysiology, 1972, 9, 60~05. MELEI, J., AND HILGARD, E. R. Attitude toward hypnosis, self-predictions and hypnotic susceptibility. Journal of Clinical and Experimental Hypnosis, 1964, 12, 99-108. MOORE, R. K., AND LAUllR, L. W. Hypnotic susceptibility in middle childhood. Journal of Clinical and Experimental Hypnosis, 1963,11, 167-174. MORGAN, A. H., AND HILGARD, E. R. Age differences in susceptibility to hypnosis. Journal of Clinical and Experimental Hypnosis, 1973,21, 78-85. MORGAN, A. H., JOHNSON, D. L., AND HILGARD, E. R. The stability of hypnotic susceptibility: A longitudinal study. Journal of Clinical and Experimental Hypnosis, 1974,22, 249-257. MORGAN, A. H., AND MACDONALD, H. EEG alpha: Lateral asymmetry related to type of task, task difficulty, and hypnotizability. Paper presented at meeting of the Society for Clinical and Experimental Hypnosis, Newport Beach, California, December, 1973. MORGAN, A. H., McDONALD, P. J., AND MACDONALD, H. Differences in bilateral alpha activity as a function of experimental task, with a note on lateral eye movements and hypnotizability. Neuropsychologia, 1971,9,459-469. NOWLIS, D., AND RHEAD, J. c. Relation of eyes-closed resting EEG alpha activity to hypnotic susceptibility. Perceptual and Motor Skills, 1968,27, 1047-1050. ORNE, M. T. The nature of hypnosis: Artifact and essence. Journal of Abnormal and Social Psychology, 1959,58,277-299. ORNE, M. T., AND EVANS, F. J. Inadvertent termination of hypnosis with hypnotized and simulating subjects. Journal of Clinical and Experimental Hypnosis, 1966, 14, 61-78. ORNE, M. T., AND O'CONNELL, D. N. Diagnostic rating of hypnotizability. Journal of Clinical and Experimental Hypnosis, 1967, 15, 125-133. PASKEWITZ, D. A., AND ORNE, M. T. Visual effects of alpha feedback training. Science, 1973,181, 360-363. ROBERTS, A. H., KEWMAN, D. G., AND MACDONALD, H. Voluntary control of skin temperature: Unilateral changes using hypnosis and auditory feedback. Paper presented at meeting of Biofeedback Society, 1972. ROSENHAN, D., AND LONDON, P. Hypnosis: Expectation, susceptibility and performance. Journal of Abnormal and Social Psychology, 1963,66, 77-81. ROSENHAN, D., AND TOMKINS, S. S., On preference for hypnosis and hypnotizability. Journal of Clinical and Experimental Hypnosis, 1964, 12, 109-114. SACHS, L. B., AND ANDERSON, W. L. The modification of hypnotic susceptibility. Journal of Abnormal Psychology, 1967,15, 172-180. SANDERS, R. S., AND REYHER, J. Sensory deprivation and the enhancement of hypnotic susceptibility. Journal of Anbnormal Psychology, 1969,74, 375-381. SARBIN, T. R., AND COE, W. C. Hypnosis: A social psychological analysis of influence communication. New York: Holt, Rinehart and Winston, 1972. SARBIN, T. R., AND MADOW, L. W. Predicting the depth of hypnosis by means of the Rorschach test. American Journal of Orthopsychiatry, 1942,12, 268-27l. SCHAEFLER, K., AND LONDON, P. Untitled mimeographed paper, Department of Psychology, University of Southern California, 1968. SCHAFER, R. A study of personality characteristic related to hypnotizability. Unpublished masters thesis, Department of Psychology, University of Kansas, 1947. SCHULMAN, R. E., AND LONDON, P. Hypnotic susceptibility and MMPI profiles. Journal of Consulting Psychology, 1963,27, 157-160. SHAPIRO, J., AND DIAMOND, M. J. Increases in hypnotizability as a function of encounter
HYPNOTIC SU5CEPTffiILITY, EEG-ALPHA, AND SELF-REGULATION
221
training: Som,e confirming evidence. Journal of Abnormal Psychology, 1972, 79, 112115. SHOR, R E. Hypnosis and the concept of the generalized reality orientation. American Journal of Psychotherapy, 1959,13, 582-602. SHOR, R E., AND COBB, J. c. An exploratory study of hypnotic training using the concept of plateau responsiveness as a referent. American Journal of Clinical
Hypnosis, 1968,10, 178-193. SHOR, R E., AND ORNE, E. C. Harvard Group Scale of Hypnotic Susceptibility. Palo Alto, Calif.: Consulting Psychologists Press, 1962. SHOR, R. E., ORNE, M. T., AND O'CONNELL, D. N. Validation and cross-validation of a scale of self-reported personal experiences which predicts hypnotizability. Journal of Psychology, 1962,53, 55-75. STEVENS, J. R, SACHDEV, K., AND MILSTEIN, V. Behavior disorders of childhood and the electroencephalogram. Archives of Neurology, 1968,18, 160-177. STUKAT, K. G. Suggestibility: A factorial and experimental analysis. Stockholm: Almqvist and Wiksall, 1958. TART, C. T. Increases in hypnotizability resulting from a prolonged program for enhancing personal growth. Journal of Abnormal Psychology, 1970,75, 260-266. TAUB, E., AND EMURIAN, C. Autoregulation of skin temperature using a variable intensity feedback light. Paper presented at meeting of the Biofeedback Research Society, 1972. TRUE, R. M., AND STEPHENSON, C. W. Controlled experiments correlating electroencephalogram, pulse and plantar reflexes with hypnotic age regression and induced emotional states. In M. V. KLINE (Ed.), Clinical correlations of experimental hypnosis. Springfield, Ill.: Thomas, 1963. ULETT, G. A., AKPINAR, S., AND ITIL, T. M. Hypnosis: Physiological, pharmacological reality. American Journal of Psychiatry, 1972,128, 799-805. WEITZENHOFFER, A. M. Hypnotism: An objective study in suggestibility. New York: Wiley, 1953. WEITZENHOFFER, A. M., AND HILGARD, E. R. Stanford Hypnotic Susceptibility Scale, Forms A and B. Palo Alto, Calif.: Consulting Psychologists Press, 1959. WEITZENHOFFER, A. M., AND HILGARD, E. R. Stanford Hypnotic Susceptibility Scale, Form C. Palo Alto, Calif.: Consulting Psychologists Press, 1962. WEITZENHOFFER, A. M., AND WEITZENHOFFER, G. B. Sex, transference and susceptibility to hypnosis. American Journal of Clinical Hypnosis, 1958,1, 15-24. WICKRAM, 1. Effects of EMG feedback training on hypnotic susceptibility: More preliminary observations. Paper presented at meeting of the American Psychological Association, Honolulu, September, 1972. WICKRAMASEKERA, 1. Effects of sensory restriction on susceptibility to hypnosis: A hypothesis and more preliminary data. Journal of Abnormal Psychology, 1970,76, 6975. WINER, B. J. Statistical principles in experimental design (2nd ed.). New York: McGrawHill, 1971. WISEMAN, R J., AND REYHER, J. A. A procedure utilizing dreams for deepening the hypnotic trance. American Journal of Clinical Hypnosis, 1962,5, 105-110. ZUBEK, J. P. Behavioral and physiological effects of prolonged sensory and perceptual deprivation: A review. In J. RASMUSSEN (Ed.), Man in isolation and confinement. Chicago: Aldine, 1973. ZUBEK, J. P., AND WELCH, G. Electroencephalographic changes after prolonged sensory and perceptual deprivation. Science, 1963,139, 1209-1210.
6
Toward a Cognitive Theory of Self-Control DONALD MEICHENBAUM
I.
INTRODUCTION
For the last 10 years we have been conducting research designed to bring together the clinical concerns of semantic, or cognitive, therapists (e.g., Aaron Beck, Albert Ellis, Jerome Frank, George Kelley) and the technology of behavior therapy (e.g., procedures such as operant and aversive conditioning, desensitization, modeling, and behavioral and imagery rehearsal). This marriage of somewhat strange bedfellows has bred a set of therapy procedures that we have come to call cognitive-behavior modification. At one time we tended to call the procedures self-instructional training, but this title was too delimiting, not permitting ample recognition of imagery- and fantasy-based factors in the change process. This program of research has been described elsewhere (Meichenbaum, 1973, 1975b; Meichenbaum and Cameron, 1974). These studies have indicated the promising outcome, in terms of generalization and persistence of treatment effects, that follow from the alteration of "standard" behavior-therapy procedures to include self-instructional and imagery processes. For example, the efficacy of behavior-therapy procedures such as modeling (Meichenbaum, 1971), desensitization (Meichenbaum, 1972), operant conditioning (Meichenbaum and Goodman, 1971), and aversive conditioning (Steffy, Meichenbaum, and Best, 1970) was enhanced by the focusing of treatment of the client's cognitive processes. See Mahoney's (1974) recent book for a review of the cognitive-behavior modification literature. The purpose of the present paper is (1) to share some general conclusions that derive from our treatment research and (2) to offer a DONALD MEICHENBAUM Waterloo, Ontario.
.
Department of Psychology, University of Waterloo,
223
224
DONALD MEICHENBAUM
cognitive theory of self-control based on these conclusions. Specifically, an attempt will be made to explain why modifying the client's internal dialogue (i.e., self-statements and images) results in behavior change.
II.
CONCLUSIONS FROM TREATMENT
Let me first state the general conclusions that have issued from our treatment research; then we can explore each of them in more detail. First, it has become increasing clear that there are a host of different ways to view our clients' cognitions but that at present we have few or no data to determine the relative merit or long-term effectiveness of these different approaches. A related point is that no matter how one views his client's cognitions, the distinction between a purely behavioral versus a cognitive intervention program is misleading and mistaken. Perhaps we can help put to rest the false distinction between behavioral and cognitive therapies by an interactional model, in which the behavioral and the cognitive processes that underlie change are interdependent. Another general conclusion is that therapeutic change comes about by means of a sequential, mediating process, in which (1) the client becomes aware of his maladaptive intra- and interpersonal behaviors; (2) this self-recognition is the occasion for the client to emit a set of incompatible images and self-statements and incompatible behaviors; (3) finally, what the client says to himself (i.e., his appraisals, attributions, self-statements, and images), following the emission of the new behavioral act and its accompanying consequences, will influence the nature and stability of the change. As far as we, as therapists, can anticipate and subsume the content of the client's internal dialogue in our treatment package, we will be that much more effective. By analogy, the neurological concept of final common pathway explains how behavior change follows from diverse therapy procedures. It is suggested that clients who see therapists of wholly different persuasions go through similar psychological processes in achieving behavioral change. The final common pathway to behavior change is the alteration in the internal dialogues in which our clients engage. The final and perhaps the most important conclusion is that we as
TOWARD A COGNITIVE THEORY OF SELF-CONTROL
225
researchers have not paid sufficient attention to what happens in therapy prior to the point of implementation of the behavior-modification procedures. Little has been written about the initial, conceptualization phase of therapy. In summary, the conclusions I wish to emphasize concern first, the variety of ways therapists conceptualize their clients' thinking processes and the interactional role of cognitive and behavioral processes; second, the suggestion that behavioral change is mediated by means of changes in our clients' internal dialogues; and finally, the significant role of the initial, conceptualization phase in the therapy process.
A. How Shall We Treat Our Clients' Cognitions? When I see a client for assessment and perform a situational analysis of his presenting problem (a la Peterson, 1968, or Kanfer and Saslow, 1969), I also ask him to share with me the feelings and thoughts he has that precede, accompany, and follow the presenting problem. I try to secure his description, his perception, and the meaning that he has fabricated to explain his behavior. I wish to discern from his point of view what is going on, what led up to his present difficulties, and what he thinks should be done to help him. Up to this point I think my clinical behavior is not atypical of clinicians, no matter of what particular persuasion or orientation. My behavior-therapy and psychoanalytic colleagues are likely to have similar concerns. But it is how we conceptualize our client's answers to these questions that will illuminate the variety of clinical orientations. More specifically, the role and the significance attributed to the client's cognitions seem to be the fulcrum that truly distinguishes the various clinical approaches. I would like to describe seven different ways in which therapists have viewed their clients' cognitions. Where appropriate, I will illustrate from our own research program the particular conceptualization.
1. Cognitions as Behaviors How shall we treat our client's cognitions? Shall we view the client's thoughts that precede, accompany, and follow the maladaptive
226
DONALD MEICHENBAUM
behaviors as "behaviors," similar in nature to other nonverbal behaviors that he emits and subject to the same "laws of learning" and social learning principles, such as reinforcement and contingency manipulations? There is a long tradition in making such a continuity assumption between overt and covert events, going back to Dollard and Miller (1950) and Skinner (1953). Homme (1965) has even offered the term coverant (that is, covert-operant) to describe our client's thinking processes. If we view the client's cognitions in this manner, then it suggests that we can affect the frequency and strength of those thoughts by systematically pairing them with the onset and offset of various consequences. Indeed, there is a host of therapeutic procedures, including covert sensitization and anxiety-relief conditioning, that are based upon this notion. However, recent treatment studies by Ashem and Donner (1968), Marshall, Polgrin, and Boutilier (1974), Marks, Boulougouris, and Marset (1971), Meichenbaun and Cameron (1973a), and Sachs and Ingram (1972) have seriously questioned the continuity assumption. When the contingency variable in procedures such as covert sensitization or anxiety-relief conditioning was inverted or made noncontingent, treatment efficacy did not deteriorate. These studies question the validity of viewing and treating our client's cognitions in the same manner as overt behaviors. (See Mahoney, 1974, and Meichenbaum, 1974a, for a more detailed review of this literature.)
2. Cognitions as Part of the Response Chain Perhaps instead we should view the client's cognitions as instances of automatic thoughts (i.e., images and self-statements), which are only part of the maladaptive response chain. According to this conception, the task of therapy is to have the client become aware of the role such thoughts play in the behavioral sequence. A number of theorists, such as Premack (1970) and Bergin (1967), have emphasized the therapeutic value of having the client interrupt the maladaptive response chain by controlling automatic thoughts and producing incompatible self-instructions and images. This viewpoint maintains that target behaviors that are habitual in nature (Le., not premeditated) should first be returned to a "de automatized" condition, in which the habitual maladaptive behaviors come to be preceded by cognitive activity occurring within the
TOWARD A COGNITIVE THEORY OF SELF-CONTROL
227
client's awareness. Such "forced mediation" increases the separation between stimuli and responses and thereby provides an additional opportunity for interrupting the behavioral sequence. In this way one can impose an interruption of the response chain, thus increasing the likelihood of termination of the sequence at an earlier stage and production of incompatible thoughts, images, and behaviors. Illustrative of this approach is a self-instructional training program developed by Meichenbaum and Goodman (1971) to treat hyperactive, impulsive children. The impetus for self-instructional training program came from the work of the Soviet psychologists Luria (1961) and Vygotsky (1962). Luria suggested that the child goes through three stages in developing internalized control of behavior. His performance is first controlled by the verbal instructions and reactions of external agents (e.g., parents). The child then begins to regulate some of his own actions through audible self-talk. Finally, these self-statements become covert (i.e., go "underground," to use Vygotsky's term) and expand their extensive regulatory influence. The Meichenbaum-Goodman self-instructional training program followed, in abbreviated form, such a developmental progression. In order to achieve covert self-instructional control of behavior in hyperactive children, the training regimen was as follows: 1. An adult model performed a task while talking to himself out loud (cognitive modeling), 2. The child performed the same task under the directions of the model's instructions (overt, external guidance), 3. The child performed the task while instructing himself aloud (overt self-guidance), 4. The child whispered the instructions to himself as he went through the task (faded, overt self-guidance), 5. And finally, the child performed the task while guiding his performance via private speech (covert self-instruction). Over a number of training sessions the package of self-statements modeled by the experimenter and rehearsed by the child (initially aloud and then covertly) was enlarged by means of response chaining and successive approximation procedures. For example, in a task that required the copying of line patterns, the examiner performed the task while cognitively modeling as follows:
228
DONALD MEICHENBAUM
Okay, what is it I have to do? You want me to copy the picture with the different lines. I have to go slowly and carefully. Okay, draw the line down, down, good; then to the right, that's it; now down some more and to the left. Good, I'm doing fine so far. Remember, go slowly. Now back up again. No, I was supposed to go down. That's okay. Just erase the line carefully . . . Good. Even if I make an error I can go on slowly and carefully. I have to go down now. Finished. I did it!
In the thinking-out-Ioud phase the model displays several performance-relevant skills: (1) problem definition ("What is it I have to do?"); (2) focusing attention plus response guidance ("Be careful ... draw the line down."); (3) self-reinforcement ("Good, I'm doing fine."); and (4) self-evaluative coping skills plus error-correcting options ("That's okay, even if I make an error I can go slowly."). Such training, provided over a number of different tasks, was successful in causing hyperactive children to learn to think before they act, to employ mediational processes, and to develop verbal control of behavior (Meichenbaum and Goodman, 1971). A number of other investigators have also successfully trained children to bring their behavior under self-instructional and imagery control (d. Bem, 1967; Palkes, Stewart, and Kahana, 1968; Blackwood, 1970; Ridberg, Parke, and Hetherington, 1971; Monahan and O'Leary, 1971; Palkes, Stewart, and Freedman, 1972; Denny, 1972; Hartig and Kanfer, 1973; Mischel, 1974; Schneider, 1974). In each of these studies self-control was enhanced as the involuntary act was made voluntary. This was accomplished as the child's behavior was brought under his own cognitive control through the emission of deliberate self-statements and images. Then, with the development of task proficiency, or what Kimble and Perlmutter (1970) call the "automatization of voluntary acts," the child's private speech became more abrupt, incomplete, and whispered and then completely vanished. This process of abbreviation and interiorization of private speech also applies to adults as they acquire skills. For example, one can imagine a similar sequence in the learning of a new motor skill such as driving a car. As Henry Murray (1938) noted some years ago; When one is learning to drive an automobile, one is, at first, aware of every accessory intention and subsequent motor movement, but later, when proficiency has been attained, the details of the activity are seldom in consciousness. (p. 51)
In other words, early in the mastery of a voluntary act, speech serves a useful supporting function. With practice, these verbalizations disappear.
TOWARD A COGNITIVE THEORY OF SELF-CONTROL
229
Thus the therapist can assess the role of his client's cognitions as part of the response chain and help his client produce intentional, adaptive cognitions (Le., self-instructions and images) that will interrupt the maladaptive behavioral chain and foster incompatible, adaptive behavior. It should be noted that the client must learn to use his own behaviors, feelings, and thoughts as well as the behavior of others as cues or signals to engage in this newly acquired, internal dialogue. The importance of self-observation in the change process will be more fully discussed below.
3. Cognitions as Instances of Irrational Thinking Styles We have viewed our client's cognitions as instances of behavior per se, or as part of the response chain. There are alternatives. Aaron Beck has been most helpful in noting some of the stylistic qualities of client's cognitions, especially depressed clients (Braff and Beck, 1974). Beck (1970a) attempts to have clients become aware of the distortions in their thought patterns. These distortions include (1) arbitrary inference--the drawing of a conclusion when evidence is lacking or is actually contrary to the conclusion; (2) magnification-exaggeration of the meaning of an event; (3) cognitive deficiency-disregard for an important aspect of a life situation; (4) dichotomous reasoning-overly simplified and rigid perception of events as good or bad, right or wrong; and (5) overgeneralization-taking a single incident such as failure as a sign of total personal incompetence, and in this way generating a fallacious rule. Such cognitive distortions result in the dient's selectively attending to and inaccurately anticipating consequences and in his making logical errors. By means of pinpointing such stylistic qualities, the client is brought to understand that his affective experiences and maladaptive behaviors are a result of his faulty thinking processes-thinking processes that the client is capable of changing and controlling. Note that the focus has shifted from treating the client's cognitions as an instance of behavior to the stylistic qualities of the client's thinking processes. It is important to underscore Beck's (1970b, 1971) observation that the client's faulty cognition may often take a pictorial form instead of, or in addition to, the verbal form. For example, Beck reported that a woman with a fear of walking alone found that her spells of anxiety followed images of her having a heart attack and being left helpless; a
230
DONALD MEICHENBAUM
college student discovered that her anxiety of leaving the dormitory was triggered by visual fantasies of being attacked. As Goldfried, Decenteceo, and Weinberg (1974) have indicated, because of the habitual nature of one's expectations or beliefs it is likely that such thinking processes and images become automatic and seemingly involuntary, like most overlearned acts. The client's faulty cognitions-negative and anxiety-engendering self-statements and images-become habitual and in many ways are similar to the automatization of thought that accompanies the proficiency of a motor skill such as driving a car. However, the therapist can make the client aware of such thinking processes and increase the likelihood that such an awareness will be the trigger that produces incompatible thoughts and behaviors. (See Meichenbaum, 1975b, for a description of how the clinician can achieve this process.)
4. Cognitions as Instances of Irrational Belief Systems Let us continue the list of alternative conceptualizations available to the clinician as he sits in his chair, listening to his client. Whereas Beck emphasized the process and style of our client's thinking, Albert Ellis (1961) emphasized the so-called underlying premises that contribute to our client's faulty thinking, emotional disturbance, and maladaptive behavior. Ellis (1961) proposed that a major core of emotional disturbances has to do with the client's preoccupation with what others think of him and the mistaken belief that an individual's selfworth is determined by others. Ellis encouraged the clinician to note the themes, the irrational premises, that underlie our patients's selfstatements, images, and cognitions. This view, that psychological problems arise from misperceptions and mistaken cognitions about what a client perceives was most SUCcinctly summarized by the stoic philosopher Epictetus (60 A.D.), who said, "Men are disturbed not by things, but the views they take of them." Therefore Ellis attempted to have clients examine the irrational ideas and beliefs, such as the following, that give rise to misperceptions: 1. I must be loved or approved of by practically every significant person in my life, and if I'm not it's awful. 2. I must not make errors or do poorly, and if I do it's terrible.
3. People and events should always be the way I want them to be.
TOWARD A COGNITIVE THEORY OF SELF-CONTROL
231
In order to counteract such irrational beliefs, the rational-emotive therapist encourages, goads, challenges, and educates by means of a Socratic dialogue; provides information; conducts rational analyses; assigns behavioral homework assignments; and so on, in order to have the client entertain the notion that his maladaptive behavior and emotional disturbance are a reflection of a commitment to irrational beliefs. As a result of such therapeutic interventions it is hoped that the client will replace the irrational beliefs described above with the following: 1. It's definitely nice to have people's love and approval, but even without it, I can still accept and enjoy myself. 2. Doing things well is satisfying, but it's human to make mistakes. 3. People are going to act the way they want to, not the way I want them to. The complexity of such cognitive restructuring therapy is illustrated in research conducted by Meichenbaum (1972, 1974c) and Meichenbaum, Gilmore, and Fedoravicius (1971). Included in the cognitive restructuring treatment regimen were the following components: 1. Didactic presentation and guided self-discovery of the role of selfstatements in subjective distress and performance inadequacies. 2. Training in the fundamentals of problem solving (e.g., problem definition and anticipation of consequences). 3. Training in the discrimination and systematic observation of selfstatements. 4. Graduated performance assignments. 5. Structured modeling of both overt and cognitive skills in the form of self-statements and images. 6. Modeling and encouragement of positive self-evaluation and of coping and attentional focusing skills. 7. Depending on the treatment package employed, the use of behavior-therapy procedures such as relaxation training, coping imagery training, and behavioral rehearsal. Thus a complex, multifaceted training procedure was employed to change the client's irrational beliefs, self-statements, images, and maladaptive behaviors. Although therapeutic procedures such as El-
232
DONALD MEICHENBAUM
lis's rational-emotive therapy (RET) have been available and professionally visible for well over a decade, there is a sparsity of controlled experimental data bearing on their efficacy. A few encouraging studies of the efficacy of RET have been offered (Baker, 1966; DiLoreto, 1971; Karst and Trexler, 1970; Meichenbaum et al., 1971; Trexler and Karst, 1972). However, after reviewing the outcome literature for RET and cognitive restructuring therapy in general, Mahoney (1974) concluded-and my own assessment of the literature is in full accord: "the clinical efficacy of Ellis' rational-emotive therapy has yet to be adequately demonstrated" (p. 182). A similar assessment could be made of the cognitive-therapy position of Beck. The cognitive restructuring procedures as conducted by Beck, Ellis, and Meichenbaum vary in several respects, most notably in terms of the relative emphasis placed on formal logical analysis (i.e., isolation and evaluation of premises), the directiveness with which the therapeutic rationale and procedures are presented, and the adjunctive use of behavior-therapy procedures. Future research is necessary to determine the Significance of such differences. 5. Cognitions as Instances of Problem-Solving Ability
Whereas the Ellis's approach sensitizes the therapist to listen for the presence of maladaptive, self-defeating, anxiety-engendering cognitions, the therapist with a problem-solving orientation listens for the absence of specific, adaptive cognitive skills and responses. Illustrative of this approach are D'Zurilla and Goldfried (1971) and Goldfried (1975). They suggest that our client's cognitions reflect a deficit in systematic, problem-solving skills. Treatment is designed to have clients learn how to specify problems, generate alternative solutions, tentatively select a solution, and then test and verify that solution. The clinical potential of such a problem-solving approach is illustrated in the treatment research of Spivack and Shure (1974) with disruptive preschool children. In previous research, Spivack and his colleagues (Shure and Spivack, 1972; Shure, Spivack, and Jaeger, 1971) found that children exhibiting maladaptive behavior are often less capable of employing means-ends thinking and frequently limit their problem solutions to impulsive and aggressive methods. By means of problem-solving training Spivack and Shure (1974) were able to train disruptive children to consider alternatives and to engage in cause-
TOWARD A COGNITIVE THEORY OF SELF-CONTROL
233
effect thinking. Such training seemed to reduce superfluous and irrelevant thinking and to help children explore nonforceful possibilities, resulting in significant behavioral change. These results take on particular significance when it is noted that adolescent and adult psychiatric patients have also been noted to manifest problem-solving deficits, most notably the absence of foresight in considering the possible consequences of various actions (e.g., see studies by Platt and Spivack, 1972a,b; Spivak and Levine, 1963). Several investigators have explored the therapeutic potential of problem-solving training. Their applications have included (1) the use of imagery and thematic fantasy play with culturally deprived and behavior-problem children (Freyberg, 1973; Saltz and Johnson, 1974; Schneider, 1974); (2) the use of problem solving in crisis clinics (McGuire and Sifneos, 1970); (3) the use of problem solving in assisting adolescents to handle various conflict situations (Kifer, Lewis, Green, and Phillips, 1973) and ex-drug addicts to remain drugfree (Copeman, 1973); and (4) the potential application of self-instructional, problem-solving skills with the aged, in the form of a "cognitive prosthesis," to overcome any age-related deficits which may appear (Meichenbaum, 1974b). A recent, encouraging application of a self-instructional problemsolving-training approach was offered by Hanel (1974), as cited by Heckhausen (1974). Hanel was working with fourth-graders who were selected for a marked fear of failure, in addition to poor academic records. Using the self-instructional training procedure developed by Meichenbaum and Goodman (1971), Hanel was successful in teaching these children to talk to themselves differently, to problem-solve, in order to change their motivational style and academic performance. The experimenter cognitively modeled for the children how to set standards, plan actions, calculate effort output, monitor performance, evaluate performance outcome, weigh causal attributions, and administer self-reward. Then the students took turns in performing tasks while emitting similar cognitions (initially aloud and then covertly). The result of the children's adopting the modeled cognitive processes was improved academic performance and changes on laboratory measures such as level of aspiration, attributional style, and patterns of self-reinforcement. The Hanel study, as well as those mentioned earlier, indicates the therapeutic promise in viewing our client's cognitions as instances of problem-solving ability.
234
DONALD MEICHENBAUM
Common to each of these problem-solving approaches is an attempt to teach the client to engage in covert problem-solving entailing symbolic stimulus-transformation, cognitive rehearsal, and tests of alternate solutions. An illustration of the role such covert problem-solving rehearsal plays is offered by Singer and McCraven (1961), who, in a questionnaire study of daydreaming behavior in a normal adult population, found that 96% of their subjects engaged daily in some form of daydreaming. Their daydreams took the form of fairly clear images of people and events. Daydreams dealing with planning for future actions, and particularly interpersonnal contacts, were in high frequency, with the largest percentage of daydreams involving fairly practical immediate concerns. For most of the respondents, daydreaming was not a matter of wish-fulfillment ideation but rather an attempt to explore the future through positing a variety of alternatives, not specifically involving satisfactory outcomes.
6. Cognitions as Instances of Coping Skills Closely akin to a problem-solving approach to our client's cognitions is a skills-oriented training approach. Whereas the problemsolving approach emphasizes the client's learning to stand back and systematically analyze a problem situation in the absence of any acute stress, the coping-skills approach concentrates on what the client must do when immediately confronted with an acute stress-situation. Indeed, the problem-solving process may include rehearsal of coping skills as clients fantasize dealing with stressful events. The increasing clinical attention given to a skills-oriented treatment approach has been noted by Mahoney (1974), who commented that a shift in behavior treatment research is underway: from a focus on discrete, situation-specific responses and problem-specific procedures to a coping-skills model, which can be applied across situations and problems. This model views the client's cognitions as instances of cognitive skills that he can employ in confronting stressors. As with the other therapeutic approaches reviewed, the supporting evidence for the effectiveness of mediationally based, coping-skills training is very sparse, but initial investigations are encouraging. Some examples of how we can teach our clients cognitive coping skills are offered. Meichenbaum and Cameron (1973b) developed a copingskills-training package, which they described as "stress-inoculation"
TOWARD A COGNITIVE THEORY OF SELF-CONTROL
235
training. The treatment regimen drew heavily on Meichenbaum's selfinstructional research, and included (1) a discussion of stress reactions (with emphasis on labeling, attribution, and arousal-inducing selfstatements); (2) relaxation training (presented as an active, coping skill); (3) instructed practice in the emission of coping self-statements (cognitive self-monitoring, preparation for stress, self-reinforcement); and (4) supervised practice in utilizing the coping skills in an actual, stressful situation (e.g., an unpredictable shock situation). Meichenbaum and Cameron (1973b) found that such a stress-inoculationtraining procedure was effective in modifying the fearful behavior of mutiphobic clients. Turk (1975) has recently sucessfully applied the stress-inoculation training to subjects who received experimentally induced pain, and Novaco (1974) to clients with extreme anger. Interestingly, Turk employed a number of cognitive training procedures, including imagery and self-instructional rehearsal, to teach pain tolerance. The imagery manipulations included changes and transformations of the pain. These cognitive coping procedures were presented in a "cafeteria" style and subjects could pick and choose those that worked best for them. A number of other investigations (Goldfried, 1971; Goldfried and Trier, 1974; Langer, Janis, and Wolfer, 1973; Suinn and Richardson, 1971; Tori and Worrell, 1973) have also successfully applied a copingskills approach. In each case the client's cognitions playa major role in the change process. The clinical potential of coping training was further illustrated in the cognitive modeling research of Kazdin (1973, 1974a,b). In a series of studies, Kazdin demonstrated that covert modeling (i.e., mental rehearsal by the subject of modeled behavior) was effective in reducing phobic behavior and engendering assertive behavior. Such symbolic rehearsal, especially when it included coping self-statements and behaviors, proved to be a most effective therapeutic intervention. Such cognitive rehearsal in preparation for a stressor is similar to the cognitive process of the "work of worrying" which Marmor (1958) and Janis (1958) have described. The cognitive-behavior modification approach suggests that clients can be explicitly taught how to worry in such a constructive fashion. The "work of worrying" can now be translated into sets of self-statements and images, which can be modeled by the therapist and rehearsed by the client. Sarbin (1972) has viewed such imagery rehearsal as a form of muted roletaking.
236
DONALD MEICHENBAUM
A research area that may prove most heuristically valuable for the behavior modifier who is interested in such cognitive manipulations is the work on mental practice. Richardson (1967a,b) has summarized a considerable body of evidence that indicates that in a variety of different physical tasks, subjects improve their performance after spending varying amounts of time in "thinking about" or imagining themselves in the act of performing. Increased motivation as a result of mental practice and increased task sophistication, analogous to test sophistication, might account for improvement. The importance of internalizing a very clear model of what a good performance of the task is like is indicated by the fact that the more familiar a task has become the greater the relative gain that can be expected from mental practice. Thus an examination of such factors as the degree of task familiarity, accuracy of anticipated outcomes, clarity and control of visual and kinesthetic imagery, degree of proficiency on the task, length of time provided for imagery, and the alternation of mental and physical practice, which have been found to be important in the mental practice area, is likely to be of importance to the cognitivebehavior therapist.
7. Cognitions as Instances of Defense Mechanisms A long tradition derives from psychoanalytic theory, in which the client's cognitions are viewed in terms of their defensive aspects. The client's cognitions are viewed as manifest signs of underlying conflicts, many of which the client will be unaware of. Illustrative of this approach is the work of Shapiro (1965), who has noted client's neurotic styles. An evaluation of this conception of our client's cognitions is beyond the scope of the present review. In summary, we can view our clients' cognitions as behaviors, automatic thoughts and thus part of the response chain, reflections of cognitive styles and faulty belief systems, inadequate problem solving and coping skills, or defense mechanisms. Indeed, the task for the research clinician is to match the most useful conceptualization with the specific client's problem and the goals of treatment (Le., an adaptive treatment approach). But what difference does it make how we view the client's cognitions? "I am a behavior therapist." It is behavior that I'm after. My approach is to have the client emit the incompatible adaptive
TOWARD A COGNITIVE THEORY OF SELF-CONTROL
237
behavior and the cognitions will take their natural course. Haven't you heard the old adage, "It is easier to act your way into a new way of thinking, than to think your way into a new way of acting"? Can behavior therapy be viewed as simply as the adage suggests? Indeed not! It is suggested that cognitions play a substantial role in the change process. It is proposed that there is an interactional process between cognitions and behaviors, one preceding and following the other. Each time we attempt to modify some aspect of the client's behavior (e.g., teaching him to become more interpersonally assertive or to exert self-control in areas of eating or smoking), the client is also changing his internal dialogue or what he says to himself. The evidence as reviewed by Mahoney (1974) and Meichenbaum (1973) suggests that treatment efficacy is enhanced when the client's internal dialogue is incorporated into the treatment regimen. Even in such behavioral therapy procedures as operant conditioning, the client's perceptions or attributions or what he says to himself about the dispensed reinforcement influence the outcome. (For example, see work of de Charms, 1968, Deci, 1971, and Steiner, 1970.) Bandura (1974), in his APA presidential address, also questioned the automaticity of reinforcing consequences in the absence of mediating cognitive processes: So-called conditioned reactions are largely self-activated on the basis of learned expectations rather than automatically evoked. The critical factor, therefore, is'not that events occur together in time, but that people learn to predict them and to summon up appropriate anticipatory reactions. (p. 2)
As Bergin (1970) has suggested, "There may be highly specific interventions which have a behavioral or cognitive focus, but these are always embedded in a multidimensional context or have multiple consequences" (p. 208). Our clients present problems that require changes in motoric, affective, and cognitive domains. The focus on behavior or cognition would thus appear to be misguided and shortsighted. In the same way that psychologists have been seduced into arguing the either-or position of heredity versus environment, trait versus situationism, we have been seduced into arguing behavioral versus cognitive change. Our job is to find out how cognitive and behavioral processes interact in leading to change. Perhaps it is time to consider changing the title of the journal Behavior Therapy to Cognitive-Behavior Therapy, with most emphasis on the,hyphen.
238
DONALD MEICHENBAUM
B. Cognitions as Final Common Pathways In order to illustrate the role cognitions play in the change process, let me offer a quote from Jerome Singer's (1974) recent book on imagery and daydreaming methods in psychotherapy and behavior modification. In describing the successful treatment of psychoanalysis, he suggests that the following process of self-awareness and change occurs: A patient experiences a sudden sense of unrest or annoyance upon entering a room. Under some past conditions he might have hastily left the room or perhaps talked rudely in response to questions raised. His analytic experience now alerts him to the fact that this sudden unease is occasioned by an irrational anticipation or transference in the situation. He replays in his mind the thoughts just previous to entering the room or what he was thinking about immediately prior to this situation. On this mental screen, he "instant replays" the thoughts and perceptions that occurred and suddenly is aware that he had been thinking about some obligation to one of his parents and that on entering the room he noticed across the wayan elderly gentleman who rather resembled his father. He now perceives that his distress is a combination of anticipatory image plus the scene occurring in the room and generally is freed of his anxiety and certainly is less likely to engage in an irrational and self-defeating bit of behavior in this new situation. (p. 64)
The Singer quote nicely illustrates two points that I would like to make. First, in order to bring about change, the client must recognize some behavior that he emits (e.g., a set of thoughts, images, and physiological and motoric responses) or the interpersonal response of someone else. This recognition is the necessary but not sufficient condition for change. This recognition, this self-awareness, acts as the cue, the bell ringer, the discriminative stimulus for producing a set of incompatible thoughts and behavior. Following therapy the client no longer responds impulsively, in a stimulus-response manner, to externally or internally generated events. Instead, a mediational process is elicited by stimuli and such internal processes now precede the emission of the overt response. Insofar as stimuli or situations elicit the same mediational processes or internal dialogue, the treatment effects will generalize. It should be noted that generalization is engineered into the treatment package. For now the client's own maladaptive behavior is always the reminder to use the coping skills that were taught in therapy. What incompatible thoughts and behavior the client emits at this
TOWARD A COGNITIVE THEORY OF SELF-CONTROL
239
point vary with the orientation of the therapy and the nature of the conceptualization that has evolved between patient and therapist. The client's internal dialogue may be in terms of pscyhoanalytic interpretations as in the Singer example, or learned response habits a la Wolpe, or faulty belief systems a la Ellis. Indeed it is suggested that our clients have sufficient life experiences to provide data consistent with any one of these therapy conceptualizations, whether psychoanalytic, Jungian, Rogerian, Gestalt, semantic, or behavioral. The human life condition provides sufficient experience to maintain the employment of a host of therapists of widely different persuasions. The important point is that our clients have a need to fabricate a meaning, some understanding, a conceptualization about what is happening to them and what can be done to help them to change. What becomes essential for the cognitive-behavior therapist is how to have the client adopt a conceptualization of his problem that will lead to specific behavioral and cognitive changes that can be transferred to real-life situations. This leads me to the important role of the initial, conceptualization phase in therapy.
C. Initial, Conceptualization Phase of Therapy The role of the conceptualization process in therapy has not received much attention by behavior-modification researchers or practitioners. It is usually subsumed under such terms as nonspecific therapy factors or included as those aspects that go "beyond" behavior therapy (e.g., Lazarus, 1971). What goes on in therapy before desensitization, or some other behavior-therapy procedure, is implemented, is rarely discussed. Few therapy studies mention the rationale that has been offered to the patient prior to treatment. There are a variety of ways to have the client and the therapist evolve a common conceptualization. Some therapists are very directive and didactic and seem to force upon the client a particular conceptualization by power of their personalities, jargon, or positions. In some cases such a "hard-sell" approach may prove successful. But the therapist must be concerned about the client's self-statements and attributions about the therapy process, as well as those concerning his presenting problems.
240
DONALD MEICHENBAUM
An alternative way to proceed is to have the client and therapist evolve a common conceptualization, so that the client feels that he is an active participant and contributor. The manner in which the therapist discusses the presenting problem, the kinds of questions he asks, the type of assessment procedures employed, the content of the therapy rationale, and the kinds of homework assignments given are all used to evolve a common client-therapist conceptualization. Once the client accepts a certain conceptualization of his problem then he readily engages in the treatment assignments. As Jerome Frank (1961) pointed out, the shared conceptual system between therapist and client is important in the change process. Insofar as the client adopts a particular paradigm, or comes to view his behavior from a given perspective, thus far are the client's "assumptive world" (to use Frank's term) and behavior open to change. Such a conceptualization evolves over the course of treatment as the therapist cognitively models particular beliefs and encourages the client to engage in an active self-examination. The role that conceptualization plays in the change process is illustrated in both our laboratory research and our clinical work. In our laboratory research most emphasis has been placed on this initial, conceptualization phase. In treating phobics or interpersonally anxious clients we have conceptualized their presenting symptoms of arousal (e.g., muscular tension, pounding hearts, sweaty palms, and heavy breathing) and accompanying task-irrelevant, anxiety-engendering thoughts in terms of Schachter and Singer's (1962) theory of emotion. Thus treatment could be directed naturally toward (1) helping the client control his physiological arousal by means of relaxation and (2) substituting positive, coping self-statements for the anxietyengendering self-statements that habitually occupy the client's mind under stressful conditions. In the treatment of pain patients we conceptualized the subjects' pain in terms of Melzack and Wall's (1962) gate-control theory of pain (Meichenbaum, Turk, and Burstein, 1975). It should be noted that the scientific validity of a given conceptualization is less important than the aura of its plausibility. The aim of the therapist is not primarily to impart precise, scientific information, but rather to provide the client with a conceptualization that will facilitate therapy of making its rationale comprehensible. Similar examples could be offered from our work with low-creativity subjects, test-anxiety subjects, and others (see Meichenbaum, 1975b). The general treatment strategy is to share with the client, in terms that he can
TOWARD A COGNITIVE THEORY OF SELF-CONTROL
241
clearly understand, the rationale that led to the present treatment procedures. More specifically, the goal of the conceptualization phase is to have the client talk to himself differently about his presenting problem. An attempt is made to have the client change his perceptions, attributions, sense of control, and sense of helplessness about his presenting problem-in short, to alter the client's internal dialogue regarding his appraisal of his maladaptive behaviors and emotional disturbance. For example, in treating multiphobic patients the therapist helped the client change his perception of how he behaved in a fearful situation. Instead of viewing his response as a massive panic reaction, the therapist suggested that the phobic client's response seemed to include several stages, namely, preparing for, then confronting or handling the stressor, possibly being overwhelmed, and finally, reinforcing himself for having coped. In this way the phobic patient no longer had a "massive phobic reactions," but rather a series of responses that went through four stages, for each of which the client could be trained to employ adaptive coping responses. For once the client views his problem from a given perspective, then the acquisition of a number of specific skills makes sense and they are actively rehearsed. One could offer a number of clinical cases to illustrate how the initial phase of therapy is designed to have the client talk to himself differently about his presenting problem. A translation process occurs. Initially the client describes what's bothering him, often with a sense that he is losing control and feeling helpless and hopeless. The therapist, with skill, has the client come to view his problem from a different perspective, to fabricate a new meaning or explanation for the etiology and maintenance of the client's maladaptive behavior. Whereas prior to therapy the client may view his compulsion to wash as a sign of his "losing his wits," being depressed, etc., as a result of therapy he may come to view his washing as a communication problem, or as a manifestation of deep-seated conflict about guilt, or as a behavioral repertoire maintained because of the secondary gains (reinforcers) that accrue, etc. Many other conceptualizations could be offered. Indeed the patient may provide enough data to support each of these conceptualizations. Most patients do! The exact conceptualization adopted in therapy will vary with the therapist's orientation, the patient's expectations and the goals of therapy. From the present perspective, the important theoretical problem is
242
DONALD MEICHENBAUM
"Why does altering the client's conceptualization-how he views his problem and what he says to himself about it-result in behavior change?" It should be apparent that the clinician has a host of alternative ways to view his client's cognitions. At present we have little empirical evidence (Mahoney, 1974) to guide us in determining which conceptualization will prove to be most efficacious. In fact, we may come to learn, by means of systematic investigation, which conceptualization works best with which clients. As Dember (1974) has stated, "Psychology has gone cognitive" (p. 161), and it is time for behavior modification to do likewise. In order to engender the shift to cognition, the following list of conclusions and implications is offered. 1. First, we can no longer compare the effectiveness of specific behavior-therapy procedures such as desensitization with an "insight" or "semantic" therapy. The uniformity myth with respect to treatment procedures that Kiesler (1966) described can no longer be applied to semantic or cognitive therapies. Instead we must encourage comparisons between different cognitive approaches in order to identify the parameters that underlie cognitive restructuring. What are the relative therapeutic merits of viewing our clients' cognitions from such perspectives as those of Ellis, Beck, Meichenbaum, D'Zurilla, and Goldfried? 2. When reading a therapy study one must carefully attend to the details of the therapist's manual, especially those phases in which an initial conceptualization is offered. The conceptualization phase must be seen as an active ingredient of the therapy process and not something beyond the researchable interests of behavior therapists. 3. As clinicians we must become more sensitive to the thoughts and feelings our clients have in the criterion situations. What would you like your client to say to himself in order to cope more adequately? Can we not teach our clients to use their own maladaptive behaviors as cues for using coping skills? Indeed we can use the technology of behavior therapy to influence our clients' internal dialogues (see Meichenbaum and Cameron, 1974). 4. Finally, behavior therapists may wish to consider how they would alter such therapy techniques as operant-conditioning programs with parents, self-control training with obese patients, desensi-
TOWARD A COGNITIVE THEORY OF SELF-CONTROL
243
tization with phobics, aversive conditioning with pedophiles, assertive training with college students, biofeedback with headache patients, and so on, in order to include self-instructional and imagery processes. It is suggested that as we become more cognitive in our orientation, we will become more effective in our practice. The need for a cognitive-hyphen-behavior therapy approach is now! In order to explain further the nature of the hyphen, a cognitive theory of self-control is offered.
III. A
COGNITIVE THEORY OF SELF-CONTROL
As already mentioned, the therapist has a variety of ways to view his client's cognitions. Each conceptualization leads to different therapeutic interventions. However, it has been suggested that even though the therapy procedures differ in emphasis and techniques, clients go through a similar cognitive process in achieving behavioral change. The research conclusions that have been outlined can now be integrated into a more general and coherent theory of a three-stage process accounting for therapeutic change.
A. A Three-Stage Process 1. Stage 1: Self-Observation
The first step in the change process is the client's becoming an observer of his own behavior. Through heightened awareness and deliberate attention, the client monitors, with increased sensitivity, his thoughts, feelings, and/or interpersonal behaviors. In the second stage of the change process, this self-observation of the client's inappropriate actions will, upon the occurrence of a maladaptive behavioral event, serve as a signal, or cue, to produce thoughts and behaviors incompatible with the continuation of the inappropriate cognitions and behaviors. The process of self-observation is a necessary but not a sufficient condition for change. On which behaviors the client focuses depends upon the conceptualization process that evolves during therapy. The important role of
244
DONALD MEICHENBAUM
this process needs once again to be underscored. The client's motivation and skill in acting as a self-observer are fostered as a result of the conceptualization process. During the course of therapy a translation process occurs, through which the client's definition, description, and fabricated meaning concerning his problems are subjected to examination and change. The exact terms of the translation vary with each therapist's orientation and each client's characteristics (e.g., client expectations and background). The important point is that the client and the therapist go through a series of steps, such as interviewing, testing, and homework assignments, whereby the client comes to entertain a different view of his maladaptive behaviors and emotional disturbances. The client may come to view his problems from a psychoanalytic, a social learning, or a semantic viewpoint. Whereas the client enters therapy with one description or language system, with particular referent terms and explanatory concepts, as a result of therapy he comes to view his complaints and dissatisfactions in different terms. The therapist uses a host of clinical tools, such as reflections, explanation, interpretation, information giving, and cognitive modeling, to achieve this translation process. The goal of each of these therapeutic techniques is to have the client view his behavior differently. The end result is that the client becomes an observer of his own behavior. One of the by-products of the translation process and the increased self-awareness is that the client gains a sense of control of his emotional state and behavior; he feels that he is an active contributor to his own experience and not a helpless victim of his thoughts and feelings and the reactions of others. A sense of hopefulness and "faith" are aroused. As Strupp (1970) pointed out, a major component of effective psychotherapy is the client's experience of having increased his control over his own emotions and overt behavior. This translation process, in the form of a new conceptualization, results in a cognitive restructuring or new internal dialogue by the client. Terms such as sense of control, hope, faith, expectancy, and cognitive restructuring have been offered by many theorists to explain the therapy process. For purposes of the present explication, each of these concepts is viewed as part of the client's internal dialogue. Our clients think; they engage in an emission of thoughts, images, and selfstatements. We, as psychologists, are interested in finding out the
TOWARD A COGNITIVE THEORY OF SELF-CONTROL
245
most theoretically useful and heuristically productive conceptual system to explain our clients' internal dialogues. The goal is a kind of ethological description of our clients' thinking processes. At present, the "marketplace" is filled with competing concepts and schemes to explain our clients' thoughts and images. In the same way that we try to fashion an explanation or fabricate a meaning for our clients' thinking processes, our clients attempt to develop an understanding of their own thoughts, feelings, and behaviors. It is postulated that both therapist and the client-and man in general-have a need to believe, to understand, to impose an explanation on events. Evidence that such a need exists may be offered from such diverse areas as the psychology of superstition (Jahoda, 1969), the universal role of religious belief (Allport, 1950), and the wide appeal of scientific investigation, and it is perhaps most interestingly documented by Pitkin's book A Short Introduction to the History of Human Stupidity (1935). The translation process in therapy serves to create a conceptual framework, which provides a basis for the client to monitor his cognitions and behavioral productions effectively in terms that will then serve as springboards for therapeutic change. 2. Stage 2: Incompatible Thoughts and Behaviors
Once the client has become an observer of his behavior and these self-observations have been reinforced by, and in tum, reinforce, the conceptualization process, the second stage in the change process occurs. The process of self-observation becomes the occasion or stimulus for the client to emit different cognitions and behaviors. This point was illustrated before, with the quote from Singer's book. The content of what the client now says to himself will vary with the conceptualization that emerged in therapy. If the client's behavior is to change, then what he now says to himself, and/or imagines, must initiate a new behavioral chain, one which is incompatible with his maladaptive behaviors. 3. Stage 3: Cognitions Concerning Change
The third step in the change process, what the client says to himself about his newly acquired behaviors, determines whether the behavioral change will be maintained and will generalize. As the
246
DONALD MEICHENBAUM
client attempts to behave differently, he will often elicit different reactions from significant others. What the client says to himself and imagines about these reactions and his own behavior change will influence the stability and generalizability of the treatment. As stated before, insofar as we as therapists can anticipate and incorporate the client's internal dialogue into the therapy process, we will increase our effectiveness. Once again, the situation is that our client is emitting a set of thoughts and images, and the question is how best to describe such mediational events. Social psychologists are prone to characterize such internal dialogue in terms of appraisals, attributions, perceived freedoms, psychological reactance, etc. Some therapists may conduct a hierarchical analysis of the client's internal dialogue in terms of selfstatements, attitudes, beliefs, values, etc. Others may engage in a cluster analysis of the client's cognitions in terms of self-concept. Our own preference is to engage in a situational analysis of the specific self-statements and images the client emits, noting their similarities and differences. It is hypothesized that consistency of behavior across situations or treatment generalization is a function of the degree to which the individual emits a set of similar self-statements andior images across situations. Parenthetically it may be noted that what is being offered is a concept-formation view of personality. Insofar as the same mediators (i.e., appraisals, attributions, self-statements, and images) are elicited across situations, one will observe behavioral similarities. Once description of the mediating events has been recently offered by Mischel (1974). From this viewpoint, it is interesting to observe that rarely, if ever, do we ask our clients, or for that matter, even less frequently our experimental subjects, how they would characterize their own thoughts and feelings. What are the summary terms and concepts that our clients employ to describe their own self-statements and images.
4. Summary In summary, a three-stage theory of behavioral change is offered. The client must first become an observer of his thoughts, feelings, and behaviors by means of heightened awareness. This process is facilitated by means of a conceptualization or translation process that
TOWARD A COGNITIVE THEORY OF SELF-CONTROL
247
evolves over the course of therapy. The process of self-observation lays the foundation for the client to emit incompatible thoughts and behaviors, which constitute the second stage of change. The third stage, which determines the persistence and generalization of treatment effects, involves the nature and content of the client's internal dialogue and images about the behavior change. Although these three stages may be viewed as occurring sequentially, they often overlap in a continual process of change. The reader may be concerned that one of the things that already characterizes some clients prior to therapy is a heightened awareness, a self-preoccupation, and general egocentrism, e.g., in the case of the obsessive. Such a description of clients may be accurate, and indeed, the therapist would incorporate such behaviors into the conceptualization process. But what the client attends to and says to himself about his behavior prior to therapy are qualitatively and quantitatively different than what he will observe and how he will appraise his behavior following therapy. Prior to therapy the obsessive's selfperception is likely to be delimited and repetitive, eliciting a sense of helplessness and despair. As a result of therapy, the client will come to view his obsessive ideation differently, with the consequence of an increased sense of resourcefulness, of control of his own behavior. Whether he views his obsessions as manifest symptoms of underlying conflicts, or instrumental acts to control anxiety, or interpersonal ploys to influence others, etc., will depend on his therapist's orientation. Behavior is sufficiently multi determined that it is likely that there will be aspects of each of the above explanations contributing to the obsessional style. Moreover, both client and therapist have the capability of entertaining anyone of these conceptualizations that can be offered to explain the client's obsessive style. The important ingredient is that the translation process provides the stimulus for the other stages of change to occur. The theory thus far provides an overview of the change process. The therapeutic picture that is offered is that one can treat a client's verbal utterances that occur between the thought and the act and by doing so reach backward to change the thoughts and reach forward to modify the behavior. Through monitoring and modifying his thinking (i.e., self-statements and images), a client can effectively change his behavior.
248
DONALD MEICHENBAUM
B. How Does Behavior Change through Internal Dialogue? But how exactly does changing the client's internal dialogue lead to behavior change? The answer to this question is surely complex. Indeed, there are likely to be a number of answers, given the heterogeneity of private speech. In part, if will depend upon which popUlation and which behaviors one is trying to change. The psychological mechanisms involved in altering the internal dialogue of hyperactive children, versus adult psychotics, versus adult neurotics, may be quite different. We may find value in developing a number of minitheories of verbal control of behavior.
1. From Related Investigations One place to begin is exploring interpersonal instructions in order to ascertain if they bear any applicability to intrapersonal self-instructions. Several investigators (e.g., Gagne, 1964; Marlatt, 1972; Simkins, 1963; Sutcliffe, 1972) have speculated about the role of interpersonal instructions in controlling behavior. They emphasized both the instigational and directive functions of such instructions in controlling behavior: instructions both initiate or facilitate performance in a general sense and direct attention to stimulus conditions and to specified performance requirements in a task. Moreover, instructions control extraneous behaviors by directing subjects not to engage in certain responses. Gagne (1964), working within a problem-solving framework, has viewed instructions as serving the following functions: (1) motivating the subject by eliciting an achievement set; (2) helping him identify the criterion performance and the salient parts of the stimulus situation; (3) aiding recall of relevant subordinate performance capabilities necessary to the task; and (4) channeling thinking in terms of taskrelevant hypotheses and controlling extraneous thoughts and behaviors. In this way, instructions provide the subject with a rule of principle by which he can mediate his behavior. In describing the role of overt self-verbalizations, or self-instructions, in a problem-solving task, McKinney (1973) offered the following list of functions: the overt self-instructions (1) increase the distinctiveness of the stimulus attributes; (2) direct the subject's attention to
TOWARD A COGNITIVE THEORY OF SELF-CONTROL
249
the relevant dimensions; (3) assist the subject in formulating a series of hypotheses; and (4) maintain information in short-term memory. The similarity between the Gagne and McKinney lists of psychological functions for inter- and intrapersonal instructions is noteworthy. This leads to the first, rather obvious, hypothesis, that selfinstructions operate in a similar fashion to interpersonal instructions. As noted earlier, Vygotsky (1962) and Luria (1961) have theorized that developmentally the child comes to exercise verbal control of his behavior by incorporating adults' instructions. To quote Vygotsky (as cited by Zaporozhets and Elkonin, 1971): Apparently, egocentric speech, besides having a purely expressive function and a function of discharge, besides merely accompanying the child's activity, very readily becomes a means of thinking in its own sense, i.e., it begins to fulfill the function of formulating a plan for the solution of a problem emerging in the course of behavior. (p. 124)
Thus it is posited that one aspect of the functioning of selfinstructions depends upon their being analogous to interpersonal instructions and serving many of the same purposes. From a different vantage point, Janis (1968) discussed the role of appraisal, or what we would call internal dialogue (i.e., self-statements and images), in the handling of stress. Janis suggested that such cognitive appraisal includes the subject's (1) making plans for coping with a number of different contingencies; (2) attempting to reassure himself; (3) warding off disturbing thoughts, and (4) noting the level of his shortcomings that becomes a cue for actions. Janis suggested that such appraisal responses also raise the client's concern for others and help him identify with a group. In some general sense the "work of worrying," or emitting self-statements, increases the likelihood that the subject will be an observer of his situation and of his ongoing behavior. By thus self-instructing, the client becomes less egocentric and develops the ability to take the point of view of others or, in Piagetian terms, the ability to decenter perceptions (Looft, 1972). Thus by training a client to talk to himself, one reduces the likelihood that he will see events from only his own perspective. Common to both Gagne's and Janis's formulations is the role of self-instructions in directing the subject'S attention to either specific aspects of the task (i.e., in the problem-solving situation) or to the viewpoints of others (i.e., in the stress situation). Another postulate of
250
DONALD MEICHENBAUM
the present theory is, then, that changing the client's internal dialogue effects behavior change specifically by causing differential attentional behaviors. The literature on the effect of instructional sets of autonomic functioning indicates that changing the client's style of self-instructions can have directive physiological effects (Barber, 1965; May and Johnson, 1973; Platonov, 1959; Schwartz, 1971; Sternbach, 1964; Zimbardo, 1969). Cognitive activity has been suggested as a mediational factor (Le., facilitator or inhibitor) in operant, autonomic conditioning. Katkin and Murray (1968) proposed that an internal source of stimulation, rather than the external, experimenter-controlled reinforcers, may be controlling the autonomic responses. The subject may be involved in arousing or inhibiting subvocal activity (thinking), which produces a previously conditioned autonomic response. An illustration of the role of cognitive set is the work on emotion by Schachter (1966), who provided evidence for the important role that the client's restructuring of a situation plays in mediating behavior. In our own research the clients, following cognitive-behavior modification treatment, came to label their physiological arousal as facilitative rather than debilitative (Meichenbaum, 1972; Wine, 1970). Sweaty palms, increased heart and respiratory rates, and muscular tension now became allies, bell ringers, cues to use the coping techniques for which they had been trained. The physiological arousal that the client had previously labeled as totally debilitating anxiety and fear, the harbinger of further behavior deterioration leading to feelings of helplessness, was now relabeled as eagerness to demonstrate his competence, as a desire to get on with a task, and as a sign to cope. The result was a change to a sense of "learned resourcefulness," replacing a sense of "learned helplessness." In other words, the client learns to respond to the same physiological cues when they do arise with different cognitions: originally he entertained cognitions that mediated further autonomic arousal (e.g., "I'm really nervous; I'm sweating; others will see it; I can't handle this"); after treatment, his cognitions have a coping orientation and move the focus away from his arousal toward response alternatives. This shift in cognitions in itself may mediate a shift in autonomic functioning. The present theory postulates that it is not the physiological arousal per se that is debilitating, but rather what the client says to himself about that arousal that determines his eventual reactions. The
TOWARD A COGNITIVE THEORY OF SELF-CONTROL
251
distinction between thought and feeling becomes obscure, for each thought has an affective component and each feeling has a cognitive component. Thus the arbitrary distinction between "feeling" and "cognitive" therapies is misguided. "Touchy-feely" therapy, to use the cliche, has as much cognitive restructuring occurring-that is, if it leads to behavioral change--as semantic therapies have affective alterations. Our original question was "How does changing the client's internal dialogue lead to behavior change?" In order to approach our question we have briefly covered such diverse research areas as problem solving, appraisal or stress, and autonomic conditioning. From these vantage points we have been able to discern several of the factors contributing to the effect that cognitive restructuring has on a client's behavior: self-instructions playa direct role, analogous to that served by interpersonal instructions; self-instructions and images affect behavior through influencing attentional direction; and they influence a client's interpretation and experience of physiological state. Perhaps another useful approach that should not be overlooked is to be found in an examination of the conditions under which a client's self-instructions do not enhance self-control nor lead to behavioral change and autonomic conditioning. (Consider New Year's resolutions. )
2. When Self-Instructions Fail "Tous les jours, a tous points de vue, je vais de mieux en mieux." In English, "Day by day, in every way, I'm getting better and better." This is what Emil Coue, the Fren<1h psychiatrist, enjoined his patient to say to themselves in order to improve themselves (Coue, 1922). The banner of Coue's approach of fostering the power of positive thinking has been take~ up by a number of individuals writing for the general public, including Norman Vincent Peale (1960), Dale Carnegie (1948), and W. Clement Stone (1962). The use of a formula such as Coue's to be repeated daily, or in times of need, has found widespread appeal in the lay literature. It has been argued that such a "general formula leaves every mind free to unfold and develop in the manner most natural to itself" (Brook, 1922). However encouraging, the use of a formula, or "psy-
252
DONALD MEICHENBAUM
chological litany," tends to lead to rote repetition and emotionless patter that has been found ineffective as a coping tool (Meichenbaum and Cameron, 1973b). They are, in effect, self-instructions that fail. Several reasons for this failure have been logically and experimentally derived. In the first place, the client's self-instructions may be too general or too broad, not specific enough nor sufficiently individualized. For instance, the client may resolve that he will give up smoking but not specify when and how or what incompatible responses he will substitute in its place. Just why such a resolution would be likely to fail can be seen from the analogous situation of a person attempting to become a vegetarian without specifically substituting anything for the meat in his diet. Moreover, the client's general self-instructions are likely to be insensitive to situational conditions and are unlikely to incorporate such contingencies or rewards and punishments as would strengthen appropriate, and weaken inappropriate, responses to his self-instruction. Instead of the client's making only a general self-statement (a la Coue), it seems necessary for him to use a more detailed set of selfinstructions, specifying the specific desirable responses he will emit, under what conditions he will emit them, and the set of reinforcement contingencies he will employ. A general, self-motivating statement may enhance the functioning of such a detailed plan, but the detail itself is likely to lead to more self-control than generalized prohibitions or exhortations, repeated mantra-like. Two other, related examples illustrate conditions under which self-instructional rehearsal may not work. First, in our research on reducing fears by modifying what clients say to themselves (Meichenbaum and Cameron, 1973a), we repeatedly compared a treatment group who received only self-instructional rehearsal with a self-instructional rehearsal group who also received application training in the form of exposure to electric shock. The results indicated that self-instructional rehearsal was a necessary but not a sufficient condition for the elimination of fears. Having clients merely cognitively rehearse the self-instructions, saying that they could overcome their fears, did not lead to consistent significant behavioral and affective change. In fact, when the phobic clients who had received only self-instructional rehearsal confronted the phobic object following treatment, they initially reported minimal anxiety, indicating that they were calm and in control. However, when the demands to handle the phobic objects
TOWARD A COGNITIVE THEORY OF SELF-CONTROL
253
increased, their self-reports of anxiety precipitously rose with the consequence of a rearousal of their fears. The initial bravado which followed from mere self-instructional training gave way when the task demands increased. In comparison, those phobic clients who rehearsed the self-controlling self-statements and who then had an opportunity to use them in confronting a stress (e.g., electric shock) significantly reduced their fears following treatment. Thus the mere rehearsal of self-instructions without the opportunity for application training on tasks other than the criterion tasks will be likely to result in those self-instructions' exerting a minimal selfcontrolling influence. Saying the "right" things to yourself may not be a sufficient condition for change. One may have to "try out" these self-statements gradually in real situations that are similar to the criterion task. In addition to our earlier excursion into the literature from related investigations, this brief review of some self-instructions that fail had indicated another field of investigation that may be productively explored in an attempt to understand better the role of alterations of internal dialogue in effecting behavioral change. Some of the complexity involved in an understanding of the psychological mechanisms underlying behavioral change is further illustrated when we learn that there are approximately 15 different ways to alter our clients' imagery in therapy (Le., see Beck, 1970b; Singer, 1974). We can proceed by systematically analyzing each of these procedures, or rather proceed by searching for a set of common principles that underlie these different approaches. The present cognitive theory of self-control was designed to help us pursue the latter course.
IV.
SUMMARY
The alternative ways the therapist may view his clients' cognitions were reviewed. These included cognitions as behaviors, automatic thoughts, and thus part of the response chain, reflections of cognitive styles and faulty belief systems, inadequate problem-solving and coping skills, and defense mechanisms. Each conceptualization leads to different therapeutic interventions. However, it was suggested that even though the therapy procedures differ in emphasis and techniques, clients go through a similar cognitive process in
254
DONALD MEICHENBAUM
achieving behavior change. A cognitive theory of self-control, which includes a three-stage process, was offered to explain the behavior change process. Briefly, the theory proposed that the client must first become an observer of his thoughts, feelings, and behaviors by means of heightened awareness. This process is facilitated by means of a conceptualization process, or translation of how the client views his problem, which occurs over the course of therapy. The language system of the conceptualization process will be influenced by the therapist's orientation. The process of self-observation acts as a stimulus for the client to emit incompatible thoughts and behaviors, constituting the second stage of the change process. The third stage, which determines the persistence and generalization of treatment effects, involves the nature and content of the client's internal dialogue and images about behavior change. The chapter also addressed the question, "How does changing the client's internal dialogue lead to behavior change?" Elements of the answer came from the literature on problem solving, appraisal and stress, and antonomic conditioning. It was theorized that (1) selfinstructions playa direct role in changing behavior, analogous to that served by interpersonal instructions; (2) self-instructions and images affect behavior through influencing attentional direction; and (3) they also influence a client's interpretation and experience of his physiological state. An examination of the conditions under which self-instructions fail further elucidated the issues. A cognitive-behavior modification treatment approach was proposed with most of the emphasis on the psychological mechanisms that underlie the hyphen.
ACKNOWLEDGMENTS
Portions of this paper were presented at the eighth annual meeting of the Association for the Advancement of Behavior Therapy, Chicago, Illinois, November, 1974. The author is grateful to Myles Genest and Roy Cameron for their helpful editorial comments. The research reported in the paper was supported by grants from the Canada Council (#S73-0024-V1) and the Ontario Ministry of Education.
TOWARD A COGNITIVE THEORY OF SELF-CONTROL
255
REFERENCES ALLPORT, G. The individual and his religion. New York: MacMillan Company, 1950. ASHEM, B., AND DONNER, L. Covert sensitization with alcoholics: A controlled replication. Behaviour Research and Therapy, 1968,6, 7-12. BAKER, J. Reason versus reinforcement in behavior modification. Unpublished doctoral dissertation, University of Illinois, 1966. BANDURA, A. Principles of behavior modification. New York: Holt, Rinehart, & Winston, 1969. BANDURA, A. Behavior theory and models of man. Paper presented at APA meeting in New Orleans, 1974. BARBER, T. Physiological effects of "hypnotic suggestions": A critical review of recent literature (1960-1964). Psychological Bulletin, 1965,63, 201-222. BECK, A. Cognitive therapy: Nature and relation to behavior therapy. Behavior Therapy, 1970a, 1, 184-200. BECK, A. Role of fantasies in psychotherapy and psychopathology. Journal of Nervous and Mental Disease, 1970b, 150, 3-17. BECK, A. Cognition, affect and psychopathology. Archives of General Psychiatry, 1971,24, 495-500. BEM, S. Verbal self-control: The establishment of effective self-instruction. Journal of Experimental Psychology, 1967,74, 485-491. BERGIN, A. A self-regulation technique for impulse control disorders. Psychotherapy: Theory, research and practice, 1967,6, 113-118. BERGIN, A. Cognitive therapy and behavior therapy: Foci for a multidimensional approach to treatment. Behavior Therapy, 1970,1, 205-212. BLACKWOOD, R. The operant conditioning of verbally mediated self-control in the classroom. Journal of School Psychology, 1970,8, 251-258. BRAFF, D., AND BECK, A. Thinking disorder in depression. Archives of General Psychiatry, 1974,31, 45Cr462. BROOKS, C. H. The practice of autosuggestion by the method of Emile Coue. In E. COUE AND C. H. BROOKS (Eds.), Better and better every day. London: Unwin, 1960. CARNEGIE, D. How to stop worrying and start living. New York: Simon & Schuster, 1948. COPEMAN, C. Aversive counterconditioning and social retraining: A learning theory approach to drug rehabilitation. Unpublished doctoral dissertation, S.U.N.Y. at Stony Brook, 1973. Com" E. The practice of autosuggestion. New York: Doubleday, 1922. DE CHARMS, R. Personal causation: The internal affective determinants of behavior. New York: Academic Press, 1968. DECI, E. The effects of externally mediated rewards on intrinsic motivation. Journal of Personality and Social Psychology, 1971,18, 105-115. DEMBER, W. Motivation and the cognitive revolution. American Psychologist, 1974, 29, 161-168. DENNY, D. Modeling effects upon conceptual style and cognitive tempo. Child Development, 1972,43, 105-419. DILORETO, A. Comparative psychotherapy: An experimental analysis. Chicago: AldineAtherton, 1971. DOLLARD, J., AND MILLER, N. Personality and Psychotherapy. New York: McGraw-Hill, 1950. D'ZURILLA, T., AND GOLDFRlED, M., Problem solving and behavior modification. Journal of Abnormal Psychology, 1971,78, 107-126.
256
DONALD MEICHENBAUM
ELLIS, A. Reason and emotion in psychotherapy. New York: Lyle Stuart Press, 1961. FRANK, J. Persuasion and healing. Baltimore: Johns Hopkins Press, 1961. FREYBERG, J. Increasing the imaginative play of urban disadvantaged kindergarten children through systematic training. In J. SINGER (Ed.), The child's world of make believe. New York: Academic Press, 1973. GAGNE, R. Problem solving. In A. MELTON (Ed.), Categories of human learning. New York: Academic Press, 1964. GOLDFRIED, M. Systematic desensitization as training in self-contro!. Journal of Consulting and Clinical Psychology, 1971,37,228-234. GOLDFRIED, M., DECENTECEO, E., AND WEINBERG, L. Systematic rational restructuring as a self-control technique. Behavior Therapy, 1974,5, 247-254. GOLDFRIED, M., AND GOLDFRIED, A. Cognitive change methods. In F. KANFER AND A. GOLDSTEIN (Eds.), Helping people change. New York: Pergamon Press, 1975. GOLDFRIED, M., AND TRIER, C. Effectiveness of relaxation as an active coping skill. Journal of Abnormal Psychology, 1974,83,348-355. HANEL, J. Der Einfluss eines Motivanderungsprogramms auf schulleistungschwachmiszerfolgsmotivierte Grundschuler der 4. Klasse. Psycho!. Institue der Ruhr-Universitat: Unveroffentlichte Diplomarbeit, 1974. HARTIG, M., AND KANFER, F. The role of verbal self-instructions in children's resistance to temptation. Journal of Personality and Social Psychology, 1973,25, 259-267. HECKHAUSEN, H. How to improve poor motivation in students. Paper presented at the 81th International Congress of Applied Psychology, Montreal, August, 1974. HOMME, L. Perspectives in psychology: Control of coverants, the operants of the mind. Psychological Record, 1965,15, 501-511. JAHODA, G. The psychology of superstition. Baltimore: Penguin, 1969. JANIS, I. Psychological stress. New York: Wiley, 1958. JANIS, I. Human reactions of stress. In E. BORGOTTA AND W. LAMBERT (Eds.), Handbook of personality theory and research. New York: Rand McNally, 1968. KANFER, F., AND SASLOW, G. Behavioral diagnosis. In C. FRANKS (Ed.), Behavior therapy: Appraisal and status. New York: McGraw-Hill, 1969. KARST, T., AND TREXLER, L. Initial study using fixed-role and rational-emotive therapy in treating public-speaking anxiety. Journal of Consulting and Clinical Psychology, 1970,34, 360-366. KATKIN, E., AND MURRAY, E. Instrumental conditioning of autonomically mediated behavior: Theoretical and methodological issues. Psychological Bulletin, 1968,70, 5268. KAZDIN, A. Covert modeling and the reduction of avoidance behavior. Journal of Abnormal Psychology, 1973,81,87-95: KAZDIN, A. Covert modeling, model similarity, and reduction of avoidance behavior. Behavior Therapy, 1974a, 5, 325-340. KAZDIN, A. The effect of model identity and fear-relevant similarity on covert modeling. Behavior Therapy, 1974b,5, 624-636. KIESLAR, D. Some myths of psychotherapy research and the search for a paradigm. Psychological Record, 1966, 65, 110-136. KIFER, R., LEWIS, M., GREEN, D., AND PHILLIPS, E., The S.O.C.S. model: Training predelinquent youths and their parents in negotiation responses to conflict situations. Paper presented at the American Psychological Association meeting, Montreal, August, 1973.
TOWARD A COGNITIVE THEORY OF SELF-CONTROL
257
KIMBLE, G., AND PERLMUTIER, L. The problem of volition. Psychological Review, 1970, 77, 361-384. LANGER, E., JANIS, 1., AND WOLFER, J. Effects of cognitive devise and preparatory information on psychological stress in surgical patients. Unpublished manuscript, Yale University, 1973. LAZARUS, A. Behavior therapy and beyond. New York: McGraw-Hill, 1972. The psychology of preschool children. Cambridge: MIT Press, 1971. LOOF!", W. Egocentrism and social interaction across the life span. Psychological Bulletin, 1972, 78, 7~9i LURIA, A. The role of speech in the regulation of normal and abnormal behavior. New York: Liveright, 1961. MAHONEY, M. Cognition and behavior modification. Cambridge, Mass.: Ballinger Publishing Co., 1974. MARKS, 1., BOULOUGOURIS, J., AND MARSET, P. Flooding versus desensitization in the treatment of phobic patients. British Journal of Psychiatry, 1971, 119, 35~375. MARLATT, G. Task structure and the experimental modification of verbal behavior. Psychological Bulletin, 1972, 78, 335--350. MARMOR, J. The psychodynamics of realistic worry. Psychoanalysis and Social Science, 1958,5, 115--163. MARSHALL, W., POLGRIN, D., AND BOUTILIER, J. Reinforcement contingencies in covert reinforcement. Unpublished manuscript, Kingston Psychiatric Hospital, 1974. MAY, J., AND JOHNSON, H. Physiological activity to internally elicited arousal and inhibitory thoughts. Journal of Abnormal Psychology, 1973,82, 239-245. MCGUIRE, M., AND SIFNEOS, P. Problem solving in psychotherapy. Psychiatric Quarterly, 1970,44, 667-673. McKINNEY, J. A developmental study of the effects of hypothesis verbalizations and memory load on concept attainment. Unpublished manuscript, University of North Carolina, Chapel Hill, 1973. MEICHENBAUM, D. Examination of model characteristics in reducing avoidance behavior. Journal of Personality and Social Psychology, 1971,17,298-307. MEICHENBAUM, D. Cognitive modification of test anxious college students. Journal of Consulting and Clinical Psychology, 1972, 39, 370-380. MEICHENBAUM, D. Cognitive factors in behavior modification: Modifying what clients say to themselves. In C. FRANKS AND T. WILSON (Eds.), Annual review of behavior therapy: Theory and practice. New York: Bruner/Maze!, 1973. MEICHENBAUM, D. Cognitive behavior modification. Morristown, N. J.: General Learning Press, 1974a. MEICHENBAUM, D. Self-instructional training: A cognitive prosthesis for the aged. Human Development, 1974b, 17, 27~280. MEICHENBAUM, D. Therapist manual for cognitive behavior modification. Unpublished manuscript, University of Waterloo, 1974c. MEICHENBAUM, D. A self-instructional approach to stress management: A proposal for stress inoculation training. In C. SPEILBERGER AND 1. SARASON (Eds.), Stress and anxiety in modern life. New York: Winston and Sons, 1975a. MEICHENBAUM, D. Self-instructional methods. In F. KANFER AND A. GOLDSTEIN (Eds.), Helping people change. New York: Pergamon Press, 1975b. MEICHENBAUM, D., AND CAMERON, R. An examination of cognitive and contingency variables in anxiety relief procedures. Unpublished manuscript, University of Waterloo, 1973a.
258
DONALD MEICHENBAUM
MEICHENBAUM, D., AND CAMERON, R. Stress inoculation: A skills training approach to anxiety management. Unpublished manuscript, University of Waterloo, 1973b. MEICHENBAUM, D., AND CAMERON, R. The clinical potential of modifying what clients say to themselves. Psychotherapy: Theory, Research and Practice, 1974,11, 103-117. MEICHENBAUM, D., GILMORE, B., AND FEDORAVIClUS, A. Group insight vs. group desensitization in treating speech anxiety. Journal of Consulting and Clinical Psychology, 1971,36, 410-421. MEICHENBAUM, D., AND GOODMAN, J. Training impulsive children to talk to themselves: A means of developing self-control. Journal of Abnormal Psychology, 1971,77, 115126. MEICHENBAUM, D., TURK, D., AND BURSTEIN, S. The nature of coping with stress. Paper presented at NATO conference on Stress and Anxiety, Korsor, Denmark, 1975. MElZACK, R. The puzzle of pain. Ringwood, Australia: Penguin Books, 1973. MElZACK, R., AND WALL, P. On the nature of cutaneous sensory mechanisms. Brain, 1962,85, 331-356. MISCHEL, W. Toward a cognitive social learning reconceptualization of personality.
Psychological Review, 1973,80, 252-283. MISCHEL, W. Processes in delay of gratification. In L. BERKownz (Ed.), Advances in experimental social psychology (Vol. 7). New York: Academic Press, 1974. MONAHAN, J., AND O'LEARY, D. Effects of self-instruction on rule-breaking behavior. Psychological Reports, 1971,79, 1059-1066. MURRAY, H. Explorations in personality. New York: Oxford Press, 1938. NOVACO, R. A treatment program for the management of anger through cognitive and relaxation controls. Unpublished doctoral dissertation, Indiana University, Bloomington, Indiana, 1974. PALKES, H., STEWART, M., AND FREEDMAN, J. Improvement in maze performance on hyperactive boys as a function of verbal training procedures. Journal of Special Education, 1972, 5, 337-342. PALKES, H., STEWART, M., AND KAHANA, B. Porteus performance of hyperactive boys after training in self-directed verbal commands. Child Development, 1968,39, 817826. PEALE, N. The power of positive thinking. Englewood Cliffs, No. J.: Prentice-Hall, 1960. PETERSON, D. The clinical study of social behavior. New York: Appelton-Century-Crofts, 1968. PITKIN, W. A short introduction to the history of human stupidity. London: Roworth & Co., 1935. PLATONOV, K. The word as a physiological and therapeutic factor. Moscow: Foreign Languages Publishing House, 1959. PLATT, J., AND SPIVACK, G. Problem-solving thinking of psychiatric patients. Journal of Consulting and Clinical Psychology, 1972a, 39, 148-151. PLATT, J., AND SPIVACK, G. Social competence and effective problem-solving thinking in psychiatric patients. Journal of Clinical Psychology, 1972b,28, 3-5. PREMACK, D. Mechanisms of self-control. In W. HUNT (Ed.), Learning and mechanisms of self-control in smoking. Chicago: Aldine, 1970. RICHARDSON, A. Mental practice: A review and discussion. Part 1. Research Quarterly, 1967a, 38, 95-107. RICHARDSON, A. Mental practice: A review and discussion. Part II. Research Quarterly, 1967b, 38, 263-273. RlDBERG, E., PARKE, R., AND HETHERINGTON, M. Modification of impulsive and reflective
TOWARD A COGNITIVE THEORY OF SELF-CONTROL
259
cognitive styles through observation of film mediating models. Developmental
Psychology, 1971,5, 369-377. SACHS, L., AND INGRAM, G. Covert sensitization as a treatment of weight control. Psychological Reports, 1972,30, 971-974. SALlZ, E., AND JOHNSON, J. Training for thematic-fantasy play in culturally disadvantaged children. Journal of Educational Psychology, 1974, 66, 62:>-630. SARBIN, T. Imagining as muted role-taking: A historical-linguistic analysis. In P. SHEEHAN (Ed.), The function and nature of imagery. New York: Academic Press, 1972. SCHACIITER, S. The interaction of cognitive and physiological determinants of emotional state. In C. SPIELBERGER (Ed.), Anxiety and behavior. New York: Academic Press, 1966. SCHACIITER, S., AND SINGER, J. Cognitive, social, and physiological determinants of emotional state. Psychological Review, 1962,69, 379-399. SCHNEIDER, M. Turtle technique in the classroom. Teaching Exceptional Children, Fall, 1974, 22-24. SCHWARlZ, G. Cardiac responses to self-induced thoughts. Psychophysiology, 1971, 8, 462-467. SHAPIRO, D. Neurotic styles. New York: Basic Books, 1965. SHURE, M., AND SPIVACK, G. Means-ends thinking, adjustment and social class among elementary school-aged children. Journal of Consulting and Clinical Psychology, 1972, 38, 348--353. SHURE, M., SPIVACK, G., AND JAEGER, M. Problem-solving thinking and adjustment among disadvantaged preschool children. Child Development, 1971,42, 1791-1803. SIMKINS, L. Instructions as discriminative stimuli in verbal conditioning and awareness. Journal of Abnormal and Social Psychology, 1963,66, 21:>-219. SINGER, J. Imagery and daydream methods in psychotherapy. New York: Academic Press, 1974. SINGER, J., AND MCCRAVEN, V. Some characteristics of adult daydreaming. Journal of Psychology, 1961,51, 151-164. SKINNER, B. F. Science and human behavior. New York: MacMillan, 1953. SPIVACK, G., AND LEVINE, M. Self-regulation in acting-out and normal adolescents. Report M-4351. Washington, D.C. National Institutes of Health, 1963. SPIVACK, G., AND SHURE, M. Social adjustment of young children: A cognitive approach to solving real-life problems. San Franscisco: Jossey-Bass, 1974. STEFFY, R., MEICHENBAUM, D., AND BEST, A. Aversive and cognitive factors in the modification of smoking behavior. Behaviour Research and Therapy, 1970, 8, 115125. STEINER, 1. Perceived freedom. In L. BERKOWITZ (Ed.), Advances in experimental social psychology. New York: Academic Press, 1970. STERNBACH, R. The effects of instructional sets on autonomic activity. Psychophysiology, 1964, 1, 67-72. STONE, W. C. The Success System that Never Fails. Englewood Cliffs, N. J.: Prentice Hall, 1962 STRUPP, H. Specific vs. nonspecific factors in psychology and the problem of control. Archives of General Psychology, 1970,23, 393-401. SUINN, R., AND RICHARDSON, F. Anxiety management training. A nonspecific behavior therapy program for anxiety control. Behavior Therapy, 1971,2,498-510. SUTCLIFFE, J. On the role of "instructions to the subject" in psychological experiments. American Psychologist, 1972,27, 755-758.
260
DONALD MEICHENBAUM
TORI, c., AND WORRELL, 1. Reduction of human avoidant behavior: A comparison of counterconditioning, expectancy, and cognitive information approaches. Journal of Consulting and Clinical Psychology, 1973,41, 269-278. TREXLER, 1., AND KARST, T. Rational-emotive therapy, placebo, and no treatment effects on public-speaking anxiety. Journal of Abnormal Psychology, 1972,79, 6(k;7. TURK, D. Cognitive control of pain: A skills training approach. Unpublished manuscript, University of Waterloo, 1975. VYGOTSKY,1. Thought and Language. Cambridge, Mass.: MIT Press, 1962. WINE, J. Investigations of an attentional interpretation of test anxiety. Unpublished doctoral dissertation, University of Waterloo, Waterloo, Ontario, 1970. ZAPOROZHETS, A., AND ELKONIN, D. The psychology of preschool children. Cambridge: MIT Press, 1971. ZIMBARDO, P. The Cognitive Control of Motivation. Glenview, Ill.: Scott Foresman, 1969.
7 Physiological and Cognitive Processes in the Regulation of Anxiety THOMAS
D.
BORKOVEC
For the past five years we have been engaged in a program of research whose ultimate goal has been the development and evaluation of therapeutic methods for reducing anxiety. A basic assumption underlying our work has been that the successful evolution of such strategies will be facilitated by advances in our knowledge about the nature of anxiety itself. Consequently the majority of the research has attempted to identify basic conditions (environmental and subject) that serve to maintain or reduce the anxiety response. In the present chapter I would like to share with you a descriptive model of anxiety process that has guided our research and studies from our program that relate to portions of that model. As will become clear later on, the central focus emerging from our work has been the role of physiological arousal, cognitive processes, their interaction, and the importance of individual differences in those variables in determining the maintenance and the reduction of anxiety. An early, behavioristic account of anxiety suggested an essentially physiological model based on simple Pavlovian conditioning. Repeated pairings of a conditioned stimulus (CS) and an unconditioned stimulus (UCS) ultimately result in the elicitation of a conditioned response (CR) by the CS alone. In the aversive conditioning situation, that CR involves both autonomic and skeletal responses. Maintenance of the CS-CR relationship is viewed as a function of the frequency and intensity of CS-UCS presentations. Extinction is a simple matter of repetitious CS presentation in the absence of the UCS. The model was THOMAS
Iowa.
D.
BORKOVEC
Department of Psychology, University of Iowa, Iowa City,
261
262
THOMAS
D.
BORKOVEC
historically exemplified by Watson and Rayner's (1920) classic demonstration of conditioned fear acquisition and jones's (1924) program for the elimination of acquired fear. Although Skinner foreshadowed a two-factor theory of learning in his distinction between Type I and Type II learning, an operant view of anxiety is essentially a behavioral model and focuses on escape and avoidance behavior and the discriminative role of stimuli for those instrumental responses. Thus certain stimuli come to signal the imminent occurrence of aversive stimuli. Those stimuli also set the occasion for escape and avoidance responses which are negatively reinforced by the termination of those stimuli and the preclusion of the aversive stimulus. Avoidance responses occur with shorter latency and to more remote discriminative stimuli because of negative reinforcement and stimulus generalization. Since extinction requires nonoccurrence of the aversive stimulus in the presence of the discriminative stimuli, avoidance maintains unless the organism is somehow prevented from making the avoidance response. Attacking the monistic theories of Pavlov and Watson, Mowrer (1947) offered a two-factor theory of learning by suggesting that neither the principle of association nor the law of effect could by themselves provide a unified theory of learning. Rather, it seems necessary to assume that there are two basic learning processes: The process whereby the solutions to problems, i.e., ordinary "habits," are acquired; and the process whereby emotionalleaming, or "conditioning," takes place. (p. 114)
The origins of anxious behavior from this model involved a twoprocess sequence: (1) eS/noxious ues pairings result in the establishment of the es as a danger signal which elicits fear, an intervening variable; and (2) fear serves as an acquired drive motivating subsequent escape and avoidance behavior and as a reinforcer via the drivereducing consequence of the instrumental behavior. Anxiety maintains subsequent to acquisition since repeated es exposure is precluded by efficient avoidance responses to more remote cues. Mowrer's (1947) theory has stimulated an enormous quantity of research. While subsequent investigations have been equivocal in supporting his mediational hypothesis (Rescorla and Solomon, 1967), Mowrer's model has come to serve as the framework for many
·PHYSIOWGICAL AND COGNITIVE PROCESSES IN THE REGULATION OF ANXIETY
263
behavior therapy researchers and has been the point of departure for our own hypothesis. Specifically, we are concerned with the responseproduced feedback arising from the CR and stressed by Mowrer in establishing the drive properties of anxiety which motivate skeletal coping responses. Four additional considerations have contributed to our developing model, however. First, the important work of Schachter (1964) has indicated that specific emotional experience and behavior may be mediated by the subject's interpretation of his own physiological arousal and the extant environmental cues. The subsequent influence of his theory and research has been broadly felt, with numerous writers such as Beck, Epstein, Lazarus, Mandler, and Spielberger (d. Spielberger, 1972) emphasizing a central role for cognitive factors in current anxiety theory. We feel that such a cognitive model offers important dimensions to the role of CR feedback in human anxiety. Second, all of the previous learning models assume that extinction of anxious behavior will take place as long as feared stimuli are presented, avoidance precluded, and the UCS absent. In clinical work, however, it is clear that the anxious client often confronts the feared situation repeatedly in his daily life and yet anxiety maintains despite the absence of skill deficits, overt coping responses, or clear aversive consequences from the environment. Among several alternatives there are two positions regarding these observations that have interested us. Mowrer's model might be extended to allow the substitution of cognitive avoidance behavior for overt instrumental behavior. Alternately there may be conditions of nonreinforced CS exposure which do not lead to extinction. Very little direct data from anxious humans have been collected relevant to these hypotheses. Third, while the physiological component has long been central in behavioral definitions of anxiety, the investigation of physiological variables has been relatively ignored. Renewed emphasis on its functional role in theories of anxiety appears to be appropriate. Finally, although individual differences have entered into various general theories of anxiety (e.g., Spielberger's A-trait; Taylor'S Manifest Anxiety Scale), little attention has been devoted to the functional importance of those differences in the behavior therapy literature. The absence of that consideration has contributed to a great deal of ambiguity regarding the theoretical mechanisms both of treatment techniques and of the anxiety response itself.
264
THOMAS
D.
BORKOVEC
1. A DESCRIPTIVE MODEL OF ANXIETY PROCESS The focus of the descriptive model emerging from these considerations is on maintenance and reduction processes. Issues regarding the origins of anxiety are relatively ignored. The variables included in the framework are schematically represented in Table 1 and described by the series of working assumptions presented below. The emphasis of the model is on (1) the sequence of events that may be elicited at any given moment upon the presentation of an anxiety-eliciting stimulus; (2) the maintaining or reducing effects that any particular sequence of events has upon response to the subsequent recurrence of the feared stimulus; and (3) individual differences in the functional relevance of any particular component of the sequence.
A. Current Stimulus Conditions In terms of current eliciting conditions, measured anxiety is considered to be a function of external and/or internal cues. External cues include aspects of the environment which, because of the past history of learning, produce fear responses. Internal cues include verbal and nonverbal images, physiological activity, and proprioceptive or perceptual feedback from skeletal behavior. Each of these internal cues may, depending on its history, serve conditional or discriminative functions (or both) for subsequent fear components. While either external or internal cues may be separately capable of eliciting anxious behavior, observed anxiety reactions are most often a function of an interaction of both sets of cues: if a fear stimulus is presented, internal cues may be elicited; if diffuse arousal occurs, the organism may search its environment for external cues. Although the influence of future events on the development of the anxiety response sequence has traditionally focused on the external cues (e.g., additional, periodic ues presentations and stimulus-generalization processes), our concern has been primarily with the effect of history on the internal cue component of this interaction. Finally, individuals are assumed to differ, because of differences in learning history or genetic consitution, in terms of the relative functional importance of external and internal cues and their interaction in determining responses to feared stimuli.
Current stimulus conditions
+
1.
Immediate anxiety reaction
~b. Cognition Lc. Overt behavior
Subsequent maintaining and reducing reactions
Intervention strategies
Cognition ~ 3. Overt behavior----.J
.!I----.--t2.
Internal fear cues Autonomic arousal Verbal and nonverbal images Proprioception from overt behavior
Past - { history
External fear cues
1
A Descriptive Model of Anxiety Process
TABLE
8;
N
~ §
o>rJ
Z
~
gr
m
'"0
~ ~ '"gJ '"z
~
no
~
8
Cl
is
266
THOMAS
D.
BORKOVEC
The origins of a fear response to either set of cues may reside in classically conditioned autonomic arousal; observational learning of autonomic, self-report, or avoidance behavior; cultural role-learning and direct reinforcement of reports and behavioral display of fear; or a combination of these factors. As mentioned earlier, the issue of origin has not been our focus. It is important only to suggest that the recurrence of these various events may playa role in strengthening the functional effects of external and internal fear cues.
B. The Immediate Anxiety Reaction Our working definition of anxiety is in partial agreement with Spielberger (1966) and Paul (1969b) in suggesting that anxiety may be conceptualized as a label denoting a complex pattern of responses characterized by subjective feelings of apprehension and tension and the occurrence of physiological arousal. However, we would add behavioral manifestations of arousal (e.g., facial grimace and trembling hands) and avoidance behavior to that definition. While there would be little argument regarding behavioral arousal effects, few definitions since Watson would include avoidance behavior. Such coping reactions are reasonably considered to be anxiety-elicited and a consequence, therefore, of anxiety itself. While this is probably the frequent case, it is important to allow for a notion of skeletal avoidance behavior not mediated by the other presumed components of anxiety. For example, it may be possible that an "avoidance" behavior is learned via direct positive reinforcement. The individual may subsequently learn to label that behavior as fear, even though no clear CR is elicited by the stimulus situation. One might argue that such behavior should not be considered within the realm of fear research. We would disagree, if only because many of the fear behaviors investigated in the behavior-therapy literature may be a function of exactly these conditions. Second, while various investigators have attempted to distinguish between fear and anxiety at both the theoretical and empirical level, we are in agreement with Spielberger that the distinction is meaningless unless the response patterns of the two emotions differ and that little attention has yet been devoted
PHYSIOLOGICAL AND COGNITIVE PROCESSES IN THE REGULATION OF ANXIETY
267
to the identification of those differential patterns (Spielberger, 1972). Thus the terms will be used interchangeably in the present context. Given this working definition of the construct, anxiety will be operationally defined by the multiple measurement of three general response components: cognitive, overt behavioral, and physiological. Notice that these are gross· categories which may contain several subsets of response measures potentially reflecting several responses within a given category. There has been considerable concern with the empirical fact that these three measurement groups are not highly interrelated. Two reactions are often offered to this state of affairs: the low correlations are a consequence of dealing with subjects at the extreme and restricted end of the anxiety distribution, and improved measurement of each response group will ultimately reduce error variance in the instruments and therefore produce higher intercorrelations. Both reactions fail to recognize an extremely important point that is at the core of our conceptualization. In agreement with Lang (1968), anxiety must be considered to involve any or all of three separate but interacting response components. Cognitive behavior, motor behavior, and physiological reactions may be separately influenced by different environmental conditions at different points in time and may even obey different learning principles. Yet because of their potential interaction, changes in one response component due to the direct manipulation of its conditions may ultimately affect subsequent changes in the response of one or both of the remaining components. Most importantly, individuals differ in terms of the learning history associated with each response component, resulting in individual differences in the intensity and/or functional importance of the response from each component in reaction to a particular feared stimulus. Some individuals, for example, will report intense distress and display rapid avoidance when confronted with feared situations, but no evidence of increases in physiological arousal can be detected. Others may show such autonomic increases but differ in the degree to which they are aware of the arousal, the degree of avoidance behavior, the level of reported discomfort, etc. The separateness of responses considered to be evidence of anxiety (either by the researcher or by the anxious person) allows for such important individual differences, and their interaction will be hypothesized to account for maintaining reactions and for the development or reduction of anxiety over time.
268
THOMAS
D.
BORKOVEC
Unfortunately such individual differences have been largely ignored in the study of anxiety and its modification. It is our conviction that theoretical anxiety research and the development of self-regulatory techniques must focus on those differences. The importance of this assumption is magnified when we consider interactions among the response components and the impact of intervention or reduction strategies on the anxiety reaction.
C. Subsequent Maintaining and Reducing Reactions It is assumed that the three response components interact. That is, the occurrence of an immediate anxiety reaction in one or more components in response to external and/or internal cues may result in the occurrence of subsequent reactions in one or more of the components. Such reactions may be a function of subject characteristics of the individual (constitutional or learned behavior) or a function of environmental manipulations during the immediate anxiety response. The effect of any interaction may be either the maintenance of anxiety or its reduction upon subsequent exposure to the feared situation. Assuming three response components potentially involved in the immediate and subsequent reactions, there are nine possible relationships. These relationships are offered as hypotheses regarding the sequential interactions which may occur in any given case of anxiety. None of the relationships are considered -necessary or sufficient for anxiety maintenance or reduction. The contribution of anyone relationship to the anxiety process and the effects such a relationship may have on the anxiety process are hypothetical. Only future research can support the reasonableness of these relationships and their effects. While the assumptions have some basis in the theories and research of various investigators, the statement of the assumptions will not represent an extensive review of the evidence in support of those assumptions. In addition, the assumptions are stated in terms of maintenance conditions only. Reduction conditions are implied either in the sense that the maintaining reactions do not occur or that other reactions antagonistic to anxiety occur. The latter conditions will be dicussed more fully in the subsequent intervention section.
PHYSIOLOGICAL AND COGNITIVE PROCESSES IN THE REGULATION OF ANXIETY
269
la. The Occurrence of a Physiological Reaction to Fear Cues May Result in Subsequent Physiological Reactions That Contribute to the Maintenance of Anxiety.
The notion of anxiety maintenance subsequent to a single exposure is similar to Lader and Mathews's (1968) suggestion of a feedback loop among highly anxious individuals. Immediate arousal cues, as response-produced stimuli, may serve as conditional stimuli for other physiological reactions. Alternately, physiological reactions may be maintained even by the occurrence of nonfeared stimuli once arousal level has been elevated by initial fear cues. The possible influence of physiological reactions resulting in maintenance over repeated exposures is similar to Eysenck's (1968) theory of incubation. The autonomic CR may hypothetically be viewed as a noxious stimulus which under certain conditions serves to strengthen the CS-CR relationship despite the absence of a UCS. lb. Immediate Physiological Reactions to Feared Stimuli May Set the Occasion for Subsequent Cognitive Behaviors Which Serve to Maintain the Anxiety Response.
This relationship focuses most directly on the role of physiological cues in maintaining human anxiety. Mowrer (1947) hinted at this notion in referring to the "cognitive restructuring" that occurs as the subject undergoing aversive conditioning learns to fear not only the CS but also the entire experimental situation. The more recent and relevant conceptualization of this process is provided by Schachter (1964). The occurrence of diffuse arousal sets the occasion for search for environmental cues to explain the arousal. The resulting cognitive interpretation strongly influences the emotional reaction. lc. Physiological Reactions to Fear Cues May Set the Occasion for Subsequent Overt Behavior Which Maintains the Anxiety Response.
The state-dependent learning literature provides clear, general support for this interaction. This relationship, of course, was the focus of Mowrer's original two-process theory implicating a mediational role for the CR in controlling instrumental avoidance. Rescorla and Solo-
270
THOMAS
D.
BORKOVEC
mon's (1967) review found weak support for the mediational hypothesis. But their conclusion was only that CRs are not required for avoidance behavior. There is evidence that CRs may serve such a mediational role, and these authors readily admitted that avoidance may be "mediated by a complex of CRs, both autonomic and skeletal; no one of these may be necessary for operant behavior, but each contributes to that behavior" (p. 163). It is sufficient for our purposes that there is some evidence for CR mediation (e.g., Gantt and Dykman, 1957; Black, 1959), not as a necessary condition but as a potentially sufficient contributor to avoidance behavior.
2a. Presentation of Fear Cues May Immediately Result in Thoughts and Images Which Elicit Subsequent Physiological Responses Contributing to Anxiety Maintenance. There is ample evidence that imagery is capable of such CR elicitation (d. Mathews, 1971). As long as cognitive events of an anxiety-eliciting nature continue to occur, physiological arousal may be maintained at high levels.
2b. Thoughts and Images in Immediate Response to the Feared Situation May Result in Further Chains of Catastrophizing Self-Verbalizations Regarding the Feared Stimulus and the Individual's Ability to Cope with It. Ellis's (1958) theory of neurotic behavior emphasizes this relationship, while Meichenbaum (this volume) has presented empirical evidence supporting the role of self-verbalizations in anxiety.
2c. Cognitive Response to Fear Cues Can Influence Subsequent Overt Behavior. Again, Meichenbaum's studies suggest that the presence or absence of self-instructions importantly contributes to problem behavior, while recent research on covert rehearsal indicates that cognitive instructions and imagery may be as powerful in improving performance as overt rehearsal (e.g., Cautela, Flannery, and Hanley, 1974; McFall and Twentyman, 1973).
PHYSIOLOGICAL AND COGNITIVE PROCESSES IN THE REGULATION OF ANXIETY
271
3a. Immediately Elicited Avoidance Behavior, Inept Coping Behavior, etc., May Serve as Conditional Cues for Increased Arousal.
The James-Lange theory of emotion represents an early, similar version of the relationship. In support of that relationship, Black (1959) found maximum heart-rate reaction to occur following the avoidance response and to maintain for some time. This finding agrees with clinical observations that some phobic individuals do not report intense physiological reactions until after they have removed themselves from the feared situation and have found safety, at which time large (and aversive) reactions occur. 3b. Behavioral Reactions to a Feared Situation May Serve to Elicit Negative, Self-Evaluative Cognitions.
This general relationship is best represented currently by Bem (1972) who postulates that: Individuals come to "know" their own attitudes, emotions, and other internal states partially by inferring them from observations of their own overt behavior andJor the circumstances in which this behavior occurs .... To the extent that internal cues are weak, ambiguous or uninterpretable, the individual is functionally in the same position as an outside observer, an observer who must necessarily rely upon those same external cues to infer the individual's inner states. (p. 2)
The Bem hypothesis may be viewed in our context as a complement to Schachter's consideration of internal states and their contribution to the subsequent cognitive interpretations mediating anxiety. 3c. Immediate Fear Behavior May Set the Occasion for Continued Fear Behavior.
Again, Mowrer's (1947) focus on CR proprioception providing response-produced stimuli for avoidance serves as an early example of this relationship. While few data have been collected on the development of behavioral response chains in the fear situation, Patterson (1975) has recently suggested that most behaviors appear to occur in "bursts," implying a facilitation effect of one response on the probability of occurrence of immediately subsequent responses of the same class.
272
THOMAS
D.
BORKOVEC
The above assumptions regarding response-component interactions only suggest that each component may serve conditional or discriminative functions for other components and that those interactions will influence the immediate anxiety reaction upon subsequent exposure to the feared stimulus. While the added consideration of individual differences in those interactions results in a highly complex model, its basic assumption is simple: Stimuli elicit one or more immediate anxiety-response components, which in tum may elicit further responses; some individuals, upon exposure to the feared situation, respond with an immediate image or thought which elicits physiological arousal, further catastrophizing images and thoughts, and!or avoidance behavior; others respond immediately with a physiological increase, leading to cognitive apprehension, avoidance, and! or further arousal; etc. The implications of this assumption relate importantly to both the maintenance and the reduction of any given case of anxiety. If, for example, the cognitive component of the immediate anxiety reaction is primarily or solely relevant to an individual's anxiety reaction and to the mediation of maintaining reactions in other components, then the second three relationships (2a-2c) are primarily or solely relevant to the maintenance of his anxiety. By the same token, the cognitive component and its interactive relationships with subsequent responses are therefore primarily or solely relevant to strategies for reducing his anxiety. This model, which considers such individual differences in combination with the interactive relationships of response groups, provides a framework for theoretical research on anxiety maintenance and reduction as well as for analyzing the effects found in the current behavior-therapy literature and suggesting a guideline for the development of appropriate self-regulation procedures. The latter analysis will be briefly presented first, while the former will be later exemplified by studies from our research program.
D. Intervention Strategies The presented description of the anxiety process allows us to view the current behavior-therapy literature from two different perspectives. The first refers to the assumed point of therapy impact on the
PHYSIOLOGICAL AND COGNITIVE PROCESSES IN THE REGULATION OF ANXIETY
273
anxiety process. The traditional behavioral approach to conceptualizing therapeutic intervention suggests that techniques aimed at reducing maladaptive behavior and increasing adaptive response to feared situations focus on the relationships between fear cues and immediate anxiety reactions. A more recent view exemplified by anxiety researchers in the area of cognitive behavior therapy and self-control suggests that it may be therapeutically wise to apply intervention strategies between immediate anxiety reactions and the subsequent maintaining reactions. Thus, for example, while systematic desensitization is often considered to influence the relationships between fear cues and the immediate anxiety reactions, recent developments suggest that the therapist should train the client to relax himself when confronted by the feared situation and to modify his self-verbalizations to promote more effective coping strategies (cf. Meichenbaum, this volume). The focus is thus shifted from the permanent treatment of a previously established stimulus-response relationship to the teaching of a generalizable coping skill that can be used when the client confronts any future, fear-producing situation. The client is therefore taught to administer his own interventions between the immediate anxiety responses and the subsequent maintaining responses. In terms of the descriptive model, such stress-inoculation training assumes that the physiological and cognitive components are relevant responses maintaining the anxiety reaction and that selfregulatory techniques attacking both of those responses would be the treatment of choice. By logical extension of the model, the therapeutic process would involve (1) identification of the relevant response components comprising the immediate anxiety reaction; (2) identification of subsequent maintaining reactions; and (3) application of the most efficient and efficacious techniques to eliminate the immediate reactions and training the client in self-control skills that preclude the occurrence of maintaining reactions. Research in behavior therapy is only beginning to develop such techniques and skills. Hopefully research on the conditions of anxiety maintenance and reduction within this descriptive framework will facilitate those efforts. Second, the role of individual differences in immediate responsegroup strength allows us to make some sense out of a current literature which seems to indicate that a variety of independent manipulations can effectively modify fear behavior. Recent reviews of animal condi-
274
THOMAS
D.
BORKOVEC
tioning (Wilson and Davison, 1971), human psychophysiology (Mathews, 1971), and the role of cognitive factors in desensitization (Wilkins, 1971) have arrived at a very similar conclusion: repeated CS exposure appears to be the only necessary condition for the ultimate elimination of fear. Beyond this, there is little agreement. Procedurally, there are many ways to present the CS. Theoretically several underlying processes have been offered to explain the effectiveness of various CS presentation techniques. Empirically many techniques have been found effective in reducing fear. However, the existing literature addressing procedural, theoretical, or even basic effectiveness questions has almost completely ignored the role of individual differences in anxiety-response components, despite Paul's (1969a) warning that the question of technique effectiveness cannot be divorced from subject characteristic considerations. Such neglect has given rise not only to invalid claims of effectiveness (Bernstein and Paul, 1971) but has resulted in erroneous conclusions regarding the underlying theories of the techniques as well as of the anxiety process itself. If, for example, the theory of a technique implies that the physiological component of fear is (1) important in the definition of fear and (2) the focus of that therapy technique, then tests of the technique are valid only if subjects displaying a strong physiological response make up the treatment sample and only if assessment of physiological responding is included in the outcome measures. If the theory of another technique implies that the behavioral component is important and the specific target of the treatment, then valid conclusions are possible only if subjects display strong avoidance and are measured primarily by behavioral scales. While this may appear to be intuitively obvious, disregard for this basic notion is common and stems from the inaccurate assumption that a technique generally influences "anxiety" and that any measure of "anxiety" is sufficient to draw conclusions relevant to that technique and its theory. Selection of subjects and measures which fit the theory never occurs to the investigator who accepts that assumption. We may categorize several common manipulations in terms of which response components are their focus. 1. Physiological Reactions: relaxation therapies, systematic desensitization, implosive therapy, biofeedback.
PHYSIOLOGICAL AND COGNITIVE PROCESSES IN THE REGULATION OF ANXIETY
275
2. Cognitions: rational emotive therapy, semantic conditioning, self-instructions, modified self-verbalizations, thought stopping, placebo and expectancy manipulations. 3. Overt Behaviors: reinforcement of approach behavior, response prevention, modeling, contact desensitization. The theory and/or procedures involved in each manipulation appear to focus on a particular anxiety-response component. This is not to say that the other components are conceptually ignored or that other components are not influenced. Rather, the focus of the techniques is at least on the specified response component, while additional focus on other components and the interactive relationships among components allows for multiple effects by any technique. However, each technique involves to a greater or lesser degree repeated, nonreinforced CS exposure. Thus each technique, if administered for a long enough period of time, will, hypothetically, effect a reduction of anxiety, however measured. But this is sufficient for the experimental analysis of neither the technique nor the anxiety process. The current desensitization literature is an excellent example of a body of therapy research which has ignored subject characteristics and as a consequence has produced endless debates regarding active mechanisms of effect. A variety of such mechanisms has been offered, each with associated empirical support. Thus systematic desensitization effects have been said to be due to: counterconditioning, extinction, expectancy, therapist reinforcement of nonfearful behavior, information feedback, controlled attention shifts, exposure to contingencies of nonavoidance behavior, modified cognitions via attribution manipulations, covert rehearsal, self-controlleaming, semantic associations, modifed self-sentences, simple demand and placebo effects, etc. The majority of the relevant studies have involved phobic college students selected on the basis of self-reported fear and behavioral avoidance on a pretest. The conclusions of the studies are ordinarily based on change in the behavioral component of anxiety after four to eight treatment sessions. As I have previously argued (Borkovec, 1973b), almost all therapyoutcome studies employ selection criteria and dependent measures which do not directly tap a response group considered of major importance in the definition of anxiety and desensitization processes (i.e., the physiological component). It is of little wonder that manipu-
276
THOMAS
D.
BORKOVEC
lations such as reinforcement, modeling, demand, placebo, etc., produce significant change in approach behavior and self-report if one assumes that avoidance behavior and reports of fear can be currently maintained by reinforcement, modeling, demand, suggestions, etc. Without matching theory, measures, and subject differences in relevant response groups, erroneous conclusions regarding any anxietyreduction techniques will continue. As long as researchers continue to select phobic subjects on the basis of only one or two non stringent criteria, they will continue to find that scores of manipulations, some of which are inherent in desensitization procedure, can be separately effective in changing avoidance behavior. Such demonstrations are indeed important. They are relevant, however, only to anxious behavior that is not mediated by a physiological response component and are irrelevant to the basic theoretical issues of desensitization.
II.
RESEARCH STUDIES ON THE MAINTENANCE AND REDUCTION OF ANXIETY
The presented descriptive framework of anxiety process predicts that everything influences everything else, making it useless as a theory. The model does, however, delineate some of the variables considered to be of importance in anxiety. The task of the researcher operating from such a model is to identify the conditions (environmental and subject) under which interrelationships among the variables do and do not occur and what effects those interrelationships have on behavior over time. This has been the goal of our program. The majority of our research has focused on two of the response components: physiological and cognitive processes. Since the model addresses anxiety maintenance/reduction, our aim has been to identify the conditions under which interrelationships of those two variables lead to the maintenance or reduction of anxiety. Since individual differences in response components are at the core of the model, our research strategy has been to assess the effects of manipulations on subjects differing in response-component strength, principally in level of physiological arousal and autonomic awareness. The specific experimental strategy we have adopted has involved repeated exposure of subjects to feared stimuli, with intervening cognitive and/or physiological manipulations. Two main target behav-
PHYSIOLOGICAL AND COGNITIVE PROCESSES IN THE REGULATION OF ANXIETY
277
iors have been employed: snake phobias and social anxiety. In the course of our research, it has been found that these two groups are quite different in two important respects. First, the behavioral component of the fear reaction of snake phobics, as typically selected in outcome research, is very susceptible to modification by demandi suggestion (Borkovec, 1973a; Bernstein, 1973). Simply ask the subject (implicitly or explicitly) to show behavioral improvement and suggest that he will be less anxious, and reduced behavioral fear will be observed. The fear behavior of socially anxious subjects has not been so influenced (Borkovec, Stone, O'Brien, and Kaloupek, 1974). Second, on rare occasions when physiological measurement has been obtained during snake avoidance tests (e.g., Borkovec, 1973a; Craighead, 1973), maximal average heart rate has been no greater than 96 beats per minute. In studies of speech and social anxiety, similar measurements have revealed increases in arousal typically between 113 and 118 beats per minute (Borkovec, Wall, and Stone, 1974; Borkovec, Stone, O'Brien, and Kaloupek, 1974; Singerman, 1974). On the basis of the demand studies, it is clear that the use of snake phobics requires controls for demand and suggestion effects. What often appears to be fear reduction in therapy outcome studies with this target behavior may often simply reflect the influence of extratherapeutic variables on avoidance behavior. On the basis of the heart-rate data, it is equally clear that the physiological response will oridinarily be a relevant fear component in studies involving social anxiety and an irrelevant component in those employing snake phobias. Since we have used both targets, we have had an opportunity to observe the effects of physiological and cognitive manipulations on two anxiety problems differing in the extent to which the physiological response is functionally relevant. In the following presentation of research studies, therefore, I will refer to investigations with low physiological reactors (snake phobics) and high physiological reactors (social- and speech" anxious subjects). These two characteristics (demand/suggestion susceptibility and level of physiological arousal) appear to be related to each other and give rise to one of the central theses of our work. To the extent that the immediate anxiety reaction involves a weak physiological component, simple manipulations of the cognitive and behavioral components of fear (such as demandisuggestion) will be effective in changing those components. To the extent that the immediate anxiety reaction in-
278
THOMAS
D.
BORKOVEC
volves a strong physiological component, such manipulations will be ineffective and will be effective only after the autonomic component is reduced. By logical extension throughout the anxiety model, to the extent that any response component is strong and immediately elicited by the feared stimulus, that component must be modified directly if efficient fear elimination is to be accomplished. The origin of this assumption stems from my dissertation (Borkovec, 1970, 1972) and a brief summary of that study will serve to set the stage for a description of our subsequent research program. In a rather typical outcome study, snake phobic college females who failed to touch a live snake on pretest were randomly assigned to one of four treatment conditions: systematic desensitization, implosive therapy, avoidance-response placebo, and no treatment. In order to assess the contribution of client expectancy for improvement, half of the subjects in each condition received therapeutic instructions regarding the purpose of the procedures, while half were presented non therapeutic instructions. The theories underlying desensitization (Wolpe, 1958) and implosive therapy (Stampfl and Levis, 1967) both suggest that phobic behavior is maintained because of negatively reinforced avoidance of feared stimuli. This common assumption directly suggested a model-relevant placebo condition. Subjects visualized the same hierarchy items presented to the two therapy conditions. However, as soon as the subject signaled the presence of anxiety, a standardized avoidance response was visualized. The placebo procedure thus analogized the phobic's behavior assumed to occur in real-life situations and served as a theoretically inert treatment. Posttest and four-week follow-up assessments were made subsequent to four weekly therapy sessions. Several important findings emerged and suggested the major variables to be explored in our developing program: (1) Expectancy instructions had a dramatic effect on the behavioral component of fear, suggesting a potentially important role for cognitive variables in fear reduction. (2) The modelrelevant placebo condition showed behavioral improvement equal to that displayed by the therapy conditions. A serious problem was apparently raised for the theories underlying desensitization and implosion. (3) Desensitization and implosion both reduced the physiological component of fear at posttest regardless of whether subjects were given positive or neutral expectation for improvement. Subse-
PHYSIOLOGICAL AND COGNITIVE PROCESSES IN THE REGULATION OF ANXIETY 279
quent avoidance behavior for these subjects was influenced, however, only when an expectation for improvement was instructionally established. The role of physiological cues and their interaction with the subject's interpretation of the procedures and the effects of the procedures were thus implicated.
A. The Role of Physiological Arousal and Cognition In our subsequent attempts to identify maintaining/reducing conditions for anxiety, we have focused primarily on the role of physiological arousal and cognition. In terms of the model, we have so far been concerned with internal cues, immediate physiological and cognitive reactions, interrelationships between those reactions and subsequent physiological and cognitive responses, and the effects that the relationships have on subsequent anxiety. Subject characteristics enter into these relationships through investigations of individual differences in magnitude of the physiological response and in the perception of that response. Studies relating to physiological reactions and cognition will be presented first, followed by investigations of the contribution of subject characteristics to such relationships.
1. The Effect of Cognitive Avoidance Behavior in Maintaining the Anxiety of Low Physiological Reactors The theoretical models underlying both systematic desensitization and implosive therapy suggest that avoidance behavior precludes CS exposure and hence extinction. The ideal control condition in a theoretical investigation of either technique would involve a group of subjects who visualize the same phobic hierarchy but visualize performing an avoidance response upon elicitation of the CR. As mentioned earlier, such a group was included in the outcome study of desensitization, implosion, and expectancy effects on snake phobic behavior (Borkovec, 1972). While basal skin conductance data collected during therapy revealed that all three treated groups displayed reductions in arousal over hierarchy-item presentations, a review of the psychophysiological research suggested that heart-rate measures provided data during imagery procedures that were more in line with theoretical predictions than were skin-conductance measures (Mat-
280
THOMAS
D.
BORKOVEC
hews, 1971). Subsequent analyses of samples from continuously monitored heart rate indicate that the avoidance-response group remained at high levels throughout imaginal exposure to the CS scenes, while both desensitization and implosion groups displayed significant decreases in heart rate (Borkovec, 1974). Thus despite repeated CS exposure frequently demonstrated to produce autonomic decreases (Mathews, 1971), the addition of imagined avoidance behavior in response to the CS resulted in maintenance of physiological arousal. In addition, outcome pulse-rate of the avoidance-response group failed to decrease relative to nontreated control subjects, while both desensitization and implosion groups displayed significantly greater reductions than no treatment. The combined process and outcome results, which occurred irrespective of whether the subject expected to improve or not because of the therapy administration, directly support Mowrer's two-factor theory of anxiety maintenance in humans for a physiological response and provide evidence for a physiological-cognitive interaction in determining that maintenance. In terms of the descriptive model, external cues (therapist verbalization of the CS hierarchy scene) and internal cues (subject visualization of the scene) elicited a conditioned fear response. Upon the subject's awareness of an anxiety response, a cognitive avoidance response was made. The effect of this relationship was the maintenance of the physiological component of fear as evidenced by both process and outcome heart-rate measures. In contrast, scene presentations in both desensitization and implosion involved repeated, nonreinforced CS exposures with a relative preclusion of cognitive avoidance, and both techniques resulted in reductions of autonomic process and outcome measures. The fact that these processes occurred in fearful subjects for whom the physiological component is not very strong is quite encouraging. The results suggest that a simple cognitive response can maintain anxiety despite repeated CS exposure.
2. Cognitive Expectancy Related to Self-Reported Outcome and Physiological Process Research by Goldstein and others (e.g., Friedman, 1963; Goldstein, 1960; Goldstein and Shipman, 1961; Piper and Wogan, 1970) has suggested that client expectancy for improvement is powerfully related to therapeutic outcome. Unfortunately the majority of the research
PHYSIOLOGICAL AND COGNITIVE PROCESSES IN THE REGULATION OF ANXIETY
281
from which this conclusion derives is based on correlational analyses of self-report data. More recent investigations have attempted to manipulate expectancy via the presence or absence of therapeutic instructions during therapy administration. Half of these studies have found such a manipulation to be unrelated to therapy outcome, while the other half (including the above study from our laboratory) demonstrated greater behavioral improvement under positive (therapeutic) than neutral (nontherapeutic) expectancy. Reviews of this literature have indicated that the differences between the two sets of studies appear to reside in important methodological considerations. Wilkins (1973), for example, found that therapists and/or observers were not "blind" to the condition status of subjects in studies demonstrating a significant expectancy effect. One aspect of positive expectancy instructions, however, is that high demand is placed on the subject for displaying less avoidance behavior on the posttest. He knows that the experimenter expects improved behavior and responds accordingly. A review of these studies has noted that investigations finding a significant difference between positive and neutral expectancy conditions (i.e., high and low demand) employed nonstringent selection criteria, while studies failing to find an expectancy effect used severe selection criteria (cf. Borkovec, 1973b). If it is assumed that the strength of the physiological component increases as fear selectioncriteria become more stringent (Bernstein and Paul, 1971), then the outcomes of these studies support the hypothesis that expectancy manipulations are effective in changing the behavioral component only to the extent that the physiological component is absent. In the only clinical trial of an expectancy manipulation where a strong physiological response is likely (Gelder, Bancroft, Gath, Johnston, Mathews, and Shaw, 1973), high- versus low-expectancy instructions had no effect on the outcome of neurotic patients treated by desensitization or implosive therapy. Despite the apparently weak contribution of expectancy to outcome change, the importance of client expectation for improvement, born in clinical impressions and Goldstein's research, remains a firmly entrenched assumption in clinical folklore. Our research has obtained two bits of further information regarding this construct. First, in three studies within a sleep-disturbance program (Borkovec, Kaloupek, and Slama, 1975; Slama, 1975; Steinmark and Borkovec, 1974) correlations obtained between expectancy ratings and outcome were generally
282
THOMAS
D.
BORKOVEC
nonsignificant. Second, however, analysis of the heart-rate data during desensitization, implosion, and avoidance response placebo conditions (Borkovec, 1974) revealed a significant main effect of expectancy during the process of treatment. This effect provides an example of a cognitive-physiological relationship in anxiety process. Subjects who were given therapeutic instructions regarding the purpose of the treatment procedures displayed lower heart rate throughout the therapy sessions than subjects given nontherapeutic instructions, relating expectancy manipulations to an objectively measured and potentially important therapy process variable. If low physiological arousal is facilitative of desensitization effects as Lader and Mathews (1968) have suggested, then positive expectancy instructions may contribute to fear extinction via its arousal-reducing effects during repeated CS exposure. If generation of intense CR is an essential part of implosive therapy process, as Stampfl and Levis (1967) have indicated, then positive expectancy may initially mitigate implosion outcome.
3. Manipulation of Physiological Cues among Low Physiological Reactors The anxiety outcome study mentioned above exemplified a strategy of monitoring naturally occurring physiological arousal during CS exposure and observing what anxiety effects occur as a function of manipulating responses subsequent to the immediate anxiety reaction. A second approach for investigating the role of physiological cues in fear involves the manipulation of false physiological feedback. This strategy directly addresses the physiological-cognitive relationship in that the subject's perception of the autonomic cues is of primary importance in determining the effects of those cues. The earliest study of this nature was conducted by Valins and Ray (1967) in an attempt to provide evidence for a cognitive interpretation of the effects of systematic desensitization. Valins and Ray presented alternating slides of the word shock and feared stimuli to snake phobic subjects. Experimental subjects were informed that sounds presented during the slide presentations were their heartbeats, while control subjects were told that the sounds were extraneous noise. No change in feedback heart-rate occurred during presentation of the feared slides, while feedback heart-rate increased upon presentation of the word shock. The experimental group displayed significantly less avoidance to a live snake than the control group. The authors argued
PHYSIOLOGICAL AND COGNITIVE PROCESSES IN THE REGULATION OF ANXmTY
283
that the subjects' interpretation of the physiological feedback resulted in a modification of cognitions regarding the feared object, i.e., "That stimulus no longer affects me internally." It is this revised cognition that presumably mediates changes in overt behavior toward the fear object. Several attempted replications (Gaupp, Stem, and Galbraith, 1972; Kent, Wilson, and Nelson, 1972; Sushinsky and Bootzin, 1970) failed to produce this phenomenon. These studies differed from the original Valins and Ray investigation in one or more procedural aspects. Some studies included a pretest to ensure the selection of truly phobic subjects. Such an in vivo experience with the feared object, however, may mitigate the effects of symbolic presentation of the feared stimulus in association with the bogus feedback. Second, Valins and Ray employed a high-demand posttest (a money incentive was offered), while some of the replications essentially involved lowdemand posttesting. Because of the susceptibility of avoidance behavior to demand!suggestion influences and without control for the types of demand inherent in the conditions, it would perhaps be desirable to ensure high posttest demand in order to control for uncontrolled demand influences. Differential avoidance improvement among treatment conditions would more clearly reflect level of anxiety than situational determinants. Borkovec and Glasgow (1973) replicated the Valins and Ray study in a Solomon four-groups design. Experimental subjects heard false heart-rate feedback during the presentation of slides of snakes. The feedback occurred to a hierarchy of slides similar to hierarchies constructed in systematic desensitization treatment. Heart rate was heard to increase to initial presentations of items and to decrease across repetitions of the same slide and across hierarchy items. Control subjects observed the same slides and heard the same sounds but were not informed that the sounds were representative of their heart-rate response to the pictures. Half of the subjects in each of the conditions were pretested in a behavioral approach test, while half were not pretested. All subjects were informed that the effects of the procedures would be to reduce their fear of snakes and that it was important for the success of the experiment that they approach as close as possible. Thus the posttest instructions were high in demand! suggestion, offering a valid design despite the use of low-fearful snake-phobic subjects. As predicted, the experimental group showed significantly less avoidance than the control group, but only under nonpretested conditions. In addition, continuous heart-rate monitor-
284
THOMAS
D.
BORKOVEC
ing during the slide presentation task indicated that pretested subjects displayed significantly greater heart-rate reactivity to the snake slides than did nonpretested subjects regardless of feedback conditions. The original predictions of the study were based on the notion that fear occurring during an in vivo pretest would mitigate the effects of feedback manipulations during symbolic CS presentations. Under those conditions the only modified cognition that might result would be, "I am afraid of the actual snake, but I am not afraid of slidesof snakes." In addition, however, the heart-rate data during slide presentations suggested that in vivo exposure to the feared situation sensitized the subject to symbolic presentations of that feared object and resulted in maintained autonomic arousal during those presentations. In terms of the descriptive model, symbolic CS presentations were paired with a tape-recorded sound. Depending on whether the sounds were labeled as noise or physiological arousal, the subject's subsequent cognitive interpretation either was irrelevant to his fear and therefore to its modification or was relevant and resulted in reduced fear. This effect importantly occurred only on the behavioral component of anxiety and only in the absence of a pretest exposure. Such an outcome suggests that cognitive manipulations regarding physiological cues will be effective only if subjects are low physiological reactors. The fact that pretest exposure both increased actual physiological arousal and mitigated the effects of the cognitive manipulation suggests that in vivo presentations can lead to arousal maintenance to both symbolic CS exposures and subsequent in vivo exposures, even among subjects who ordinarily display little physiological reactivity to the feared stimulus. Thus while false feedback experiments may supply evidence for the role of physiological cues in maintenance or modification of fear, actual physiological arousal may produce an overriding effect.
4. Manipulation of Physiological Cues among High Physiological Reactors Unfortunately previous studies of false feedback effects on fear employed feedback tasks intervening between pre- and posttest in vivo exposures. The critical issue regarding the role of physiological cues centers on the occurrence of feedback during the subject's confrontation with the actual feared situation. During the conduct of
PHYSIOLOGICAL AND COGNITIVE PROCESSES IN THE REGULATION OF ANXIETY
285
two studies investigating the effects of demand instructions on snake phobic avoidance (Borkovec, 1973a), increasing versus decreasing false heart-rate feedback was presented to two groups of subjects during the postlest snake exposure. No main effects of feedback occurred on any of the dependent measures. Yet the Valins and Ray (1967) and Borkovec and Glasgow (1973) results suggested that avoidance behavior may be modified upon exposures subsequent to feedback experience. Consequently a study was designed in which false physiological feedback was presented during the actual exposure to feared situation, and its effects were assessed on a subsequent exposure (Borkovec, Wall, and Stone, 1974). Speech-anxious subjects prepared and presented three consecutive speeches. During the second speech, one of five feedback conditions was presented: heart rate increasing, heart rate decreasing, heart rate no change, and two control conditions. Subjects in the heart-rate-increasing condition heard what was reported to be their own heart rate during the last minute of the preparation period of the second speech and throughout the presentation of the second speech. Feedback indicated fluctuating but increasing rate in heartbeats throughout the task. Similarly subjects in the heart-rate-decreasing condition heard gradual decreases in heart rate throughout the speech. Subjects in the heart-rate-no-change condition were given heart-rate feedback indicating relatively little fluctuation at a relatively low level throughout the second speech. Subjects in a feedback control condition heard the same heart-rate tape that was employed in the heart-rate-increasing condition but were told that the sounds were extraneous noise. Subjects in a no-feedback control condition wore the headphones but heard no sound during the speech. Pronounced heart-rate increases did in fact occur during all three speeches. There was, however, no effect of feedback conditions on actual heart rate. As predicted, no significant differences were found during the feedback speech among any of the conditions. Also as hypothesized, the heart-rate-decreasing and no change conditions resulted in significantly less self-reported and behavioral manifestations of anxiety on the posttest (third) speech than did the heart-rateincreasing condition. The combination of high physiological arousal which occurred in all groups and increasing false feedback in the heart-rate-increasing condition resul~ed in the maintenance or facilitation of fear behavior upon exposure to the feared situation subsequent to the feedback experience. Despite repetitious presentation of the
286
THOMAS
D.
BORKOVEC
feared situation, fear behavior maintained. Given the Valins and Ray and the Borkovec and Glasgow studies demonstrating posttest behavioral differences subsequent to feedback tasks and the above study demonstrating a lag in false feedback effects during repeated exposures to the feared situation, some process must occur between the feedback experience and the subsequent exposure to the feared situation to result in maintenance of behavioral anxiety. If phYSiological feedback is considered to be a response-produced stimulus and if, in agreement with Eysenck (1968), such conditioned responses may serve as aversive stimuli, then it can be hypothesized that awareness of the physiological component of the CR can in itself serve to maintain fear behavior during nonreinforced CS presentations. False physiological feedback may then facilitate or maintain fear because of its similarity to actual history of experience with real feedback, assumed to contain aversive properties. An alternative hypothesis in line with Valins and Ray's theory regarding systematic desensitization is that internal cuing sets the occasion for verbal mediators generated by the subject in response to perceived autonomic functioning. In the case of the speech-anxiety study, subjects in the heart-rate-increasing condition had an opportunity during the preparation period prior to the third speech and subsequent to the feedback speech to verbalize to themselves interpretation and/or evaluation of their arousal as represented by the false feedback. Focusing of attention on such cognitions stimulated by the false feedback experience may have disrupted the preparation of a subsequent speech, or the cognitions themselves elicited additional arousal which interfered with subsequent speech performance and increased reported anxiety. It is important to note that self-report and behavioral-anxiety measures of the heart-rate-decreasing and no-change conditions did not show significantly greater reductions relative to the control conditions. In contrast to snake-phobic samples, in which cognitive manipulations may, under limited circumstances, result in reduced avoidance behavior relative to controls, it appears necessary to modify the strong physiological component present in speech anxiety before cognitive interpretations can result in decreases in fear behavior. The facilitated behavioral and self-reported fear in the heart-rate-increasing condition, however, indicates that fear behavior can be maintained under conditions of both high actual arousal and a cognitive interpretation of high arousal.
PHYSIOLOGICAL AND COGNITIVE PROCESSES IN THE REGULATION OF ANXIETY 287
5. Cognitive Manipulations on Naturally Occurring Arousal in High Physiological Reactors Recently research from the Schachter tradition has resulted in the development of an attribution therapy, whereby cognitive manipulations attempt to minimize aversive or dysfunctional emotional states by leading the individual to attribute his arousal to some nonemotional source. While a variety of studies have found significant effects (e.g., Dienstbier and Munter, 1971; Nisbett and Schachter, 1966; Ross, Rodin, and Zimbardo, 1969; Storms and Nisbett, 1970), our suspicion that simple cognitive manipulations are likely to be ineffective when there is a strong physiological component to the fear reaction led to the conduct of an attribution study of speech anxiety (Singerman, 1974). In this study 60 speech-anxious subjects (30 moderately fearful and 30 highly fearful) presented two consecutive speeches. Meaningless noise, used as the misattribution stimulus, was presented from the time the subject entered the laboratory to the end of the first speech but was absent during the second speech. Prior to the first speech subjects were randomly assigned to one of three attribution conditions. Subjects in the arousal condition were told that the noise would increase physiological activity. Attribution theory would predict that these subjects should be less anxious during the stressful situation, since they would attribute their naturally occurring arousal cues to the noise rather than to the speech situation. Subjects in a sedation condition were given a placebo statement indicating that the noise would have a decremental effect on physiological activity. Control subjects were given no expectation regarding the likely effects of the noise. No evidence was found for the reverse placebo effect predicted by attribution theory. In fact, among highly fearful subjects the arousal condition produced greater behavioral anxiety than the sedation and control conditions, and this difference maintained to the second speech. The nature of the target behavior and its associated physiological activity appears to have been critical in the failure of the cognitive manipulation. If subjects have an extensive past history of association between a setting and strong physiological responses, successful cognitive misattribution appears unlikely. In terms of the descriptive model and similar to the conclusions of the false feedback study on speech anxiety, the combination of high physiological arousal and a cognitive set indicating that an extraneous stimulus
288
THOMAS
D.
BORKOVEC
would produce increases in arousal (the arousal condition) appears to have an additive effect resulting in increased saliency of arousal cues and therefore increased fear behavior.
6. The Development of a Physiological Fear Component in Nonanxious Subjects and Possible Cognitive Contributions In the study of social-anxiety measurement (Borkovec, Stone, O'Brien, and Kaloupek, 1974), high- and low-anxious males had undergone pre- and posttest exposures to a laboratory interaction with a female research assistant. Half of the subjects received high-demand! suggestion posttest instructions at posttest, while half were given lowdemand instructions. The demand manipulation failed to produce improvement on any measure. There were two surprising outcomes, however, which may relate to the origins of an anxiety response and the role of cognition in facilitating anxiety over time: (1) low anxious subjects displayed a significant increase in heart rate from pre- to posttest and (2) low anxious subjects in the low-demand condition showed that increase during the anticipatory phase prior to the beginning of the interaction, while subjects in the high-demand condition evidenced that increase only during the interaction phase itself. It is reasonable to conclude that confronting a nonreceptive female is functionally an aversive experience for ordinarily nonanxious males, an experience which can result in increased physiological activity over repeated exposures. In addition, it may be hypothesized that cognitive appraisals of that experience between testings may contribute to the heightened arousal. The significant interaction between demand, phase of interaction, and testing supports such a cognitive interpretation. In high demand these subjects were reassured that they would be less anxious, a suggestion that indeed mitigated anticipatory arousal but failed to mitigate reactive arousal upon presentation of the interaction situation. In low demand, on the other hand, the aversive experience led to facilitated anticipatory arousal. Once the subjects were interacting, the facilitative effect of the aversive pretest experience on arousal was absent. Thus given the increased arousal that occurs in nonanxious subjects as a function of the aversive nature of the pretest, the presence or absence of an expectation of reduced
PHYSIOLOGICAL AND COGNITIVE PROCESSES IN THE REGULATION OF ANXIETY
289
posttest fear established via the demand manipulation may influence the point of inflection of arousal (anticipatory or reactive). These hypotheses require further research. It would be important to determine whether the general physiological increase among nonanxious subjects ultimately results in increased behavioral and self-reported anxiety and whether cognitive manipulations or point of arousal inflection contribute to such increased anxiety.
B. The Role of Individual Differences in Physiological Arousal and Autonomic Perception Assessment of ongoing physiological activity and the manipulation of cognitive and physiological cues via instructional set and false feedback have given us several means for investigating the contribution of physiological and cognitive reactions to fear maintenance and reduction. An additional method involves the investigation of individual differences in physiological reactions and the perception of those reactions. The potential importance of the physiological fear component was discussed earlier with regard to response group patterns. In the present section, studies relating subject characteristic differences in that component to anxiety and manipulation of its conditions will be presented. Since our interest has resided in the functional role of physiological feedback as a response-produced stimulus, it appeared that either direct assessment of an autonomic response or measurement of perceived physiological activity would provide a potentially useful dimension of individual differences. Certainly the greater the autonomic activity, the greater the probability that physiological cues would be functionally related to subsequent behavior. On the other hand, mild or moderate arousal may provide sufficient feedback for some individuals to be functionally important in their fear behavior. In addition, measurement of the perception of autonomic feedback has the advantage of providing access to a response that relates to both physiological and cognitive response groups of the individual. The majority of our studies, therefore, has focused on autonomic perception measurement.
290
THOMAS
D.
BORKOVEC
1. The Autonomic Perception Questionnaire Reports by Mandler and his associates in the late 1950s indicated the existence of a self-report measure of perceived autonomic functioning and provided initial evidence of its potential utility as a bridge between physiological activity and self-reports of anxiety. Their Questionnaire on the Perception of Feeling consisted of 28 items which requested subjects to indicate the degree to which they noticed various bodily reactions during two emotional states. The first 21 items (Autonomic Perception Questionnaire (APQ); see Table 4) related to subjective experiences when the individual was anxious; the last 7 items were rated in terms of experiences when the individual was happy. Ratings on the original scales were made by the checking of a point on a continuous line, with the ends of each line representing absence versus extreme presence of each bodily reaction. Scoring involved the use of a transparent overlay dividing the line into 10 equal parts. Our program has used the same 21 APQ items, although the item scales have been rated by a simple circling of the appropriate number (0-9) representing the rater's experience of that reaction. In both Mandler's and our own research, high and low perceivers have been defined in terms of the sum over individual item scores. While Mandler has employed a general form of the APQ ("When you are anxious, do you notice ... "), we have variously used both general and task-specific ("During the experiment, did you notice ... ") forms. Normative data have not been previously presented. Consequently the general APQ has been administered to two samples of introductory psychology students (spring and fall semesters, 1971) at the University of Iowa and to all adult patients seen in the university's psychological clinic from 1971 to 1973. Table 2 presents the mean scores on each item of the APQ and its total for females and males from these three groups. Females tended to obtain higher total scores than males, although the difference was significant only in one of the normal samples. Females scored significantly higher than males in both normal samples on items reflecting awareness of cold hands, shallow breathing, lump in throat, upset and sinking stomach, and the bothersomeness of bodily reactions. Differences in the clinic sample again favored the females but were restricted to experience of face becoming hot,
PHYSIOLOGICAL AND COGNITIVE PROCESSES IN THE REGULATION OF
ANxmn 291
perspiring, and headache. When the normal samples were pooled and compared to the clinic sample, neither female nor male patients were found to differ significantly from normal subjects of the same sex on total APQ score. Clinic females did score significantly higher than their n6rmal counterparts on frequency of awareness of bodily reactions and muscle tension. Clinic males displayed significantly higher scores than normal males on frequency of awareness, lump in throat, upset stomach, and bothersomeness of the bodily reactions, while they scored significantly lower on perception of hot face and perspiration. The greater frequency of awareness of bodily reactions among the clinic samples suggests the particular relevance of those cues to clinical problems. TABLE 2 APQ Item Means for Females and Males from Normal and Clinic Samples
APQ Item
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 Total
Spring sample 1971
Fall sample 1971
Females Males (N = 239) (N = 181)
Females Males (N = 215) (N = 211)
4.87 4.85 4.30 4.32" 5.08 3.56 4.84 3.49 4.54 4.78 4.68 3.25 3.43 2.99 2.77" 3.01 3.97" 4.41" 4.58" 3.92 3.88" 85.52
4.82 5.05 3.95 2.75 5.36 3.78 4.79 1.99 4.70 4.92 4.91 3.39 3.45 3.24 2.20 3.02 3.37 3.35 3.80 3.63 3.39 79.86
5.35 5.16 4.26" 4.22" 5.51 4.00 5.15 3.28" 4.96 5.19 5.09 3.36 3.36 2.90 3.0Qb 3.47" 4.46" 4.76" 5.31" 4.37" 4.15" 91.35"
5.09 5.18 3.74 3.21 5.42 3.89 4.90 2.11 4.78 4.90 4.77 3.44 3.35 3.11 2.46 2.95 3.59 3.43 3.99 3.79 3.62 81.73
Clinic sample 1971-73 Females (N = 37)
Males (N = 38)
5.49 5.84" 4.16" 3.54 5.49" 3.46 6.32" 4.05" 4.35 4.65 4.78 3.81 3.95 3.16 3.35 3.46 4.35 4.86 5.27 4.76 4.76 93.86
5.34 6.00" 2.55" 2.89 4.29" 3.55 5.53 2.53 4.53 4.76 4.55 3.71 3.34 3.21 2.92 2.63 4.39" 4.87" 4.68 4.00 4.79" 85.08
a I-tests (p < .05) for same-sex scores between clinic sample and pooled normal samples. • I-tests (p < .05) for between-sex scores within a sample.
292
THOMAS
D.
BORKOVEC
In the first study to employ the APQ (Mandler, Mandler, and Uviller, 1958) subjects selected for reporting high levels of autonomic perception displayed significantly greater autonomic reactivity (heart rate, PGR, and respiration) during stress than low perceivers and overestimated that feedback relative to the underestimates of the low perceivers. The general APQ was found to correlate significantly with Taylor's Manifest Anxiety Scale (MAS) and was related to the number of physiological cues reported during the stress situation itself. Small but significant correlations between perceived and actual autonomic reactivity during stress were found in a second study (Mandler and Kremen, 1958), which employed an unselected sample of subjects. In contrast to the earlier study, in which high- and low-perception groups were selected on the basis of extreme scores, low-perception subjects overestimated and high perceivers underestimated actual arousal. The APQ again correlated significantly with the MAS. Performance on a vocabulary test was unrelated to actual autonomic arousal but negatively related to the perceived arousal. Our fall sample has also completed the primary items of Ullmann's (1962) facilitationinhibition scale, which correlates significantly with Byrne's (1961) repression-sensitization scale. Correlations between the APQ and F-I scales were significant for both females (r = .31, P < .01) and males (r = .37, P < .001). As would be expected, high perceivers are high facilitators (sensitizers). Mandler's replicated relationship between perceived autonomic feedback and both actual physiological measures and a general selfreport anxiety index suggested that the APQ might be a promising bridge between cognition and physiology. In addition, the performance differences between low and high perceivers suggested that the subject characteristic of autonomic perception may be related to important overt behavior. In addition, recent investigations have found that APQ is related to performance in heart-rate-control tasks (Bergman and Johnson, 1971; Blanchard, Young, and McLeod, 1972). There appeared to be sufficient evidence, then, that the instrument would provide a meaningful dimension of individual differences potentially related to fear, its maintenance, and its reduction. 2. Perceived Arousal, Actual Arousal, and Demand/Suggestion
The first attempt to obtain evidence for a relationship between subject characteristics in arousal, anxiety, and the conditions influenc-
PHYSIOLOGICAL AND COGNITIVE PROCESSES IN THE REGULATION OF ANXIETY 293
ing anxiety occurred in the two early demand/suggestion studies (Borkovec, 1973a). Instructional demand and suggestions of improvement were chosen as the external stimuli to be manipulated because of our concurrent interest in the general issue of demand characteristics in fear research. Individual differences in physiological cues were defined by (1) actual pulse-rate response in anticipation of exposure to the feared object and (2) level of autonomic perception as measured by the general APQ. Snake-phobic subjects were categorized into high or low levels on both the reaction and the perception factors. It was predicted that the presence of strong physiological cues would reduce the effects of the demand manipulation. In the first of two studies, reaction and perception were unrelated to any outcome measure among nonphobic subjects. Phobic subjects who were high in reaction, however, displayed less behavioral improvement than phobic subjects low in reaction. Contrary to the prediction, high-perception subjects showed greater behavioral improvement than low-perception subjects. On the pulse-rate outcome measure, phobic subjects who were high or low in both reaction and perception displayed significant reductions in pulse-rate response to the phobic stimulus from pre- to posttest, while subjects in the other two groups showed no change, that is, they showed a maintenance of the physiological fear component. The second study involved two manipulations: increasing versus decreasing heart-rate feedback during the posttest exposure and lowversus high-demand instructions. Subject-characteristic groupings were made in the same manner as in the first study. Under lowdemand instructions, no effects of false feedback, perception, or reaction were found. Under high demand, however, high perception was again related to greater approach improvement than low perception. Three of the four low-reaction groups displayed greater improvement than three of the four high-reaction groups. One of the discrepant groups (low-perceptionllow-reaction subjects under decreasing false feedback) showed little change due to ceiling effects on the avoidance scale. Substantial and unpredicted improvement occurred in the second discrepant group (high perception/high reaction under increasing feedback). In the high-demand condition and under increasing feedback, pulse-rate improvement differences among subjectcharacteristic groups were almost identical to those occurring in the first study: accurate perceivers (subjects high or low in both percep-
294
THOMAS
D.
BORKOVEC
tion and reaction) displayed reductions over tests, while inaccurate perceivers showed a maintenance of pulse-rate response. Several implications follow from these two studies. First, while demand characteristics were found to exert strong effects on avoidance behavior, their influence depended on arousal-cue factors. In general, the greater the actual reaction, the less effective was the demand manipulation. The principle exception was the striking approach improvement for subjects high in perception and reaction under the increasing false-feedback and high-demand condition. This same feedback and demand condition, however, resulted in significant pulse-rate improvement among these subjects, providing further support for the hypothesis that the physiological fear component must be modified before demand or cognitive manipulations will affect overt behavior. Second, high-perception subjects were influenced by demand to a greater extent than low-perception subjects, contrary to predictions. Either autonomic perception does not reflect the functional role of physiological feedback, or its relation to behavior is more complex than was originally conceptualized. Third, the interaction of perception and reaction determined the probability of extinction of a physiological component of fear response. Whether accuracy of perception or a more basic attention-focusing characteristic underlies this effect remains unknown. 3. Perceived Arousal and the Incubation Phenomenon
Eysenck (1968) has suggested that under certain conditions, repeated es exposure in the absence of ues presentations may result in increases, rather than decreases, in fear. The noxious character of the eR itself is highlighted in contributing to this incubation effect. In addition to studies cited in his review, other investigations involving repeated or prolonged exposure to feared stimuli (Boulougouris, Marks, and Marset, 1971; Bresnitz, 1967; Miller and Levis, 1971; Rankin, Nomikos, Opton, and Lazarus, 1964; Rohrbaugh and Riccio, 1970; Rohrbaugh, Riccio, and Arthur, 1972) have supported the occurrence of the phenomenon in both animal and human subjects. Three studies implicated duration of es exposure as an important variable determining the occurrence of the incubation process. Specifically, in each study brief es exposure resulted in greater fear relative to zero or long exposure. Stone and Borkovec (1975) attempted to
PHYSIOLOGICAL AND COGNITIVE PROCESSES IN THE REGULATION OF ANXIETY 295
replicate the effect and to determine the relationship of autonomic perception to that effect. Because of Eysenck's stress on the noxious effects of the CR in facilitating fear increases during CS presentation, it was reasonable to expect that the incubation process would be greater among high than among low autonomic perceivers. Phobic college students observed a snake for 0, 15, or 45 minutes subsequent to a behavioral pretest. In line with earlier suggestions regarding the valid use of snake-phobic subjects, posttest was administered under high demand.· In a replication of the earlier studies, the results revealed less avoidance for the 0- and the 4S-minute conditions relative to the lS-minute condition. In addition, analysis of continuously monitored heart-rate data indicated reductions in heart-rate reaction to the phobic stimulus from pre- to posttest among zero and long-exposure subjects and an increase among brief-exposure subjects. Analyses of high and low autonomic-perception groups revealed that the incubation effect on avoidance behavior occurred only among high perceivers. The results of this study supported the existence of an incubation phenomenon in humans and the role of both CS duration and perception of autonomic cues in producing that effect. 4. Perceived Arousal Related to Overt Behavioral Anxiety
During our exploratory studies in the measurement of anxiety, socially anxious and nonanxious college males were exposed to three increasingly stressful phases of social interaction with male confederates (Borkovec, Fleischmann, and Caputo, 1973). Each anxiety group had been further divided into high and low autonomic perceivers on the basis of the general APQ. Similar to the results of the Borkovec, Stone, O'Brien, and Kaloupek, (1974) study, self-report measures discriminated the anxiety groups, while behavioral measures did not. Although the anxious and the non anxious groups did not differ on the general APQ given prior to the study, socially anxious subjects reported greater awareness of autonomic cues during the test than nonanxious subjects. The APQ was thus demonstrated to be situationally sensitive. In addition, the APQ factor interacted significantly with social-anxiety group on the number of words produced during the test and with phase of the test on observer-rated overt signs of anxiety. Specifically, nonanxious high perceivers produced the greatest number of words, while anxious high perceivers produced the least. With
296
THOMAS
D.
BORKOVEC
increasing stress, high perceivers displayed increasing signs of anxiety, while low perceivers displayed a general decrease. These results suggest that the autonomic perception characteristic may play a significant role in determining the form of overt anxiety behavior during stress. Furthermore, given that the interaction test lasted about 15 minutes, the similarity between APQ effects on word production in this study and approach behavior in the Stone and Borkovec (1975) study was striking. In both investigations, high perceivers displayed increases in behavioral indicants of anxiety as length of CS exposure increased to 15 minutes.
5. Perceived Arousal and Attention-Focusing Manipulations An alternative methodological approach to false physiological feedback manipulations involves increasing or decreasing the saliency of autonomic cues through instructions to focus attention toward or away from those cues. In an unpublished pilot study, 42 snake-phobic males (half high and half low in APQ score) were randomly assigned to one of three instructional conditions in a posttest-only design. All subjects received instructions that were very high in demand/suggestion for approach behavior. One group was additionally told to concentrate on the autonomic cues that might occur during the snakeexposure test. A second group was instructed to concentrate on the external stimuli associated with their fear. The third control group received no attending instructions. The high-demand instructions (plus the use of a female experimenter) resulted in a ceiling effect on approach behavior. Of the subjects 78%, evenly distributed across instructional conditions, achieved the maximal approach score. The main effect of instructional set was significant, however, on a specific APQ measure asking the subjects to report their awareness of autonomic cues during the snake exposure (F = 10.23, P < .01). Quite contrary to expectations, subjects instructed to attend to autonomic cues reported significantly less awareness of those cues, and subjects told to focus on external fear cues indicated significantly greater awareness, than the control subjects. In addition, the general APQ factor interacted significantly with instructional set on both latency to touching the snake (F = 4.30, P < .05) and pulse-rate change from baseline to the termination of the posttest (F = 3.40, P < .05). Table 3 presents the mean latency and pulse-rate change scores for high and
PHYSIOLOGICAL AND COGNITIVE PROCESSES IN THE REGULATION OF ANXmTY
297
TABLE 3 Mean Latency in Touching Feared Object (in Seconds) and Pulse-Rate Change Scores (Beats/Minute) for Low and High Autonomic Perceivers in the Three Attention-Focusing Conditions
Attention-focusing condition Autonomic perception group Latency Low High Pulse-rate change Low High
Autonomic cues
External cues
Control
11.3 33.6
14.0 17.8
29.5 10.3
-1.16 5.72
-4.56 0.0
6.28 -3.44
low autonomic-perception subjects in each of the three instructional conditions. Pulse rate decreased among high perceivers when no attending instructions had been given, while pulse rate showed no change under external cue attention and increased under autonomic cue attention. Latency to touching the snake paralleled these findings exactly. Among low perceivers, pulse rate decreased when the subjects were instructed to attend to external cues, increased under control instructions, and showed no change when the subjects were instructed to attend to autonomic cues. Either type of attending instructions resulted in lower latencies than the control instructions. No satisfactory interpretation of this outcome is apparent. We are currently replicating the study with speech-anxious subjects, and additional data will hopefully clarify the functional effects of attending instructions.
6. Patterns of Autonomic Perception Several studies have provided evidence that autonomic perception is an important subject characteristic related to anxiety process. It is somewhat surprising that these studies found such relationships despite dependence on only the total APQ score. Individuals differ in
298
THOMAS
D.
BORKOVEC
TABLE 4 APQ Items and Labels Associated with Extreme (0 and 9) Scores
1. 2. 3. 4. 5. 6. 7. B. 9. 10. 11. 12. 13. 14. 15. 16. 17. lB. 19. 20. 21.
Awareness of many bodily reactions (very few-very many). Frequency of awareness of those reactions (never-always). Face becoming hot (no change-very hot). Hands becoming cold (no change-very cold). Perspiration (not at all-a great deal). Mouth becoming dry (never-always). Muscles becoming tense (none-a great deal). Headache (never-always). Changes in heart action (never-always). Increases in rate of heartbeat (no change-great acceleration). Increases in intensity of heartbeat (no change-increases to extreme pounding). Changes in breathing (never-always). Breathing becoming more rapid (no change-very rapid). Breathing becoming more deep (no change-much more deep). Breathing becoming more shallow (no change-much more shallow). Blood rushing to head (never-always). Lump in throat (never-always). Stomach becoming upset (not at all-very upset). Sinking or heavy feeling in stomach (never-always). Difficulty in talking (never-always). Bodily reactions being bothersome (not bothered-bothered very much).
the reactiveness of various autonomic responses during stress, and some individuals show relatively consistent reactions over time and over different stressors (Lacey, Bateman, and Van Lehn, 1953). It might be expected, therefore, that different individuals may report awareness of quite different autonomic cues as well. In an exploration of this question, the APQ item scores from the normal and clinical samples tested earlier were recently submitted to a Q factor analysis (Stephenson, 1953). Each person's item scores were correlated with each other person's scores. The intercorrelation matrix was then factor-analyzed with persons as the variables and items as the observations. Following a principle-axis solution, a varimax rotation produced orthogonal factors which represented groups or types of persons having a similar pattern of item scores. Each pattern of item scores associated with each type was then estimated by the weighting of each item score of each of the persons most highly associated with a given type by the degree to which they were loaded on that type. The weighted scores summed across each item produced an item array of
PHYSIOLOGICAL AND COGNITIVE PROCESSES IN THE REGULATION OF ANxmTY
299
weighted scores for each type. The arrays were then converted to zscores, which represent the degree to which each item contributes to the characterization of a given type (cf. Talbott, 1971). In Sample 1, the APQ item scores from 100 female and 100 male students from the 1971 spring semester were randomly selected and submitted to the Q analysis. Sample 2 (100 females and the remaining 81 males from the same semester) was similarly analyzed. Finally, cross-validation Sample 3 of 100 females and 100 males was randomly selected from the 1971 fall semester. Because of significant differences between sexes on specific APQ items (see Table 2), analyses of female and male data were performed separately. While various solutions were calculated, a three-factor solution was chosen for simplicity of presen tation. Table 4 provides a summary of the 21 APQ items and the labels associated with extreme scores (0 and 9) as a reference for subsequent figures and tables. Figures 1-5 present the z-score profile types for
3.0 0-0
2.0
Sample # I
. - . Sample # 2 . - - . Sample # 3
/.0 ILl
a:
0
(.) rJ)
0
N
-/.O
-2.0
-3.0
2
3 4
5
6
7
8
9
10 II 12 13 14 15 16 17 18 19 20 21
APQ FIGURE 1.
ITEM
Type I z-score profiles of APQ items from the three normal-female samples.
300
THOMAS
D.
BORKOVEC
3.0
2.0
1.0 I
UJ 0:
0
(.) (J)
I
",e---\ 1I
I
\
Y"
• .,-
\
II
\
0
\
0.,
:::~\.~I /"J{./~ ,,\'6/j~ I\ 0-l,v../--· \ / • 0 \
,
I \
Q
\
•
1/ •
I \ I II
I
-2.0
¥
-3.0
I
2
I
I
3 4
I
5
I
6
I
7
I
8
I
9
I
2.
I
I
I
I
I
I
I
I
I
1---'
10 II 12 13 14 15 16 17 18 19 20 21
APQ FIGURE
\
."
\\ ,,
-1.0
I I
I
II
N
I \
I \
I
0
~
I
ITEM
Type II z-score profiles of APQ items from the three normal-female samples.
females and males which resulted from the Q analysis and which were replicated in all three samples. In the description of each profile, those APQ items whose z-scores were above + 1.0 and were replicated at that z-score level across emerging factors in all three samples were considered to be particularly characteristic of that profile, while items whose z-scores were below -1.0 and were replicated across samples were considered uncharacteristics of that profile. Figure 1 presents Type I female profiles from the three samples. The displayed profiles, emerging as the first factor type in each sample analysis and accounting for the greatest amount of variance in each sample (.155%, .139%, and .152% in Samples 1, 2, and 3, respectively), are highly similar across the samples. Type I females were found to be associated with high awareness of stomach activity and perspiration when anxious and with low awareness of breathing and blood rushing to the head. Figure 2 displays Type II female profiles. The sample profiles emerged as the second factor type in Samples 1 and 2 and as the third
PHYSIOLOGICAL AND COGNITIVE PROCESSES IN THE REGULATION OF ANXIETY
301
factor type in Sample 3 (% variance = .150, .128, and .116, respectively). Type II females were characterized by high awareness of heart activity and muscle tension and by low awareness of headaches. Figure 3 presents Type III female profiles. These highly similar profiles emerged as the third factor type in Samples 1 and 2 and the second type in Sample 3 (% variance = .092, .114, and .136, respectively). Awareness of heart and stomach activity was characteristic of Type III females, while no item was particularly uncharacteristic of that group. Figure 4 displays Type I male profiles from the three samples. As was the case with the female analyses, the first extracted factor types from the male samples (% variance = .176, .205, and .190, for Samples 1, 2, and 3) bear marked similarity to one another. Unlike Type I female profiles, however, Type I males were characterized solely by high awareness of heart activity and low awareness of headaches and shallow breathing. Type II males profiles are prese~ted in Figure 5. These similar 3.0
2.0
1.0 UJ 0:
0
(.) (J)
0
N
-1.0
-2.0
-3.0 2 FIGURE 3.
3 4
5
6
7
8
9 10 II 12 13 14 15 16 17 18 19 20 21 APQ ITEM
Type III z-score profiles of APQ items from the three normal-female samples.
302
THOMAS
D.
BORKOVEC
3.0
2.0
1.0 LLI
0:::
0
U
en
0
N
-1.0
~Jrv ~\~! ~ \ •
I I I I
lit
o
1/
•
-2.0
-3.0 2 FIGURE 4.
3
4
5
6
7
8
9 10 II 12 13 14 15 16 17 18 19 20 21 APQ ITEM
Type I z-score profiles of APQ items from the three normal-male samples.
profiles emerged as the second factor type in Samples 2 and 3 and as the third type in Sample 1 (% variance = .150, .142, and .119, respectively). Stomach activity, perspiration, and frequency of noticing bodily reactions when anxious were characteristic of Type II males, while headaches and breathing activity were uncharacteristic of that type. The remaining factor types emerging from the male analyses were unlike other extracted types, nor were they replicated across samples. A third profile type does not appear to be present among normal males. The results of these analyses indicate that normal subjects clustered themselves into distinct groups differing in the pattern of autonomic cues perceived when anxious, that those group clusterings were replicated across separate samples, and that females and males differed not only in absolute APQ items scores but also in their patterns of autonomic-cue perception.
PHYSIOLOGICAL AND COGNITIVE PROCESSES IN THE REGULATION OF ANXIETY
303
Q-factor analyses were similarly applied to the clinic APQ data collected from 1971 to 1973. Table 5 presents the z-score profiles of the three emerging factor types for female and male patients. Type I female patients were characterized by awareness of muscle tension, headaches, and stomach activity. Breathing activity and blood rushing to the head were relatively uncharacteristic of that group. In addition, they reported many and frequently occurring bodily reactions when anxious. Type II female patients were associated with high awareness of perspiration, muscle tension, and increasing heart rate and with low awareness of cold hands, headaches, and deep breathing. Cold hands, awareness of many bodily reactions, and being bothered by those reactions characterized Type III, while headaches, deep breathing, and upset stomach were particularly unassociated with that profile. Type I male patients were associated with reports of stomach and heart activity and awareness of many and frequently occurring bodily
3.0
2.0
1.0 UJ 0::
0
t)
0
0
~,,
•
/.,/I
, .0 - .
'
\ \
•
-1.0
...
,
o
I
•
~_-. ,\ 'I
•
0
FIGURE
5.
2
3
4
/0
\
..p /'
'i~
•
•
I
I
,\
•
-2.0
-3.0
/..~ ./~\ / \ /• // ):Q
A
I-
N
p,
0 __· ,
5
6
7
8
9 10 II 12 13 14 15 16 17 18 19 20 21 APQ ITEM
Type II z-score profiles of APQ items from the three normal-male samples.
304
THOMAS D. BORKOVEC
TABLE 5 Z-Scores for Each of the Three Types for Female and Male Clinic Patients Male patients
Female patients APQ item
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 N % Variance:
Type I 1.0 1.1 -0.2 -0.9 0.5 -0.8 1.6 1.3 -0.6 -0.2 -0.1 -1.4 -1.4 -0.8 -1.6 -1.0 0.0 1.1 1.3 0.9 0.4 17 .175
Type II
Type III
0.3 0.5 -0.1 -2.8 1.3 -0.9 1.1 -1.8 0.8 1.1 0.9 0.6 0.5 -1.1 0.6 -0.5 -0.1 0.3 0.4 -0.2 -0.8 8 .152
1.2 0.8 0.8 1.7 0.4 0.0 0.9 -1.6 -0.3 -0.2 -0.1 -0.8 -0.3 -2.7 -0.1 -0.1 -0.2 -1.2 0.3 0.2 1.2 8 .137
Type I 1.1 1.5 -1.4 -1.2 -0.2 -0.7 0.7 -1.9 1.1 1.1 1.0 -0.4 -0.7 -0.8 -1.1 -1.3 0.4 0.4 1.2 0.2 0.8 19 .239
Type II 0.8 1.2 0.2 -0.3 0.8 -0.3 1.8 -0.6 -1.9 -1.8 -1.7 0.0 -0.1 -0.8 0.1 -0.7 0.9 0.9 0.0 0.8 1.1 9 .124
Type III 0.1 0.8 -2.2 0.7 0.4 1.1 0.6 0.7 0.4 0.7 0.3 -0.1 -0.2 -0.2 -1.6 -2.1 -0.4 1.5 1.0 -1.3 -0.1 6 .119
reactions, and they were unassociated with awareness of hot face, cold hands, headaches, shallow breathing, and blood rushing to the head. Muscle tension and frequent bodily reactions were characteristic of Type II male patients, while heart activity was uncharacteristic. Type III involved stomach activity and mouth dryness, with low awareness of hot face, shallow breathing, blood rushing to the head, and talking difficulty. Unfortunately a cross-validation sample has not yet been obtained from the clinic population. Less confidence can, therefore, be placed on the reliability of the patient profiles in contrast to the five (of six) replicated types found in the normal samples. The available profiles do suggest, however, that psychiatric patients fall into autonomicperception types which differ between sexes and are characterized by
PHYSIOLOGICAL AND COGNITIVE PROCESSES IN THE REGULATION OF ANXIETY
305
patterns of subjective autonomic activity that do not match those of normal samples. Of particular interest was the fact that four of the six patient types loaded significantly on the three general APQ items (awareness of many reactions, frequency of awareness, and bothersomeness of reactions). In contrast, only frequency of awareness was loaded in normal profiles, and its loading occurred only in one of the five profile types. We would tend to interpret this finding as supporting our assumption that autonomic perception, whatever specific cues may be involved, is an important variable contributing to clinically distressing behavior. We have argued that individuals differ in relative patterning of physiological, cognitive, and behavioral fear components. In addition, Lacey's research indicates that individuals display different patterns within the physiological component. The factor analytic studies presented above suggest that autonomic perception, a mixture of physiological and cognitive components, is also characterized by various patterns. It will no doubt turn out, therefore, that the model of anxiety presented earlier is highly oversimplified. However, the importance of individual differences in multiple response-group components and their interactions will remain. Hopefully, the APQ profiles will facilitate more precise research regarding the role of autonomic perception and physiological cues in anxiety maintenance and reduction.
III.
SUMMARY AND CONCLUSIONS
Anxiety has been assumed to involve three separate but interacting components (physiological, cognitive, and overt behavioral) in response to external and internal fear cues. The separateness of the responses allows for individual differences in anxiety-response patterns and hence in the conditions of anxiety maintenance and reduction. The interactions among the three response groups allow for nine sequential relationships, each relationship potentially contributing to the maintenance or the reduction of anxiety. The model provides a perspective from which to view the current behavior-therapy literature and recent developments in self-control procedures. Our anxiety-research program has focused on a subset of the response-component interactions (the role of physiological and cognitive responses) and the contribution of individual differences in
306
THOMAS
D.
BORKOVEC
perceived and actual physiological activity to those relationships. Manipulations of physiological and cognitive response have included the use of cognitive avoidance responses, expectancy, false physiological feedback, misattribution, and demand/suggestion. Individual differences in physiological cues have been investigated primarily by the use of an Autonomic Perception Questionnaire. To date, the effects of demand/suggestion, duration of CS exposure, and attention-focusing instructions on subjects scoring high or low in autonomic perception have been assessed. In terms of anxiety maintenance and reduction, the studies have resulted in the following tentative conclusions: 1. Cognitive avoidance responses can maintain the physiological fear component even in subjects for whom physiological reactions to the feared stimulus are relatively weak (snake phobics). Systematic desensitization and implosive therapy procedures, on the other hand, lead to decreases in process and outcome autonomic measures with such subjects. 2. Expectancy for improvement is unrelated to outcome improvement in the treatment of sleep disturbance. Positive expectancy, however, results in lower physiological activity during therapy than neutral expectancy in the treatment of snake phobia. 3. Pretesting sensitizes snake phobics, resulting in maintained physiological reactivity to symbolic CS presentations and a mitigation of false physiological-feedback effects on approach behavior. False feedback manipulations will be effective in increasing the approach behavior of snake phobics only if no pretest is administered. 4. False feedback indicating increased physiological arousal among subjects for whom the physiological fear component is strong (speech phobics) results in maintenance of behavioral and self-reported fear. False feedback indicating decreased arousal for such subjects has no effect on fear beyond repeated exposure to the feared situation. 5. False feedback effects on either snake or speech phobics do not occur during the feedback presentation but during a subsequent exposure to the feared situation, suggesting a cognitive mediation of those effects. 6. Misattribution manipulations are ineffective if a strong physiological fear component is present. The combination of high physiologi-
PHYSIOLOGICAL AND COGNITIVE PROCESSES IN THE REGULATION OF ANXIETY 307
cal arousal and attribution of high-arousal effects to an irrelevant stimulus results in increased behavioral signs of fear. 7. Socially nonanxious males will display increased physiological activity upon repeated presentations of a nonreceptive female, and demand instructions influence whether that increase occurs in anticipation of, or in response to, the social-interaction situation. 8. The effects of repeated CS exposure and demand characteristics on snake phobia are influenced by subject characteristics of perceived and actual physiological activity. In particular, accurate perceivers display reductions of physiological reactivity to repeated CS exposures, while inaccurate perceivers show a maintenance. 9. Duration of CS exposure can result in an incubation of behavioral and physiological fear components, especially among high autonomic perceivers. 10. Instructions to attend to autonomic cues during exposure to the feared situation have different effects on high and low perceivers. In particular, such instructions result in increased behavioral and physiological fear among high perceivers. 11. There appear to be replicable patterns of autonomic perception among normal subjects, and the patterns differ from those found among clinical subjects. From these specific conclusions, two more general conclusions can be drawn. First, we have obtained no data which suggest a revision of one of our central hypotheses. As long as the physiological component is strongly present in the individual's immediate anxiety reaction, simple manipulations of the other two components will be ineffective or, at the very least, inefficient; reduction of that component allows the other manipulations to influence those remaining components. This is not to suggest that intensive training in modified selfverbalizations, for example, will not produce therapeutic benefit. The data do argue, however, that we have probably not yet identified the best means of reducing relevant response components and that consideration of individual differences in those components may be necessary to facilitate our search. Second, it is remarkable how many cognitive manipulations (cognitive avoidance responses, false feedback, misattribution, demand/expectancy, attention focusing), as well as the subject characteristic of autonomic awareness, contribute to anxiety maintenance. Particular cognitive styles, already existing or
308
THOMAS
D.
BORKOVEC
experimentally induced, are quite capable of precluding the extinction of anxiety, despite repeated, nonreinforced CS exposure. While we have been very adept at formulating ingenious methods of CS exposure (e.g., desensitization, implosive therapy, and flooding), we knew very little about the degree of functional exposure taking place. What the client does, cognitively as well as behaviorally, during that exposure appears to be of major importance. In addition to direct modification of relevant, immediate anxiety components, therefore, elimination of identifiable maintaining components and their conditions must occur concurrently.
ACKNOWLEDGMENTS
Several of the research studies cited in this paper were supported in part by Biomedical Sciences Support Grant #FR-07035 from the Bureau of Health Professions, Education and Manpower Training, National Institutes of Health, made available in the form of a small grant awarded by the Graduate College of the University of Iowa and Grant #MH-24603-01 from the National Institute of Mental Health. The author wishes to express deep appreciation to his students, who made numerous contributions to the research and conceptualizations presented in this paper: Jon B. Grayson, Gerald T. O'Brien, Sidney D. Nau, Kathy M. Slama, Shan W. Steinmark, Norman M. Stone, and Robert L. Wall. They are the "we" mentioned throughout the text. Special thanks are due to Michael G. Grisham for his thoughtful comments and support during the preparation of the manuscript.
REFERENCES BEM, D. J. Self-perception theory. In L. BERKOWITZ (Ed.), Advances in experimental social psychology (Vol. 6). New York: Academic Press, 1972, pp. 1-62. BERGMAN, J. S., AND JOHNSON, H. J. The effects of instructional set and autonomic perception on cardiac control. Psychophysiology, 1971,8, 180-190. BERNSTEIN, D. A. Behavioral fear assessment: Anxiety or artifact? In H. ADAMS AND P. UNIKEL (Eds.), Issues and trends in behavior therapy. Springfield: Thomas, 1973. BERNSTEIN, D. A., AND PAUL, G. L. Some comments on therapy analogue research with small animal "phobias." Journal of Behavior Therapy and Experimental Psychiatry, 1971,2, 225-237.
PHYSIOWGICAL AND COGNITIVE PROCESSES IN THE REGULATION OF ANXIETY
309
BLACK, A. H. Heart rate changes during avoidance learning in dogs. Canadian Journal of Psychology, 1959,13, 229-242. BLANCHARD, E. B., YOUNG, L. D., AND McLEOD, P. Awareness of heart activity and selfcontrol of heart rate. Psychophysiology, 1972,9, 63--68. BORKOVEC, T. D. The comparative effectiveness of systematic desensitization and implosive therapy and the effect of expectancy manipulation on the elimination of fear. Doctoral dissertation, University of Illinois (Urbana), 1970. BORKOVEC, T. D. Effects of expectancy on the outcome of systematic desensitization and implosive treatments for analogue anxiety. Behavior Therapy, 1972,3,29--40. BORKOVEC, T. D. The effects of instructional suggestion and physiological cues on analogue fear. Behavior Therapy, 1973a,4, 185-192. BORKOVEC, T. D. The role of expectancy and physiological feedback in fear research: A review with special reference to subject characteristics. Behavior Therapy, 1973b,4, 491-505. BORKOVEC, T. D. Heart-rate process during systematic desensitization and implosive therapy for analogue anxiety. Behavior Therapy, 1974,5, 636--641. BORKOVEC, T. D., FLEISCHMANN, D. J., AND CAPUTO, J. A. The measurement of anxiety in analogue social situation. Journal of Consulting and Clinical Psychology, 1973,41, 157161. BORKOVEC, T. D., AND GLASGOW, R. E. Boundary conditions of false heart-rate feedback effects on avoidance behavior: A resolution of discrepant results. Behaviour Research and Therapy, 1973,11, 171-177. BORKOVEC, T. D., KALOUPEK, D. G., AND SLAMA, K. The facilitative effect of muscle tension-release in the relaxation treatment of sleep disturbance. Behavior Therapy, 1975,6,301-309. BORKOVEC, T. D., STONE, N. M., O'BRmN, G. T., AND KALOUPEK, D. G. Evaluation of a clinically relevant target behavior for analogue outcome research. Behavior Therapy, 1974,5, 504-514. BORKOVEC, T. D., WALL, R. L., AND STONE, N. M. False physiological feedback and the maintenance of speech anxiety. Journal of Abnormal Psychology, 1974,83, 164-168. BOULOUGOURIS, J. c., MARKS, I. M., AND MARSET, P. Superiority of flooding (implosion) to desensitization for reducing pathological fear. Behaviour Research and Therapy, 1971,9, 7-16. BRESNITZ, S. Incubation of threat: Duration of anticipation and false alarm as determinants of the fear reaction to an unavoidable frightening event. Journal of Experimental Research in Personality, 1967,2, 173-179. BYRNE, D. The repression-sensitization scale: Rationale, reliability, and validity. Journal of Personality, 1961, 29, 334-339. CAUTELA, J., FLANNERY, R, AND HANLEY, E. Covert modeling: An experimental test. Behavior Therapy, 1974,5, 494--502. CRAIGHEAD, W. E. The role of muscular relaxation in systematic desensitization. In R D. RUBIN, J. P. BRADY, AND J. D. HENDERSON (Eds.), Advances in behavior therapy (Vol. 4). New York: Academic Press, 1973, pp. 177-197. DmNSTBmR, R A., AND MUNTER, P. O. Cheating as a function of labeling of natural arousal. Journal of Personality and Social Psychology, 1971, 17, 208-213. ELLIS, A. Rational psychotherapy. Journal of General Psychology, 1958,59,35-49. EYSENCK, H. J. A theory of the incubation of anxiety/fear responses. Behaviour Research and Therapy, 1968,6, 309-321. FRmDMAN, H. J. Patient-expectancy and symptom reduction. Archives of General Psychiatry, 1963,8, 61-67.
310
THOMAS
D.
BORKOVEC
GANTT, W. H., AND DYKMAN, R. A. Experimental psychogenic tachycardia. In P. H. HOCK AND J. ZUBIN (Eds.), Experimental psychopathology. New York: Grone and Stratton, 1957. GAUPP, L. A., STERN, R. M., AND GALBRAITH, G. G. False heart-rate feedback and reciprocal inhibition by aversive relief in the treatment of snake avoidance behavior. Behavior Therapy, 1972, 3, 7-20. GELDER, M. G., BANCROFT, J. H. J., GATH, D. H., JOHNSTON, D. W., MATHEWS, A. M., AND SHAW, P. M. Specific and nonspecific factors in behavior therapy. British Journal of Psychiatry, 1973,123, 445--462. GOLDSTEIN, A. P. Patient's expectancies and nonspecific therapy as a basis for (un) spontaneous remission. Journal of Clinical Psychology, 1960,16, 399-403. GOLDSTEIN, A. P., AND SHIPMAN, W. G. Patient expectancies, symptom reduction, and aspects of the initial psychotherapeutic interview. Journal of Clinical Psychology, 1961,17, 129--133. JONES, M. C. The elimination of children's fears. Journal of Experimental Psychology, 1924, 7, 382-390. KENT, R. N., WILSON, G. T., AND NELSON, R. Effects of false heart-rate feedback on avoidance behavior: An investigation of "cognitive desensitization." Behavior Therapy, 1972,3, 1-6. LACEY, J. I., BATEMAN, D. E., AND VAN LEHN, R. Autonomic response specificity. Psychosomatic Medicine, 1953,15, 8-21. LADER, M. H., AND MATHEWS, A. M. A physiological model of phobic anxiety and desensitization. Behaviour Research and Therapy, 1968,6, 411-421. LANG, P. J. Fear reduction and fear behavior: Problems in treating a construct. In J. M. SHLIEN (Ed.), Research in psychotherapy. Washington, D.C.: American Psychological Association, 1968, pp. 9~102. MANDLER, G., AND KREMEN, I. Autonomic feedback: A correlational study. Journal of Personality, 1958,26, 388-399. MANDLER, G., MANDLER, J. M., AND UVILLER, E. T. Autonomic feedback: The perception of autonomic activity. Journal of Abnormal and Social Psychology, 1958,56,367-373. MATHEWS, A. M. Psychophysiological approaches to the investigation of desensitization and related procedures. Psychological BUlletin, 1971, 76, 73-91. McFALL, R. M., AND TWENTYMAN, C. T. Four experiments on the relative contributions of rehearsal, modeling, and coaching to assertion training. Journal of Abnormal Psychology, 1973,81, 199--218. MILLER, B. V., AND LEVIS, D. J. The effect of varying short visual exposure times to a phobic test stimulus on subsequent avoidance behavior. Behaviour Research and Therapy, 1971,9,17-21. MOWRER, O. H. On the dual nature of leaming-a re-interpretation of "conditioning" and "problem-solving." Harvard Educational Review, 1947,17, 102-148. NISBETT, R. E., AND SCHACHTER, S. Cognitive manipulation of pain. Journal of Experimental Social Psychology, 1966, 2, 227-236. PATTERSON, G. R. The aggressive child: Victim and architect of a coercive system. In L. HAMERLYNCK, L. HANDY, AND J. MASH (Eds.), Parenting: The change, maintena-nce, and directions of healthy family behavior. Champaign, Ill.: Research Press, 1975. PAUL, G. L. Behavior modification research: Design and tactics. In C. M. FRANKS (Ed.), Behavior therapy: Appraisal and status. New York: McGraw-Hill, 1969a, pp. 29-62. PAUL, G. L. Outcome of systematic desensitization I: Background, procedures, and uncontrolled reports of individual treatment. In C. M. FRANKS (Ed.), Behavior therapy: Appraisal and status. New York: McGraw-Hill, 1969b, pp. 63-104_
PHYSIOLOGICAL AND COGNITIVE PROCESSES IN THE REGULATION OF ANXIETY
311
PIPER, W. E., AND WOGAN, M. Placebo effect in psychotherapy: An extension of earlier findings. Journal of Consulting and Clinical Psychology, 1970,34,447. RANKIN, N. 0., NOMIKOS, M. S., OPTON, E. M., AND LAZARUS, R. S. The roles of surprise and suspense in stress reaction. Paper presented at the meeting of the Western Psychological Association, June, 1964. RESCORLA, R. A., AND SOLOMON, R. L. Two-process learning theory: Relationships between Pavlovian conditioning and instrumental learning. Psychological Review, 1967,74, 151-182. ROHRBAUGH, M., AND RICCIO, D. Paradoxical enhancement of learned fear. Journal of Abnormal Psychology, 1970,75, 210--216. ROHRBAUGH, M., RICCIO, D., AND ARTHUR, A. Paradoxical enhancement of conditioned suppression. Behaviour Research and Therapy, 1972, 10, 125-130. Ross, L., RODIN, J., AND ZIMBARDO, P. G. Toward an attribution therapy: The reduction of fear through induced cognitive-emotional misattribution. Journal of Personality and Social Psychology, 1969, 12, 279-288. SCHACHTER, S. The interaction of cognitive and physiological determinants of emotional state. In L. BERKOWITZ (Ed.), Advances in experimental social psychology (Vol. 1). New York: Academic Press, 1964, pp. 49--80. SINGERMAN, K. Misattribution and placebo effects in speech anxiety. Unpublished masters thesis, University of Iowa, 1974. SLAMA, K. Studies comparing stimulus control and progressive relaxation procedures in the treatment of sleep disturbance. Unpublished masters thesis, University of Iowa, 1975. SPIELBERGER, C. D. Theory and research on anxiety. In C. D. SPIELBERGER (Ed.), Anxiety and behavior. New York: Academic Press, 1966, pp. 3-20. SPIELBERGER, C. D. (Ed.). Anxiety: Current trends in theory and research (Vol. 1 and 2). New York: Academic Press, 1972. STAMPFL, T. G., AND LEVIS, D. J. The essentials of implosive therapy: A learning theory based psychodynamic behavioral therapy. Journal of Abnormal Psychology, 1967,72, 496-502. STEINMARK, S. W., AND BORKOVEC, T. D. Active and placebo treatment effects on moderate insomnia under counterdemand and positive demand instructions. Journal of Abnormal Psychology, 1974,83, 157-163. STEPHENSON, W. The study of behavior. Chicago: University of Chicago Press, 1953. STONE, N. M., AND BORKOVEC, T. D. The paradoxical effect of brief CS exposure on analogue phobic subjects. Behaviour Research and Therapy, 1975,13, 51-54. STORMS, M. D., AND NISBETT, R. E. Insomnia and the attribution process. Journal of Personality and Social Psychology, 1970,16, 319-328. SUSHINSKY, L. W., AND BOOTZIN, R. R. Cognitive desensitization as a model of systematic desensitization. Behaviour Research and Therapy, 1970,8,29-34. TALBOTT, A. D. Q technique and its methodology: A brief introduction and consideration. Paper presented at the annual meeting of the American Educational Research Association, New York, February, 1971. ULLMANN, L. P. An empirically derived MMPI scale which measures facilitationinhibition of recognition of threatening stimuli. Journal of Clinical Psychology, 1962, 18, 127-132. VALINS, S., AND RAY, A. A. Effects of cognitive desensitization on avoidance behavior. Journal of Personality and Social Psychology, 1967,7, 345-350. WATSON, J. B., AND RAYNER, R. Conditioned emotional reactions. Journal of Experimental Psychology, 1920,3, 1-14.
312
THOMAS
D.
BORKOVEC
WILKINS, W. Desensitization: Social and cognitive factors underlying the effectiveness of Wolpe's procedure. Psychological Bulletin, 1971, 76, 311-317 WILKINS, W. Expectancy of therapeutic gain: An empirical and conceptual critique.
Journal of Consulting and Clinical Psychology, 1973,40, 69-77. WILSON, G. T., AND DAVISON, G. C. Process of fear reduction in systematic desensitization: Animal studies. Psychological Bulletin, 1971, 76, 1-14. WOLPE, J. Psychotherapy by reciprocal inhibition. Stanford, Calif.: Stanford University Press, 1958.
8
Dreaming: Experimental Investigation of Representational and Adaptive Properties DAVID
B. COHEN
This chapter organizes into three sections diverse strategies for the empirical investigation of dreaming. The first is addressed to the question of how the dream becomes known to the individual and the investigator though the process of recall. The second deals with the assumption that dreaming reflects or represents in an explicable fashion factors such as personality traits, interpersonal events, and physiological activity. The third discusses two hypotheses: (1) that dreaming is a psychological process which can be studied apart from the neurophysiological process of sleep and (2) that dreaming is a functional process, specifically that dreaming is adaptive. For purposes of discussion, dreaming refers to a psychological process (analogous to thinking) presumably inherent in the neurophysiological activity of the sleeping nervous system. Dream or dream experience refers to conscious awareness of dreaming while it is occurring. The dream report refers to the communication about a dream.
1.
DREAM RECALL
A. The Role of Repression Prior to the rise of modem sleep research the most common explanation of dream recall rested on the notion of repression. Freud DAVID B. COHEN
.
Department of Psychology, University of Texas, Austin, Texas.
313
314
DAVID
B. COHEN
(1953) concluded "that during the night the resistance loses some of its power ... [and] ... having gained its full strength at the moment of awakening, it at once proceeds to get rid of what it was obliged to permit while it was weak" (p. 526). Modem empirical research has attempted to validate this general hypothesis about dream recall by testing the following kinds of hypotheses: (1) infrequent dream recallers should be repressors and (2) presleep stress designed to increase the probability of generating dreams that are "candidates for repression" (Goodenough, 1967, p. 139) will, especially for repressortype subjects, be associated with a diminution of dream recall. The first hypothesis, that infrequent dream recall is a sign of repressiveness, has faired rather poorly (Cohen, 1974c). Measures of frequency of dream recall (derived from questionnaire and home dream-diary methods) have been shown to correlate at best around .25 with various measures of repressiveness (e.g., Byrne repressionsensitization, Barron ego strength, field dependence, leveling, etc.), thus leaving most of the variation in dream-recall frequency unaccounted for. The second hypothesis has fared little better. It assumes that stressful presleep conditions will generate more threatening dreams (Goodenough, Witkin, Lewis, Koulack, and Cohen, 1974), which are more likely to be repressed (Cartwright, Bernick, Borowitz, and Kling, 1969). While some studies tend to support this hypothesis (Cartwright et al., 1969; Cohen, 1972), most do not (Cohen, 1974a,b; Cohen and Cox, 1975; Cohen and Wolfe, 1973; Goodenough et al., 1974). There is, in fact, evidence that subjects classified as "repressors," as well as "sensitizers," tend to have more dream recall under distressing presleep conditions. This finding has obtained whether classification was made on the basis of established measures (e.g., repression-sensitization) or presumptive measures (e.g., dream-recall frequency). Figure 1 shows two sets of idealized relationships, one derived from the results summarized above and one from results to be described in the next section. In the first set the dependent variable is situational diary dream recall, and in the second set it is dream affect (relative unpleasantness reported for both laboratory and home dreams). For both sets, the predictor variable is presleep condition (positive versus negative) either subjectively reported or experimentally manipulated. Note that the first set of relationships (a) shows an
315
DREAMING
a
Infreq Rec-". ~
~',
Freq Rec""
......... 10 Erno
~"
\
High Erno,"\ ..... ~
.... "
.,,/'
,,;'''""
,,~ ~
•
._-""..._-----_._--
Negative
Positive
PRESLEEP CONDITION b
Low
---
ErnO~
---
/ I
FIGURE 1. Empirically derived idealized relationships denoting the differential effects of presleep condition on situational dream recall (a) and dream affect (b) for subjects preselected on the basis of habitual dream recall frequency or for the trait of emotionality.
I
I
I
I
/
/
/
I
Negative
Positive
PRESLEEP CONDITION
316
DAVID
B.
COHEN
interaction effect of condition and habitual dream-recall frequency on situation dream recall, but no interaction effect for condition and the trait of emotionality or sensitization. For the second set (b), however, there is an interaction effect of condition and emotionality on dream affect, but no interaction effect for condition and habitual dream-recall frequency. That is, specifying condition and habitual dream-recall frequency allows better prediction of situational dream recall than does specifying condition and the trait of emotionality. Conversely, specifying condition and emotionality allows better prediction of dream affect than does specifying condition and habitual dream-recall frequency. While these conclusions are derived from an investigation of a limited set of conditions and dispositional variables, they are consistent with the low or zero correlations obtained between measures of emotionality and dream-recall frequency, thus supporting the contention that these dimensions are relatively independent. What all these data suggest is that the quantitative aspect of dream recall (amount, frequency) is usually determined by factors largely independent of defensive processes that may contribute to dream content (Cohen, 1974c). Consider one further example of the attempt to draw a theoretical link between repression and dream recall. Antrobus, Antrobus, and Singer (1964) found a positive relationship between attempts to avoid thinking about a topic and the amount of eye-movement activity. Antrobus, Dement, and Fisher (1964) reported that infrequent dreamrecallers had significantly more eye-movement activity during REM sleep (REM density) than had frequent dream-recallers. On the basis of these findings, the investigators suggested that REM density might be a sign of oculomotor avoidance, that is, a manifestation of a defensive process that could account for infrequency of dream recall. However, Cohen and Cox (1975) found no relationship between REM density and habitual frequency of dream recall. Further, while lowneuroticism ("repressor") subjects had significantly greater REM density under conditions of presleep stress than under positive conditions, there was no corresponding difference between these two groups in dream recall elicited from REM and NREM sleep. In addition, other investigators (e.g., Molinari and Foulkes, 1969) have found no differences in REM dream recall for awakenings made during eye-movement activity or quiescence. In short, there is little or no reliable evidence for predictions based on the oculomotor-avoid-
DREAMING
317
ance hypothesis which requires a positive relationship between repressiveness, REM density, and dream forgetting. This conclusion has no bearing on the validity of the repression concept in other areas of psychological functioning (including special situations like dream reporting in psychotherapy), nor on the question of defensive distortion of reality in the dream experiences of certain individuals under certain conditions. Finally, this conclusion does not rule out the hypothesis that certain formal properties of dream experiences (bizarreness, emotional intensity, personal relevance, etc.) affect the probability that the dream will be recalled.
B. Alternative Factors: Salience and Interference What alternative explanations for dream recall have emerged from recent research? One is that dream recall is partially determined by the salience of the dream experience itself. Salience is defined as the subjective impact (e.g., emotionality or vividness) of the experience of the dream (Cohen and MacNeilage, 1974). The assumption is that the more salient the experience, the more readily it will be recalled. Since the actual dream experience can not be directly assessed, evidence for the salience hypothesis is somewhat indirect. First, the difference in quantity and quality of dream reporting from REM and NREM awakenings is consistent with the hypothesis. Assuming at least a gross correspondence between subjective experience and physiological activation, it is not surprising that REM experiences are more vivid, exciting, and salient; REM sleep physiology includes (with some notable exceptions such as lowered tonus in certain muscle groups) marked heightening of cortical (EEG) activation and autonomic variability. However, evidence of correlations between physiological activation and dream recall predicted by the salience hypothesis is somewhat equivocal. Some studies have found greater recall for awakenings made after indications of phasic activity (e.g., eye movements and respiratory irregularity) than after periods of phasic quiescence (e.g., Goodenough et al., 1974), while others have failed to obtain commensurate relationships (e.g., Molinari and Foulkes, 1969). However, it would be premature to conclude that these data constitute definitive evidence either for or against the salience hypothesis
318
DAVID
B.
COHEN
because evidence that phasic activation is associated with more salient dream experiences is equally equivocal (see Cohen, 1974c, p. 142). The effect of presleep stress on the dream recall of infrequent recallers (Cohen, 1974a,b; Cohen and Cox, 1975) is consistent with the salience hypothesis. There is evidence for the assumption that infrequent recallers tend to have less intense imagery than do frequent recallers during wakefulness (Hiscock and Cohen, 1973) and during sleep (Cohen and MacNeilage, 1974). Then it is reasonable to hypothesize that pre sleep stress more readily enhances the salience and thus the recallability of the dream experiences of infrequent recallers. A more direct experimental test of this hypothesis was made from data obtained in a different study (Cohen and Cox, 1975). Dream content was assessed for a subset of most-frequent and least-frequent dream-recallers who were exposed to a positive or negative presleepcondition manipulation in the laboratory. On the basis of results from two previous studies (Cohen, 1974a,b), it was predicted that infrequent recallers would have higher salience scores under the negative than under the positive condition, while the presleep manipulations would have relatively little effect on the salience scores of the frequent recallers. This prediction was borne out. In the positive condition (similar to the experimental conditions of the Cohen and MacNeilage study which yielded a significant difference in salience between two comparable groups) a significant difference was again obtained. However, under the negative condition, this difference washed out. The effect was due to higher salience scores for the infrequent-recaller group in the negative compared to the positive condition. These findings, together with evidence that frequent recallers are not better than infrequent recallers in short- or long-term memory (Cohen, 1971; Taub, 1970) but do appear to generate more recallable dreams (Barber, 1969), suggest that the subjective impact of the dream experience partially determines the probability of dream recall. Since the subjective impact of the dream experience is known to be associated with stage of sleep, and since it is well established that recall after REM awakening is consistently higher than NREM recall, differences between recallers and nonrecallers might be a function of the habit of awakening from REM versus NREM sleep. One reasonable hypothesis is that infrequent recallers tend to awaken from NREMsleep, while frequent recallers tend to awaken from REM sleep. Data from the study by Cohen and MacNeilage (1974) tend to support this
DREAMING
319
hypothesis. Each subject slept four nights in the laboratory; the second night was a base-line condition of uninterrupted sleep. Although Night 2 data are available for only seven frequent and seven infrequent recallers, a trend was observed: at the end of the night, five of the seven frequent recallers awakened from REM sleep, while five of the seven infrequent recallers awakened from NREM (p = .13). These findings are consistent with the argument that dream recall on any night is a function of the probability of awakening from REM sleep multiplied by the probability of REM dream recall (Webb & Kersey, 1967). How would these presumed differences in sleep-stage awakenings between frequent recallers be explained? Perhaps infrequent recallers have a relatively lower awakening threshold for NREM than for REM sleep. There is simply no direct evidence to support such an assumption. Perhaps there are no major differences in REM versus NREM awakening thresholds of infrequent recallers; there might be a difference in the probability of being in REM versus NREM sleep at the moment that sleep is terminated either by an external (alarm) or internal ("biological") clock. Perhaps infrequent recallers are in the habit of terminating REM sleep prior to awakening. Data reported by Fiss (1969) suggest a possible mechanism and explanation. Under certain conditions, repeated interruption of REM sleep over a number of nights caused a shortening of REM periods even during the latter half of those recovery nights when REM sleep usually accounts for a significant percentage of total sleep time. Fiss suggested that there is a need to complete REM dreams, a need that is relatively independent of a need for REM sleep per se. Are infrequent recallers more likely than frequent recallers to require REM dream completion, and therefore more likely to avoid being in REM sleep at the end of the night? This question suggests two testable predictions. First, there is greater qualitative difference (salience?) in the REM versus the NREM dreams of infrequent than of frequent recallers. There is currently no experimental evidence for or against this prediction. Second, there is a greater difference between the REM (more) and the NREM (less) dream recall of infrequent than of frequent dream-recallers. Some preliminary evidence supporting such a prediction has been reported (Cohen and MacNeilage, 1974). What is suggested by these speculative remarks is that compared to frequent recallers infrequent recallers (1) show a greater discrepancy between REM and NREM dream salience
320
DAVID
B.
COHEN
and (2) show a correspondingly greater deficit of dream recall after NREM than after REM awakening. Empirical support for these hypotheses would be consistent with the underlying speculative hypothesis that REM dreaming is functionally more "important" to the psychological regulation of infrequent than of frequent recallers. There are many other unanswered questions regarding the salience factor in dream recall. For example, it is known that the recall of information presented to sleeping subjects is a function of the frequency of EEG (cortical activation) (Lehmann and Koukkou, 1973, 1974). One would predict that there is a positive relationship among measures of habitual frequency of recall, dream salience, and cortical activation. Another possibility is that differences in habitual and situational dream-recall depend to some degree on the ration of cortical activation in the left versus the right hemisphere. Does the recall of having been dreaming but the inability to report details (contentless report) depend on the ratio of right (imagery) to left (verbal) activation? These kinds of hypotheses about the psychophysiology of the dreaming experience and dream recall have not been systematically tested. However, there are data suggesting a relationship between cerebral hemisphere, dream content, and dream recall. Galin (1974) notes the parallel between descriptions of "primary-process" modes of thought and information-processing characteristics of the right hemisphere. For example: there is considerable evidence that the right hemisphere does possess words, but the words are not organized for use in propositions ... Because it deals more effectively with complex patterns taken as a whole than with individual parts taken serially, we might expect metaphors, puns, double-entendre and rebus, i.e., word-pictures. The elements in these verbal constructions do not have fixed single definitions (are not clearly bounded) but depend on context, and can shift in meaning when seen as parts of a new pattern. This is the sort of language that appears in dreams. (p. 574)
While it would be clearly misleading to downplay the role of the left hemisphere in the construction of many dreams, the following excerpt of a REM dream nicely illustrates the kind of properties Galin describes for the right hemisphere: I remember being in prison, and there was one particular girl ... every day she would make up a new game ... it would be names, you know, of different sounds, and it was all so odd ... I can't remember exactly what
DREAMING
321
she did. She kept making these songs. She put them all together ... and they spelled Coca Cola. Isn't that weird?
likewIse Galin reviews evidence that dream recall may be influenced by dream-salience characteristics mediated by right-hemisphere activity. Patients with right posterior lesions or injuries have been reported to experience a diminution or cessation of dream recall. In addition, there is evidence that convergent thinkers (who perform better on test of rational thinking than on tests of fluency and imaginativeness and who are presumably more left-hemisphere dominant) are poorer REM dream recallers than are divergent thinkers (Austin, 1971). Finally, it would be interesting to test the right-hemispheredream-salience-dream-recall hypothesis by studying the dream recall of alcoholics. There is evidence that chronic alcoholism eventuates in a relatively greater deterioration of right- than left-hemisphere mediated functions (Jones and Parsons, 1975). Were it also established that frequent recallers who subsequently became alcoholics also experienced a diminution of REM dream recall, a more direct link between sleep physiology and dream recall could be established. A third process hypothesized to account for dream recall is interference. Interference refers to events that distract attention from the dream material. For example, Cohen and Wolfe (1973) manipulated postsleep interference by having half their subjects dial the weatherinformation number and record the expected temperature immediately after awakening and just prior to recording any dream material they could remember. The subjects of the control group were asked to lie quietly in bed for the roughly 1.5 minutes required by the experimental subjects to carry out the distracting task. The results were clear-cut. Of the 40 control group subjects, 63% reported some dream content. The corresponding figure for the 46 experimental subjects was 33%. This finding was replicated within the context of a second study. In addition, it was found that the experimental condition markedly increased the percentage of con tentless reports, that is, reports of having been dreaming of something without being able to describe details. Con tentless reporting has been hypothesized to be a phenomenological representation of the repression process (Witkin, 1969, p. 32). However, marked elevation of contentless reporting after an un threatening, psychodynamically neutral task suggests the more parsimonious conclusion that the phenomenon is the result of interference.
322
DAVID
B.
COHEN
On a more speculative level, Cohen (1974c) has suggested the possibility that under certain conditions the transition from the sleep to waking state may interfere with consolidation, or at least immediate retrieval, of the dream experience in a manner consistent with a statedependent learning phenomenon. That is, for some subjects under certain conditions, information available in one stage (e.g., NREM) may be incompatible with retrieval mechanisms activated in a different state. (For a discussion of the relationship between performance, sleep, and state-dependent phenomena, see Overton, 1973.) The idea that often the immediately postsleep waking state may be psychobiologically "incompatible" with the prior dreaming state and thus interfere with retrieval of the dream is supported by evidence that under certain conditions individuals may be "reminded" of a dream later in the day. This could occur if one or both of the following events took place: perception of dream-related cues and! or a shift in the waking state toward greater congruence with the sleeping state during which the dream occurred. The latter factor is not an unreasonable hypothesis since it is known that during wakefulness as well as during sleep there are periodic increases and decreases in psychobiological activation making up the so-called basic restactivity cycle (BRAC), a roughly 90-minute fluctuation superimposed on the circadian rhythm (Kleitman, 1969). Perhaps, then, there are periods during the day that are psychobiologically "closer" to the dreaming state and that, in the presence of dream-related cues, make it more likely that a forgotton dream will be retrieved. Psychobiological "distance" between states could be operationally defined in any number of ways. On the assumption that dream recall might be employed as a rough but potentially useful estimate, some preliminary data suggest a potential relationship between sleep duration, REM dream recall, and notions about efficient or good sleep. In a recent (unpublished) study on problem solving, sleep, and dreaming, we were able to divide our sample of college men into short sleepers (who typically get less than 7 hours) and long sleepers (more than 7.5 hours). Each subject was awakened 5 minutes after the onset of each REM period of a single night to obtain dream reports. Dream recall was scored in terms of subjective reports on both (1) the percentage of the total dream recalled (on a three-point scale) and (2) the clarity of recall (three-point scale). Figure 2 shows REM dream recall for short sleepers (N = 7) and long sleepers (N = 7) in time since lights out.
323
DREAMING
FIGURE 2. Quality (amoul1t plus charity) of REM dream recall for short sleepers (N = 7) at various points during the 7.S-hour night in the sleep laboratory.
::l
6
~
4
~
2
~5/ w
Short Sleepers Long Sleepers
:s 3 0:
3-4 5 7 6 HOURS AFTER LIGHTS OUT
Note that with the exception of the last hour, REM dream recall of short sleepers is lower that than of long sleepers. These data were obtained on small samples of subjects sleeping only one night in the laboratory. However, they do support some interesting speculations. Do short sleepers get "better" sleep, sleep that is "deeper" and more restorative during the early part of the night, sleep that is psychobiologically more "distant" from wakefulness until the last hour when these subjects are normally ready to wake up? Do long sleepers require more sleep because their sleep is psychobiologically "closer" to wakefulness throughout the night? Evidence that certain conditions like presleep exercise reduce both REM dream recall throughout the night and spontaneous duration of sleep in long sleepers would reinforce the hypothesis that dream recall is a useful indication of "efficient sleep." In addition, it might be possible to discriminate between long sleepers who have "deep" (low-recall) sleep-that is, long sleepers for whom the last few hours of sleep provide no extra restorative benefits-and long sleepers who require long sleep because their sleep is less efficient. The former group would presumably find it easier to reduce sleep time gradually over a period of weeks or months without adverse psychological effects, while the latter group should find this change more disruptive. The soundness of these speculations could be tested by the establishment of individual dream-recall curves for each of a number of long sleepers and the prediction of (1) which subjects could most easily reduce their sleep time and (2) at what point in time the adverse effects of further reduction of sleep time would manifest themselves (e.g., in terms of fatigue, dysphoric mood, and loss of intellectual efficiency). The data briefly reviewed above lend themselves to the conclusion that the principles of salience and interference can account for most of
324
DAVID
B.
COHEN
what is known about the quantitative aspects of dream recall (Cohen, 1974c; Goodenough et al., 1974).1 What follows now is a discussion of the possible implications of dream recall for theoretical statements about the adaptive properties of dreaming.
C. Implication for Theory Dream recall is the process through which we come to know of, and make inferences about, dreaming. There are at least four characteristics of dream content that are presumably affected by the recall process (through omission, substitution, and elaboration) which influence theorizing: what we dream about, how we dream about it (direct versus symbolic transformation, condensation, etc.), sequential organization, and the adequacy of the ego functioning of the dreamer. The first two characteristics will be discussed in some detail in the second section of the chapter. Some preliminary comments about the latter two will be useful at this point to introduce the reader to a fundamental question-the adaptive capacity of the dream-that will be discussed in more detail in the third section of the chapter. The content of dreams may be analyzed for the presence of those cognitive functions presumably necessary for adaptation. These include evaluation and discrimination (judgment), perception (reality testing), capacity to delay and to inhibit impulse, concentration, curiosity, and competence. (These properties do not have to be characteristic of all dreams.) Some investigators believe that many or all of these properties are typically absent during dreaming (Hartmann, 1973; Klinger, 1971), but it should be noted that these kinds of properties are generally more difficult to assess (and perhaps, in the absence of explicit request from the investigator, harder to recall) (Cohen, 1974c). 1
However, salience and interference are theoretical conveniences of heuristic value, not clearly defined processes with empirically established a electrophysiological markers. The discussion of interference has focused on events occurring while awakening and during distraction. However, with respect to events occurring within the dream process, the distinction between salience and interference is clearly more difficult to maintain. Despite working assumptions, it is not known if the dream-generation process is truly independent of, and antecedent to, the attentional and memory factors that make dream recall possible. Can the psychological processes associated with salience be conceptualized as different from those psychological processes that determine attention or interference? Does the dream really "attract" attention, or does attention "create" the dream? Are there electrophysiological measures (EEG, EKG, EMG, REM, GSR) that could be used to denote salience versus interference? In fact, we have no answers to these questions.
DREAMING
325
Typically, dreams are relatively poorly recalled and often inadequately communicated. Therefore the typical dream report gives the impression that dreaming is often sequentially disorganized. For example, a subject might describe a scene in which he and a group of people are congregated in a room when" all of a sudden" he is driving his car through a blinding downpour. Next he finds himself alone in a room looking at some photographs. At first glance there seems to be no "logic" to the sequences of an apparently chaotic process that is alien to wakeful thought. However, a closer analysis of sleep and waking mentation suggests two things. First, it suggests that there is a great deal of "discontinuity" in the sequencing of waking thought, both in response to the shifts in conversational themes and in private fantasy stimulated at different points in any dialogue between two people. Thus abrupt sequential shifts in dream imagery may differ from those during wakefulness more in terms of the degree of visual dramatization, fluidness, and salience rather than in response to qualitatively different processes peculiar to sleep. In other words, the apparent sequential disarray of dream imagery cannot be invoked as evidence that the normal ego functions presumably necessary for meaningful, even logical, thought are in abeyance. Second, it is quite likely that much of the "discontinuity" of dream imagery is due to both faulty recall and faulty communication. Transitional points in the dream may be harder to recall, and subjects (especially early in the night) are prone to communicate parts of a dream in a sequence that is more a function of the recallability rather than the actual sequencing of the parts. In short, the potential adaptive capacity of the dream cannot be ruled out on the basis of apparent thematic disorganization. A more direct assessment of ego functioning during dreaming suggests that the requisites for problem solving are often retained. We are currently collecting REM dream reports from subjects in a study of the effect of sleep and dreaming on performance on a conceptformation problem. After each dream report is obtained, the subject estimates both the amount of the total dream he thinks he has recalled and the adequacy of his ego functioning during the dream on seven dimensions (e.g., concentration, judgment, and impulse control). Preliminary results have yielded the following. First, as time since initial lights out increases, estimates of both amount of recall and adequacy of ego functioning increase. Second, for any point in time there is a correlation between estimates of both amount of dream recall
326
DAVID
B.
COHEN
and adequacy of ego functioning. Third, after 5 or more hours of sleep subjects' estimations of their ego functioning are as high or higher than their estimations for 5 minutes of wakeful activity prior to arriving at the laboratory the previous evening. What do these preliminary data suggest? First, despite the fact that reality testing (ability to discriminate the "real" from the "fantasy") is generally in abeyance, ego functioning is often adequate or excellent during dreaming, especially during the last two or three REM periods of the night. Second, the impression that ego functioning is poor during dreaming may be to some extent an artifact of the dreamrecall process and other sources of bias that come into play with respect to dream reports. More specifically, the impression that dreaming is largely disorganized, chaotic, regressive, and alien may largely be based on (1) fragmentary recall and the inference made by the subject that these fragments represent the actual nature of the dream process; (2) poor communication and the inference made by the experimenter that this represents the true nature of the dream;2 (3) poorly recalled REM and especially NREM dreams and the inference made by the layman that these fragmentary memories are accurate representations; and (4) biased sampling (of dreams of clinical populations) and the inference made by clinicians that such reports are representative of those produced by members of the nonclinical population. Some dreams involve logical, even creative modes of thought. Some appear to be simple replays of rather pedestrian events. And some are a potpourri of exotic littles. A closer look at well-recalled dreams, especially those from later REM periods, suggests that ego functions are largely intact. Thus, from a purely observational point of view, the evidence suggests that the necessary, though not sufficient "raw material" for meaningful thought and action, even problem solving, exists during sleep. More direct, experimental strategies to assess this capacity will be discussed in the third section of the chapter. In the following section of the chapter, we will consider two general hypotheses about the nature of dream experiences: (1) that they reflect the combined effects of disposition and situation and (2) 2
It is interesting to note that when subjects elaborate on reports given earlier in the night, their descriptions indicate that the dream actually had a more logical sequence, and that ego functions were more adequate, than the earlier report suggested.
DREAMING
327
that they represent fundamental psychological processess which (perhaps aside from the specific physiological context in which they occur) may have a significant impact on the individual regardless of whether they are recalled or not.
II.
REPRESENTATIONAL PROPERTIES OF DREAMING
A. Validity of Dream Reports Since dream reports comprise much of the data base for a psychology of dreaming, two methodological issues relevant to the validity of such reports will be noted briefly. First, both the interpretation and the analysis of dreams and inferences about the function(s) of dreaming rest largely on the report of the memory of dreaming. The report is thus a rather indirect source of information. 3 The problem of the validity of the dream report has been discussed by a number of investigators (Rechtschaffen, 1967; Stoyva and Kamiya, 1968). However, most have given the problem little attention because it is assumed that the report is a fairly accurate representation of actual experiences. One strategy to assess the validity of the dream report is to observe the degree of convergence of objective "behavioral" indices (Stoyva and Kamiya, 1968). If, for example, the length of a REM period, irregularity of concomitant ANS events, and content of sleep-talking episodes correspond to the estimated length, subjective impact (salience), and content of the reported experience, one is reasonably confident that the report is a fairly adequate datum. For purposes of discussion, we will assume, with others (Bertini, !973; Hartmann, 1973), that there is a systematic if not isomorphic correspondence between the experience of dreaming and the psychological and neurological levels of dreaming. Further, we will assume that the dream experience is fairly adequately represented by the reasonably detailed dream report, though such an assumption should by now leave the reader with some misgivings. Second, in the discussion that follows, we will not be concerned a
One notable attempt to get at the dream experience more directly by the training of subjects to report while dreaming (Bertini and Pontalti, 1971) could not be replicated (Hauri, 1972).
328
DAVID
B.
COHEN
with the possible distinction between reports obtained in the laboratory and at home. There is some evidence that the laboratory setting tends to inhibit certain sleep-related phenomena, e.g., insomnia, sleep talking, nightmares, and flagrantly erotic dreams (including nocturnal emission). However, more careful attention to subject selection, reporting bias, and the use of within-subject comparisons of home and laboratory reports over a sufficiently representative period of time would provide a more useful indication of how" artificial" the laboratory setting actually is. Some investigators argue that laboratory dreams are more limited, drab, and pedestrian, while home dreams are more rich and revealing (Domhoff, 1969).4 However, with regard to certain dimensions of content (e.g., hedonic tone, aggression, and sexuality), evidence that home and laboratory reports differ markedly is equivocal (e.g., Hall and Van de Castle, 1966; Weisz and Foulkes, 1970). The distinction between home and laboratory reports may be more important for some questions than for others. Clearly investigation of the physiological correlates of dreaming requires a laboratory setting. However, a reasonable expectation is that empirical relationships supporting the validity of hypotheses about dreaming obtained via home reports should be validated in the sleep laboratory.
B. Two Strategies for Investigating Dreaming There are basically two major strategies used in the investigation of the representational properties of dreaming, one primarily used by clinicians, the other primarily used by academic psychologists. Interpretation of dream content is defined here as primarily a social process by which symbolic meaning is derived from, and often considered more important than, the manifest content of the report. The validity of the interpretive strategy derives from cultural tradition, clinical observation, common sense, and occasionally, empirical testing. Analysis of dream content, on the other hand, is defined here as essentially an empirical strategy by which objective elements (categories of dream content) are correlated with objective criteria. The meaning of a dream therefore derives from the reliable patterns of 4
Demonstrating a "pedestrian bias" in laboratory reports would not necessarily counterindicate their use within the context of experimental research paradigms.
DREAMING
329
associations between dream content and events (e.g., dispositional characteristics, social events, and physiological activity) that are external to it. The validity of such analyses derives from significant and replicable (predictable) relationships. Typically interpretation is oriented toward the so-called latent content (unspoken, inferable theme), while analyses is oriented toward the manifest content. The basic advantage of the interpretive approach is its recognition of the symbolic richness of human experience, while the desirability of the analytic approach rests on its objectivity (Van de Castle, 1969). A major disadvantage of the interpretive approach is that it is often refractory to verification (Van de Castle, 1971, p. 18). A major disadvantage of the analytic approach is its all-too-frequent superficiality. Recent work by clinical and academic psychologists is beginning to make less tenable an absolute distinction between the interpretive and the analytic approaches. Erikson (1954) has argued that dream interpretation should be based on both manifest and latent levels. Hall (1966) offers a method for inferring latent content from reliable (recurrent) themes expressed in manifest content. Others have begun to explore the dream process at both a manifest and a latent content level within the framework of experimental strategies (Breger, Hunter, and Lane, 1971; Cartwright et al., 1969; Fiss, 1969; Witkin, 1969). However, the major strategy of most empirical work on dreams is the assessment of relationships between objectifiable categories of manifest content and external events within the framework of correlational and experimental designs. An example of the capacity of empirical dream-content analysis to reveal clues about "deeper meaning" is illustrated by Van de Castle's analysis of the animal-content category (Van de Castle, 1969). Content analysis of dreams obtained in three different studies showed that younger children, peoples of nonindustrialized cultures, and women who had a relatively late onset of menarche have a higher percentage of dreams with animal content. Van de Castle speculated that animal content reflects "behavioral or emotional immaturity." This inference was further tested in a study by Cohen and Cox (1975). It was predicted that a larger proportion of dreams with animal content would be reported by a high- than by low-neuroticism subjects. A trend consistent with prediction was obtained but it was not statistically significant. The point, however, is that empirical strategies of
330
DAVID
B.
COHEN
content analysis are capable of generating hypotheses of theoretical interest; they do not necessarily restrict the investigator to superficial, theoretically banal correlations.
1. Clinical Approaches The bulk of neo- and post-Freudian theorizing about the nature of dreams by clinical investigators emphasizes the "meaning" of dream content rather than the function of dreaming. This is not to say that the question of function is ignored. That dreaming reflects defensive (e.g., resistance) and adaptive (e.g., information processing, creativity) aspects of personality is usually assumed, but the accent is on what the dream tells us about the individual. Assumptions about the function of the dream seem to emerge more readily in the form of interpretive emphasis and the kinds of examples that are selected to illustrate theoretical points. Thus the dream is treated like an existential statement through which a better appreciation of the person is derived for both intellectual and therapeutic reasons. The dream, then, is considered to be an alternative means of expressing motivation, defensiveness, and interpersonal preoccupations characteristic of the waking individual. In fact, some theorists (e.g., Fromm, 1951) tend to emphasize the intellectual flexibility, creativity, and source of insight inherent in some dreams. Fromm (1951, pp. 37-38) gives an example of a dream which provided the dreamer with more incisive perception of "a very important person" than was available to him during the day. It is the kind of insight which leads Fromm to explain why dreaming often seems to predict future events more accurately than does thinking. Bonime (1962) gives an example of the capacity of the dream to represent in pithy metaphor an important but vaguely understood preoccupation. A patient dreams of the analyst as a "kindly family doctor" sitting by her bed and observing her through a set of eyeglasses whose lenses have Y-shaped slits in them. To Bonime this dream represents ambivalent feelings of trust and suspicion. The glasses, which look like New York City subway tokens, suggest "token kindness" or superficial interest. Another example is the dream of a dog, half of whose face is eaten off by another dog. This dream is interpreted to represent the "dog-eatdog" evaluation by the patient of her relationship with her husband.
DREAMING
331
A third example is the dream of a female patient who continues to elude a photographer. This dream is interpreted as an expression of the motive to resist the probing therapist. The validity of such interpretations is said to rest on knowledge of the characteristic modes of personality functioning and the immediate personal and interpersonal problems of the dreamer. However, there are no scientifically established criteria to assess whether an interpretation is a valid indication of the dream's "real" meaning or an example of the therapist's inventiveness, or both. This oft-heard criticism is relevant to an understanding of the potential contribution of alternative, empirical approaches to dream content that are described below. In essence the criticism suggests that we do not have specific rules to translate between the knowing level (deep structure, central process, latent content) and the expressing level (surface structure, peripheral process, manifest content). Hypotheses about processes like condensation and displacement (Freud, 1953), readily verifiable from informal observation, provide too much latitude for subjective interpretations that are themselves not often verifiable in any scientifically acceptable sense. Eventually empirically derived correlations may yield rules that will enable us to predict that a specific event will appear in a dream. However, this would require more extensive knowledge about how events are understood (Bartlett, 1932) and how that understanding (which involves ideas more than icons, transformations more than reproductions) generates hallucinatory expression during different stages of sleep. This requirement seems all the more urgent if we, like most dream theorists, assume that the meaning level is more fundamental than the imagery or verbal level (Attneave, 1974). It therefore seems reasonable to anticipate that systematic relationships emerging from empirical strategies will eventually contribute to more objective and testable theories of the meaning, process, and function of dreams. Hall (1966) emphasizes the need for a series of dreams from which reliable inferences about personality can be made. In his research, he found that each personal dreamed about much the same sort of thing from year to year even when there were radical changes in his waking life. We attribute this consistency to the unchanging character of the unconcious from which dreams are believed to come . . .. Dreams appear to be variations on a few basic infantile wishes and fears that have not been fulfilled or resolved. (p. xvii)
332
DAVID
B.
COHEN
Hall's theorizing represents a methodological advance over other clinical approaches to the meaning of dream content in that he has applied empirical techniques which permit a more objective assessment of the reliable features of the material. This strategy yields more credible evidence of the properties of the dreams of individuals and of dreams in general, regardless of one's faith in the validity of his interpretation of those regularities. However, what Hall shares with his contempories and predecessors is a belief in the continuity of personality functioning during the dreaming process, that is, a belief that the interpretation of dreams reveals much about the waking individual. This assumption underlies much of the empirical research on the representational properites of dream content to be reviewed below. After presenting some of these findings, we will discuss their implications for theory of dream function.
2. Empirical Approaches Much of the research on dream content is predicated on the assumption that the dream reflects chronic or situational events in some more or less systematic, measureable, and predictable manner. The capacity of the dream to reflect presleep and sleep events external to it has been explored in a number of ways. (For an excellent discussion of important, unresolved methodological and theoretical problems in the study of dream content, see the symposium on dream content in Chase, 1972, also see Hall, 1969; Hall and Van de Castle, 1966; Van de Castle, 1969). One method is to study correlations between categories of dream content (e.g., emotionality) and concurrent indicators of sleep phYSiological activity (e.g., heart-rate variability). Since an excellent review of the results of psychophysiological correlates is available (Rechtschaffen, 1973), a detailed discussion of this area of research will not be pursued here. In general, evidence for consistent patterns has been somewhat equivocal. One obvious explanation is that the same physiological event may have markedly different psychological correlates across individuals. For example, I have some unreported results suggesting that under "positive" presleep conditions, the eye-movement activity of REM sleep (REM density) is greater for subjects scoring high on a measure of neuroticism (emotionality) than for subjects scoring low on this measure. Conversely, REM density was greater for the low- than for the highneuroticism subjects under "negative" (stressful) presleep conditions.
DREAMING
333
The comparison of preselected groups of subjects with respect to different patterns of physiological activity might increase the reliability of findings. However, person-by-situation interaction methodology has not been fully exploited in the psychophysiological research area. One other problem is that psychophysiological correlations usually can not answer the question of causality. If one's major interest is in factors that affect dream content rather than the effect of dream content on sleep physiology, strictly correlational data have a limited usefulness. Another approach is to explore the transformations in dreams of presleep events such as stressful films, hypnogogic presleep fantasies, or hypnotically induced material (Baekeland, 1971; Bertini, Lewis, and Witkin, 1972; Breger et al., 1971; Cartwright et al., 1969; Cohen and Cox, 1975; Foulkes and Rechtschaffen, 1964; Hauri, 1970; Tart and Dick, 1970). A number of questions can be explored by this method. For example, do dreams reflect personality differences? Under what conditions are pre sleep events openly or symbolically dreamed about, or not dreamed about at all? Can subjects follow presleep instructions to dream about a particular theme? Are these patterns related to presleep events and to significant changes in the individual? This latter question will be explored in more detail at the end of this chapter. An interesting question bearing on both theory and methodology is whether dream content reflects presleep events in a continuous or a compensatory manner (Foulkes, 1970). The continuity hypothesis is associated with Adler, while the compensatory hypothesis is associated with Jung. Evidence that certain presleep conditions (e.g., excessive exercise) tend to be dreamed about in a compensatory fashion (e.g., relaxation) (Hauri, 1970) suggests a sort of homeostatic regulatory function for the dream. However, such an example does not do justice to the sophisticated concept of compensation in dreams. (See Dallett, 1973, for a discussion of Jung's hypothesis.) Therefore evidence for compensation in dreaming is relatively sparse, mainly because the focus of most research has been on the continuity hypothesis and because evidence for compensation comes from studies that generally do not preselect subjects for specific traits (Dallett, 1973). In addition, it is possible that in some instances the notion of compensation (dreaming the inverse of a presleep event or the "opposite" of what would be expected on the basis of measurements
334
DAVID
B.
COHEN
of waking personality) may be confounded with the notion of symbolic transformation, and the problem may be compounded by problems in content-scoring strategies. For example, Cartwright et al. (1969) predicted that the effect of an erotic movie would be to decrease direct references and to increase symbolic allusions to that presleep stimulus. Results were consistent with prediction. A scoring system not sensitive to symbolic transformation might yield results suggesting a compensatory process that reduces the quantity of presleep preoccupations. Another example comes from a recent study by Cohen and Cox (1975). Under conditions of presleep stress, there was less evidence of direct incorporations of experiment-related events than observed under positive presleep conditions. Conversely, the stressful pre sleep condition generated more symbolic allusions to the experimental situation. These findings, more or less consistent with those reported by Cartwright et al. (1969), suggest that the question of complementarity versus continuity needs to be separated from the question of direct versus symbolic transformation through more careful attention to rationale for predictions, specification of presleep conditions, and clarification of content-scoring methodology. An illustration from Witkin's research (Witkin, 1969, p. 24) is instructive. After viewing a particularly distressing film, one subject "turned the tables" as follows. "After the quite bloody birth film, he dreamed of an idyllic park scene; after the subincision film, in which all the men are naked, he dreamed about characters who were elaborately dressed." Apparently compensatory transformations such as this might serve habitual personality dispositions. From a continuity point of view, one would not be surprised if waking measures of defensiveness revealed that this subject were prone to use denial. Thus it is quite likely that compensatory processes may serve the need for continuity of personality functioning in some subjects. Most of the reported research on the relationship between personality and dream content has focused on, and obtained support for, the continuity view. Waking measures of creativity, social dominance, sex-role orientation, and psychopathology have been shown to correlate with theoretically relevant dream-content measures (Adelson, 1959; Carrington, 1972; Cohen, 1973, 1974a; Cohen and Cox, 1975; Foulkes, 1967; Foulkes and Rechtschaffen, 1964; Kramer, Baldridge, Whitman, Ornstein and Smith, 1969). For example, Carrington (1972) found that the dreams of schizophrenics were "replete with mutilation
DREAMING
335
imagery and morbid themes. They were more aggressive, more bizarre, and more often reflective of ego dyscontrol" (p. 347). These results find some support in other studies (Kramer et al., 1969; Okuma, Sunami, Fukuma, Takeo, and Motoike, 1970). Likewise Adelson found that the dreams of college women rated as creative writers were more interesting and innovative than those of less impressive endowment, and these findings are supported by other observations (e.g., Schechter, Schmeidler, and Staal, 1965). A recent study by Starker (1974) demonstrates the apparent continuity between daydreaming and dreaming styles. Individuals with a "positive daydreaming style" (absorbant and pleasant daydreaming) tend to have less bizarre and more pleasant dreams than individuals with a "negative daydreaming style" (involving guilt, fear, and conflict). Cohen (1973) reported data on the correlation between masculinity-femininity and dream content that also support the continuity hypothesis. Two groups of male and female college students were divided into masculine and feminine orientation subgroups on the basis of each group's mean score on the femininity (Fe) scale of the California Personality Inventory. This scale does not measure effeminateness but rather an orientation toward activities and interests characteristic of the male or female stereotype in our culture. Therefore masculinity refers to hard-headed, outgoing ambitiousness, opportunism, robust activity, and impatience with delay. Femininity refers to helpfulness, gentleness, moderation, sincerity, and sympathy. Dreams were recorded at home over a 3-day period, and individuals producing two dreams were selected as subjects. Dreams were scored independently and reliably by two judges blind to the hypotheses for three major (and theoretically relevant) dimensions: "agency" and "communion" (as defined by Bakan, 1966, and empirically developed by Carlson, 1971) and aggression. Agency is meant to refer to masculinity and is defined in this study by surgency (self expansion, assertiveness, competition), instrumentality (mastery, problem solving), and libido (raw sexuality). Communion is meant to reflect femininity and was defined in terms of active social cooperation (altruism, helping), passive social connectiveness (reunion, receiving support), and eros (expressions of warmth, affection). Aggression was defined in terms of verbal and physical categories. In general there was a significant relationship between sex-role orientation and relevant dream dimension, especially for males. That is, masculinity predicted
336
DAVID
B.
COHEN
agency, femininity predicted communion, and there was a tendency for larger percentages of masculine males to have aggression, especially physical aggression, in their dreams. The capacity of content to reflect personality disposition-in this case, masculinity-femininity in males-is readily demonstrated by the following two rather obvious examples: I see G and L in Hermann Park in Houston. I want G to go the Winnipeg Royal Ballet with me that night. She already has tickets to go to some other club. I blow it off and am somewhat disappointed. I dreamt that I was some kind of merchant dealing with both sides in the American Civil War. I took my payments in sex with northern and southern women.
The difference between the first dream with its expression of aesthetic interest and emotional (communion) rather than self-assertive reaction to rebuff, and the second dream with its flamboyant, libidinal, exploitive quality (agency) is quite obvious. Another quality that seemed to differentiate the dreams of the feminine from the masculine male dreams was a sense of vulnerability expressed in terms of physical harm, psychological stress, and identity alterations. A more obvious example of this is the following dream of a feminine male subject:
J comes home with her parents in a convertible. She has dyed her hair red and cut it in a long shag. Her house is next door to mine. C is present but
turns into B. Apparently, Vs parents have gone to Colorado to get her. It is a very awkward situation. Her stepfather doesn't remember B or me. Mrs. S, J's mother, introduces us to Mr. S. B is wearing a bathing suit. He is turned on by J. He gets an erection, which is very visible. His penis falls out of his bathing suit and starts bleeding and deteriorating.
The italics are added to denote the identity diffusion, vulnerability, and helplessness expressed in the report. The temptation to interpret some of this material as reflecting" castration anxiety" is reinforced by examples of variants of such themes in the dreams of other feminine male subjects, e.g., "a dog with its head cut off" and "turning into a woman." The most parsimonious interpretation of these recurrent themes is that they reflect an unstable masculine identification and psychological difficulties. In this study the relationship between sex-role orientation and relevant dream-content category was stronger than the relationship between sex and category. That is, personality rather than sex was the stronger predictor for the categories of agency or communion. Of
DREAMING
337
course this is not to deny the fact that college men tend to dream of football games more often than do college females. What is different is the way in which the dream is constructed; the way that reality is experienced is somewhat predictable on the basis of waking personality. It is noteworthy that the obtained differences between sex-role orientation groups in agency and communion were significantly greater for males than for females. The findings strongly support the contention that socialization of sex-role orientation is a more difficult "problem" for boys than for girls (Bardwick, 1971; Seyfried and Hendrick, 1973). For example, in our culture it is generally more distressing for adults to witness the behavior of a sissy than that of a tomboy. Girls seem to have more latitude in their orientations, and they appear less clearly differentiated and less conflicted over masculine and feminine orientations within their personalities. The dreams of feminine males tended to be more unpleasant than those of masculine females, suggesting that sex-role orientation contrary to stereotype is more disturbing for males than for females. It should be pointed out that unless the dreams of subjects classified according to major personality dimensions are obtained over a long period of time, studies like the ones just reviewed can give only limited support for the continuity hypothesis. Over the short run there is a great deal of variability within subject groups, variability which might have been considered as "error variance" but which reflects meaningful personality-by-situation interactions that could not be assessed in the study. This leads us to research which considers the effect on dream content of both personality and presleep conditions. In a major ongoing investigation, Witkin, Goodenough, and their colleagues (Witkin, 1969) are exploring the effect of field dependence and presleep stress on dream content. Only preliminary results have been reported. But the discussion of procedure and initial observations indicates that a personality-by-condition experimental design can be useful to define factors that influence the process of dreamwork. The present author has carried out a number of studies along somewhat similar lines whose results are relevant to the continuity hypothesis. In the first (Cohen, 1974b) college women reported dreams at home over a period of a week during which they also recorded pre sleep mood. Results suggested that the positive or negative quality of dream affect was correlated with the quality of presleep mood. This
338
DAVID
B.
COHEN
seemed especially true for subjects who rated their presleep mood throughout the study as generally dysphoric, but limited sample size made such a conclusion tentative. In the second study (Cohen, 1974a) another set of college women provided additional data. This time they were preclassified into "sensitizers" and "repressors" on the Byrne scale (1964). It was predicted that only for the sensitizers would there be a correlation between pre sleep mood and dream affect. This prediction was clearly supported by the data. A more elaborate followup experiment has confirmed these findings (Cohen and Cox, 1975). The results of these three studies are summarized by the dashed lines in Figure l(b). Before we describe that experiment, an important theoretical point should be made. For the personality dimension of "emotionality" (estimated by sensitization, anxiety, or neuroticism tests), we expect that affect should be more strongly and lastingly affected by conditions for high-emotionality than for low-emotionality individuals. When subjects are used who characterize the extremes of this emotionality dimension, contrasting predictions with respect to affect can be made. The results of the study should make this point clear. Cohen and Cox (1975) preselected male college students on the basis of high or low Maudsley Personality Inventory Neuroticism (emotionality) scores. The subjects of each group were randomly assigned to one of two laboratory presleep conditions. The positive condition included treating the subject in a friendly and personal fashion, giving him a great deal of information about procedures and techniques of sleep research, spending a lot of time with the subject, and presenting him with easy items from an IQ test. The negative condition included treating the subject in a perfunctory and impersonal fashion, giving him no explanation about procedures, isolating him for at least 15 minutes after the recording electrodes had been placed, and exposing him to difficult items on the IQ test. Postsleep questionnaire data clearly indicated that subjects in the two conditions had perceived the major components of the manipulation. For the high-neuroticism group, there were positive and significant intercorrelations among condition, pre sleep affect (subjective ratings made prior to sleep), and affect associated with dream content retrieved from REM and NREM awakenings throughout the single night that subjects slept in the laboratory. None of these correlations was significant for the low-neuroticism group. The affect of presleep
DREAMING
339
condition on affect valence could be tested at different points in time: in the evening (presleep mood), during the night (dream affect), and in the morning (postsleep affect). If emotionality is defined as labile and laj)ting affective reactivity (Eysenck and Rachmann, 1965), then the correlation between condition (positive or negative) and affect ratings of the high-neuroticism group are similar and statistically significant. It would appear that the dreams of the high-neuroticism group, and the affective reaction elicited by these dreams, continue to reflect the impact of the presleep conditions. The postsleep correlation is positive though not significant. For the low-neuroticism group none of the correlations is Significant. The shape of the curve suggests that what little relationship there was in the immediate presleep situation quickly dissipated to chance level (as reflected in the dream and postsleep affect correlations). (See Figure 3.) Complementing the differences in the distribution of these correlations within each group are differences in the correlations between dream affect and other dimensions of dream content. Thus for the high-neuroticism subjects there was a positive correlation between unpleasant dream affect and the following: bizarreness, excitement, personal significance, and noncontemporaneous setting. For the lowneuroticism subjects there was a negative correlation between dream unpleasantness and the following: bizarreness, excitement, and significance; there was virtually no correlation between dream unplesantness and temporal setting. The difference in direction and the magnitude of these correlations suggests that dream affect "means" something qualitatively different for the subjects of these two groups. For example, for the high-neuroticism subjects the more unpleasant dreams were rated as more personally signficant, while for the lowneuroticism (defensive?) subjects the more unpleasant dreams were rated as less personally significant. This difference (p < .05) is consistent with what one might have predicted on the basis of personality differences between these two groups (Byrne, 1964). That is, high-neuroticism individuals are more aware of and concerned about emotionally negative experiences, while low-neuroticism types are less aware of and less concerned about such experiences. The difference between the two groups in apperception of the experimental conditions may be illustrated by one further difference in dream content. Under the negative presleep condition, a larger percentage of high-neuroticism subjects (50%) than of low-neuroti-
340
DAVID
B.
COHEN
• High N o Low N .60
.~
I:
:eo
•
.50
'"C
I:
8
a. (I) (I)
.40
o
en
~
c..
.30
.I:
'i
I:
o
•
. 20
:;:; «I
~ ...o
()
-0
.10
Presleep
Dream
Postsleep
AFFECT 3. Degree of association (point-biserial correlation) between valence of presleep condition and valence of dream affect at three points in time for high- versus lowneuroticism (N) groups separately.
FIGURE
cism subjects (10%) reported that their dreams had taken place in the past (p = .05). How can such a difference be interpreted? One possibility is that certain subjects, when stressed (e.g., experience failure, separation, and rejection, are more susceptible to the reactivation of thought processes (represented by conscious associations, fantasies, and dreams) relevant to earlier attempts to adjust to stress. That is, they are more susceptible to a kind of "regression" by which current perceptions are altered by temporally more remote schemas. The reactivation of the latter may account for the tendency of highneuroticism subjects to report more dysphoric mood on the presleep
DREAMING
341
questionnaire, and more dysphoric dream affect, under conditions of stress. It is as though they were reacting to the stress in a somewhat more childish manner than were the low-neoroticism subjects. This is an admittedly speculative hypothesis that requires further testing, but it does account for the tendency for high-neuroticism subjects to appear more "immature" (Byrne, 1964). The hypothesis might account for the intensity of emotion that seems more appropriate in children who are cognitively and interpersonally less competent than adults and who are thus more susceptible to perceptions of threat and helplessness. Thus while the two groups were indistinguishable in terms of their perception of the components of the pre sleep manipulation, they were clearly different with respect to their experience (personal meaning) of these conditions. These kinds of data illustrate how dream reports can be used to test hypotheses derived from personality theory and, conversely, how personality theory can be utilized to reveal something about dream content. In order to provide a more graphic illustration of some of the unsolved, but empirically testable, questions about representational and adaptive properties of dreaming and their interrelationship, some examples of the REM dreams of negative condition subjects in the Cohen and Cox (1975) study will be given. Recall that these subjects were made to experience difficulty or failure on an ostensibly easy IQ test by the selection of difficult items. One high-neuroticism subject dreamed something about checking circuits: "turning them back off, cycling circuits through a couple of times." He also had the vague impression of another "presence" (the experimenter?) in the room and believed that his activity had something to do with "psychological reactions or something." He went on to describe the task in terms of "light patterns rather than EEG." The dream suggests the importance of the presleep condition, the challenging testing atmosphere, the potentially threatening presence of the experimenter, and a preoccupation with the electrical recording procedures. Concern about the performance adequacy and imminent failure is more clearly expressed in the following REM dream of another high-neuroticism subject: I was having some sort of psychology test on me for the perception of chemicals, and it was a strange situation ... there was a battered-up truck set out ... I was supposed to perceive some differences. The chemical he was testing was tetramethyl chloride. I don't even know if that exists ... I came up wrong. I couldn't tell the difference.
342
DAVID
B.
COHEN
Compare these examples to those reported by low-neuroticism subjects. The following is the gist of one. The dreamer is seated in a football stadium and has in his possession a strength-giving potion. It seems natural for him to have it, and he is eager to take it. The main reason for taking it is for experimentation purposes and the desire to have extraordinary powers, including a possible increase in intelligence. It is clear that the subject is dealing with the presleep challenge to his self-confidence. The "solution" seems to be that he is in control, that he has resources to improve his situation, and that the resolution of the problem is assured. But is this really the case? When it is said that an individual is "dealing with" a problem, does it mean simply that he is expressing it, or does it mean that he is better off for having done something about it, or both? To answer such questions would require an extensive analysis of the relationship between estimates of self-esteem, measures of certain apsects of dream content, and psychological change scores. For example, suppose that subjects of similar levels of self-confidence and similar reactions to presleep task difficulty (e.g., mood ratings) were divided into those who dreamed about the "problem" directly, symbolically, or not at all. A reliable pattern of relationships between problem representation in the dream and psychological change might be useful to validate the notion that certain kinds of dreaming (e.g., symbolic representation?) confer an adaptive advantage over the other kinds of dreaming. Were a sufficient mount of dream-content material available, one could also explore the relationship between the progression of events across dreams and the degree of psychological change. Now consider the following dream of another low neuroticism subject: OK, there was a baseball game, and you (the experimenter) were in it, and you were wearing those shoes you are wearing now. I think you were the pitcher ... it was the World Series or something. It was the big game and we were in the dugout talking.
So far, the dream is more directly about the experimenter, and both he and the subject are together during an important event. The experimenter is portrayed as the person in control (pitcher). Later the subject says that the pitcher is upset, that things aren't going so well, but that it doesn't seem to bother the subject; in fact he is happy just to be in the game. It appears then that the subject recognizes the fact that the game (testing) isn't going well, but he feels all right. So far, this dream
DREAMING
343
is consistent with results from a comparison of the affect scores of lowneuroticism subjects across presleep (positive versus negative) conditions; that is, that there is a tendency for these subjects to report, and to be rated by judges as experiencing, little dysphoric affect under negative conditions relative to that experienced by high neuroticism subjects. (See Figure lb.) But what about the fact that it is the experimenter, not the dreamer, who is portrayed as having dysphoric affect? Does this reflect ego strength rather than defensiveness (Cohen, 1974a)? At this point, we can only guess, and the latter can not be ruled out. Here is another example of a dream from a low-neuroticism subject: Something to do with a baseball game ... real strange. The perspective is as if inside a baseball. All I got out of it was one scene of coming at, across the plate, as a guy is swinging his bat. It's like being among giant people or something ... I see the catcher (through the window in the bal!), and I feel the motion through the air ...
This is a delightful phenomenological metaphor of the experience of being a subject in a particular experiment. The subject feels small, confined, and definitely not in control of the situation. However, there is no sense of danger or feeling of anxiety. In fact, the subject indicates that his only reaction is interest and curiousity. Is this an example of creative symbolism or of defense (repression)? Again, we cannot yet answer this question because we do not have enough reliable information about the subject or about the relationship between this kind of dream versus other kinds and psychological change from presleep to postsleep. Until such experimental data are available, the most definitive conclusions that can be made will concern the capacity of the dream to reflect personality disposition (e.g., emotionality and cognitive style). The emphasis of this section of the chapter has been on the relationship between individual differences in personality and differences in dream content. On a more fundamental level, dreaming can be thought of as a processing of "current concerns" (Klinger, 1971, 1975). Research on sleep and human performance (Evans, Gustafson, O'Connell, Orne, and Shor, 1970; Koella and Levin, 1973, p. 36-67; Langford, Meddis, and Pearson, 1974) suggests that while learning and performance during electrophysiologically defined sleep is limited, the ability to discriminate meaningful from irrelevant (unimportant) stimulation is relatively unimpaired. In other words, the sleeping
344
DAVID
B.
COHEN
person continues to monitor, or at least to be potentially responsive to, changes in the environment that are significant. Significance can be manipulated experimentally. For example, subjects are capable of making simple motor movements specific to certain cues (e.g., a name) that are embedded in a series of "relevant" and "irrelevant" stimuli presented during sleep. In addition, and this is important to our discussion, "current concern" can be represented by the content of the dreaming process. For example, subjects newly exposed to the novel conditions of the sleep lab often dream directly or symbolically about the experimental situation (e.g., apparatus and experimenter). In short, there is good evidence, not only from clinical observation but also from experimental work, that dreaming is part of a more fundamental information-processing that is relatively continuous throughout the night. A discussion of the evolutionary significance of this process would take us far afield. However, it should be mentioned that there is some evidence that information processing during sleep, and perhaps even "dream" representation of this processing, is present in higher subhuman animals. Vaughn (1964) was able to demonstrate that rhesus monkeys could make motor movements during (REM?) sleep that had been previously learned during wakefulness as a means of avoiding shock associated with visual information. This finding is consistent with (though not unambiguous proof of) the hypothesis that higher animals are capable of some sort of dreaming experience. In order to test the hypotheses that higher animals dream and that dreaming is a representation of the operation of current concerns, the following experimental paradigm is suggested. A monkey raised in isolation could be trained to make a morphologically specific hand gesture in order to gain access to a sibling or a companion at the presentation of a signal. The companion would be an especially social object to an otherwise isolated monkey. At some point, an extinction period would be instituted. Presumably a sequence of disengagement would ensue (Klinger, 1975) during which agitation and "grief" (prior to a state of apathy) would be manifested. The prediction is that during sleep, especially REM sleep, the monkey should make the incentive-related gesture at a higher frequency than during (prefrustration) base line. Were there a correlation between separation and frequency (or intensity) of companion-related gestures, there would be more definitive support for the hypotheses that animals continue to
345
DREAMING
experience the articulation of current concerns during sleep and that these concerns are most likely to be represented in some sort of imagery. This kind of evidence would support the idea that dreaming is a reflection of a reconstruction or recreation process adaptive to problems of a purely intellectual or an interpersonal nature. It would also serve to link theories of dreaming in particular and theories of fantasy in general to theories of the development (through phylogeny as well as ontogeny) of intelligence, play, and socialization (Klinger, 1971). That representational and functional properties of dreaming are not unequivocally distinguishable on the basis of most empirical studies of dream content is aptly expressed by Meier. Referring to data on the relationship between dream content and the menstrual cycle, data which have been construed as evidence for adaptive function of dreaming, Meier says: All we can really say is that these subjects have such and such dreams ,-,{hile they are menstruating. We cannot say that these dreams serve the purpose of adapting them to the physiological condition. All we can say is that the two occur simultaneously. I think we are stipulating causal connections much too soon. I mean that we would then have to say that one can only develop psychologically because we have such and such dreams. All we really known is that we dream all the time, and that we are also developing. (In Chase, 1972, p. 260)
Hauri goes further, speculating that were a person under stress somehow made to dream more directly about the stress, that person might be better off (in Chase, 1972, p. 261). Investigating the causal relationship between what a person dreams about and how that person changes psychologically should further our understanding of how representational and adaptive properties of dreaming are separate but interrelated.
III.
FUNCTIONAL PROPERTIES OF DREAMING
The present section will describe strategies for investigating dream function. Special emphasis will be given to the following question: How can dream content be used to make inferences about adaptive psychological processes that occur during sleep? More specifically, can it be demonstrated that the individual is in a better state of mind" in the morning than he was presleep because of the way he II
346
DAVID
B.
COHEN
dreamed? Anecdotal evidence (Krippner and Hughes, 1970) and the weight of psychological theory favor such a demonstration. Before we review some examples of more "rigorous" approaches to the question of adaptive properties of dreaming, consider the following relatively informal demonstration described by Dement (1972). In the demonstration 500 graduates were given a problem to work on 15 minutes prior to sleep and were asked to record dream content the next morning. In one problem the subject was asked to figure out the word that represents the following sequence: H, 1, I, K, L, M, N, O. The solution is water or "H to 0." One student, who solved the problem incorrectly with the word alphabet reported several dreams that had water in them (ocean, swimming, raining). Keeping in mind the limitations of this kind of study, the findings suggest the following: (1) Problem representation is a relatively frequent characteristic of dreams. (2) Problem solution can occur but is extremely infrequent (and probably requires emotional and intellectual commitment and a reasonable level of intelligence). (3) Solutions may be rather indirect! symbolic (as in the famous snake dream of Kekule). (4) The person may not be aware of the relevance of dream elements to the problem or to the solution presented in the dream. That is, in some cases the dream provides a better solution than waking thought. As Dement (1972) says: perhaps only the most perceptive dreamers possess the ability to recognize a solution that is presented in disguised or symbolic fashion. Most of us, most of the time, are like the student who failed to recognize the word "water" as the solution to his problem even though he was deluged by water in his dreams! One can easily imagine Kekule shrugging as he awakened from the dream of the six circling snakes: "What nonsense! I must forget about snakes and concentrate on chemistry." (p. 101)
A. Functions of REM versus NREM Sleep Sleep includes both physiological and psychological processes, but for our purposes, hypotheses about psychological function are more relevant. Prior to the advent of modem sleep research, the function of dreaming was described within the context of the drivedefense model of Freud or in terms of the problem-solving models of neo- and post-Freudians. The dream was conceptualized as primarily
DREAMING
347
a defensive process by which infantile motives were transformed into pithy "cover stories" through a process of distortion (Freud, 1953). Or the dream was conceptualized as an alternative way of expressing disturbing personal defects, interpersonal problems, or intellectual preoccupations in a manner that facilitated self-deception (Ullman, 1962), insight (Bonime, 1962; Fromm, 1951), or creativity (Krippner and Hughes, 1970). What these theories have in common is an emphasis on the adaptive significance of the psychological processes underlying the dreaming experience. Dreaming is viewed as an important, if not necessary, function for the preservation of personality characteristics (Cartwright, 1974). Presumably, defective dreaming or the absence of the opportunity to dream would cause significant psychological disruption. Hypotheses arising out of current sleep research have much in common with these ideas despite their new data base. In addition, many of these new hypotheses have included assumptions about the differential function of REM versus NREM sleep. The distinction between REM and NREM sleep is in an empirical tradition of great heuristic value originally stemming from observations of differences in the following areas: (1) EEG characteristic (amplitude, frequency, etc.); (2) autonomic activity; and (3) dream recall (amount, frequency, and content). One of the more popular theories emphasizes the importance of NREM sleep for physiological restoration and of REM sleep for psychological adaptation (Hartmann, 1973). Specific hypotheses regarding the latter have attributed to REM sleep important functions such as (1) periodic reafferentation of the CNS (Ephron and Carrington, 1966); (2) endogenous stimulation of the oculomotor system during ontogeny (Roffwarg, Muzio, and Dement, 1966); (3) facilitating processes underlying binocular vision (Berger, 1969); and (4) information processing underlying memory, concept formation, defense, emotional problem solving, and creativity (Breger, 1967; Bryson and Schacher, 1969; Dewan, 1969; Krippner and Hughes, 1970; Pearlman, 1970; Shapiro, 1967). These hypotheses are not mutually exclusive and gain support from both clinical experience and laboratory study of sleep. "Problem-solving, information-processing, and ego consolidation theories shade into one another and often appear to be discussing similar or identical processes, despite the use of different languages" (Dallett, 1973, p. 414). These hypotheses have implica-
348
DAVID
B.
COHEN
tions for research methodology, and the most prevalent strategy has been the use of REM deprivation with NREM sleep interruption (or NREM-Stage 4 deprivation) controls. While a comprehensive review of REM deprivation work on animals is clearly beyond the scope of this chapter, some examples of research strategy and findings will focus attention on the central theme of dreaming as an adaptive psychological process. A number of studies (usually with mice or rats) have used REM deprivation prior to learning (to test hypotheses regarding acquisition) or after learning (to test hypotheses regarding memory consolidation). For example, Hartmann and Stem (1972) trained a group of REM-deprived and a group of stress-control rats in an active avoidance task. The REM-deprived rats required more trials to learn the task than the stress-control group or a REM-deprived group administered L-Dopa (a catecholaminergic precursor). The investigators concluded that REM sleep facilitates learning and that this process is catecholaminergically mediated. (See Hartmann, 1973, for a discussion of this hypothesis.) In another study Pearlman and Greenberg (1972) trained three groups of rats on an active avoidance-learning task. One group was then allowed 5 hours of REM-deprived sleep and then 5 hours of recovery sleep. One group was allowed 5 hours of normal sleep and then 5 hours of REMdeprived sleep. A third group was allowed 10 hours of normal sleep. The first group had the worst retention scores, which suggested to the investigators that REM sleep is implicated in the early phases of memory consolidation. Other studies of the effect of REM deprivation on latent learning and preextinction exposure to discriminative stimuli support their hypothesis. Pearlman has provided support for an interesting theoretical advance of the REM-learning hypothesis with evidence that "biologically prepared" (i.e., instinctively prepared, see Seligman, 1970) learning is less affected by immediate REM deprivation than is biologically unprepared learning (Pearlman and Becker, 1973). In a third variant of the REM-learning hypothesis, Fishbein and Kastaniotis (1973) demonstrated that after active avoidance learning REM time increased over that of yoked controls, while both groups showed a (stress-related) increase in NREM sleep time. While these psychobiological investigations are of great heuristic value, they are somewhat difficult to interpret in the light of more equivocal results arising from studies of human sleep. For example, most recent investigations have not substantiated Dement's now-"classic" find-
DREAMING
349
ings (1960) that REM deprivation disrupts personality. In addition, there is little evidence that REM deprivation disturbs schizophrenics and some evidence that it may be therapeutic for psychotic depressives (Vogel, 1968; Vogel and Traub, 1968; Vogel, Traub, Ben-Horin, and Meyers, 1968). Evidence for hypotheses about the importance of REM sleep for information processing have been both supportive (Empson and Clarke, 1970) and nonsupportive (Chemik, 1972; Johnson, 1973). On the basis of research on animals and humans, Greenberg has argued that the effects of REM deprivation, and therefore inferences about the psychological function of REM sleep, must take into account (1) the personal significance of presleep conditions; (2) the adequacy of measurement (e.g., what aspect of behavior is being measured, and how sensitive those measures are to subtle changes); and (3) individual differences which, if ignored, may cancel out real effects (R. Greenberg, in Chase, 1972, pp. 245, 250-251; Greenberg and Pearlman, 1974; also, personal communication, May, 1973). Attention to these methodological points may well advance the theory of sleep. Currently the research evidence does not unequivocally support the idea that significant adaptive psychological processes occur, or occur exclusively, in REM sleep or that the physiological and psychological functions of REM sleep differ from those of NREM sleep. The more optimistic view is that the evidence does permit such conclusions (Dewan, 1969; Greenberg and Pearlman, 1974). The more conservative view is expressed by Rechtschaffen (1971), who says that "it may not be too extravagant to describe the function of sleep as one of the major unsolved biological puzzles of our age" (p. 88).
B. REM Psychology versus REM Physiology To a large extent, hypotheses about the psychology of REM sleep have been inferred from observation and experimental intervention with the organismic (psychophysiological) state of REM. Two major attempts to separate out a psychological function more directly have been explored with respect to two questions: (1) Is there a need for a certain kind of fantasy which is served by REM sleep but which, under certain conditions, may transcend the bounds of REM? (2) Is
350
DAVID
B.
COHEN
there a need to complete a REM dream which is separate from the need for REM sleep? With regard to the first question, Cartwright and Monroe (1968) REM-deprived subjects during the first half of the night only. After each awakening subjects were given one of two types of task: to describe mental content (fantasy or primary process) or to carry out a digit-span task (supposedly a secondary-process task antithetical to fantasy). They found the usual REM rebound (higher than normal percentage of REM sleep) during the second half of the night in subjects given the digit span, but not in subjects given the fantasy task. These findings suggest that engaging in fantasy partially substituted for REM experience. In addition, they have implications for REMdeprivation methodalogy. If indeed there is a need for fantasy that is partially independent of REM sleep, then it is possible that at least some REM-deprived subjects may compensate for REM deprivation with more intense REM experiences later or may experience their fantasy quota during NREM sleep or even during waking. Under such conditions, REM rebound would not be expected. Some evidence for this derivation has been obtained (Cartwright, 1969; Pivik and Foulkes, 1966), though it is not unequivocal (Foulkes, Pivik, Ahrens, and Swanson, 1968). The need to complete a REM dream separate from the need for REM sleep has been investigated by Fiss and his colleagues (Fiss, 1969). Compared to a REM-deprivation condition, a REM-interruption condition yielded more salient, personally revealing, emotionally intense, and thematically continuous stories told to TAT cards immediately upon awakening. There was evidence of increased anger in the stories, suggesting that REM interruption was more frustrating and ps,y'chologically disruptive than either REM deprivation of REM completion. These findings suggest a need to complete a certain kind of dream experience independent of a possible need for REM sleep per se. However, they do not resolve the question of need for dreaming or even a need for REM dreaming. A follow-up study (Fiss and Ellman, 1973) included a variation of the REM-interruption method. Instead of responding to TAT stimuli, subjects were asked to respond to difficult IQ-type questions after each experimental awakening from a REM period. Fiss and Ellmann report that REM periods shortened (in the manner of a conditioned avoidance response) to the point where
DREAMING
351
subjects would return to NREM sleep prior to scheduled awakenings. Further, this diminuition of REM duration persisted during recovery sleep, even during the morning hours when REM periods usually account for a large proportion of total sleep time. Note that this effect was not obtained in two subjects who were asked to report dreams rather than to engage in small talk. This latter finding is roughly commensurate with the results of Cartwright and Monroe (1968), showing a difference of REM rebound during the second half of nights after REM deprivation with dream reporting or digit-span performance. That is, for both experimental strategies, interference with the psychological process (fantasy) induced changes in a REM sleep parameter (duration) that would not necessarily be predicted solely on the basis of a hypothesis of a physiological need but that are commensurate with the hypothesis of a psychological need. Together these data support the hypothesis that certain functional properties of dreaming are distinct from those of sleep physiology.
C. Dream Content and Psychological Change This hypothesis is further supported by some findings from a study by Cohen and Cox (1975). Rather than asking whether dreaming in general or REM dreaming in particular is necessary, or whether waking fantasy can substitute for REM dreaming, or whether there is a need to complete a dreaming experience, one can ask whether certain kinds of dreaming (like certain kinds of thinking) are more effective in promoting psychological change. The emphasis here is on dream content as a data base to permit inferences about the functional properties of the psychological process of dreaming, rather than the biological state of dreaming sleep. This question requires a somewhat different experimental strategy, specifically exploring the association between certain kinds of dream content and measures of psychological change from presleep to postsleep. The experimental strategy used in the Cohen and Cox (1975) study described below evolved from earlier work reported by Greenberg and Pearlman (1972), Cartwright (1971), Kramer and his colleagues (Kramer and Roth, 1973; Roehrs, Kramer, Lefton, Lutz, and Roth, 1973; Roth, Kramer, and Roehrs, 1973), and Witkin, Goodenough, and their colleagues (Witkin, 1969). A number of these investigators have tested
352
DAVID
B.
COHEN
the hypothesis that dreaming is adaptive by using REM-deprivation techniques or by looking at associations between behavioral and REM sleep variables. If one assumes further that adaptive dreaming may occur in NREM sleep and that dream content reflects the hypothetical adaptive process of interest, then it seems reasonable to investigate the relationship between adaptive changes and what individuals are dreaming about (Kramer and Roth, 1973). Dream experiences may be pure epiphenomena, but it is likely that they are systematically related to the psychological processes which give rise to them. Thus dream content serves as a vehicle or "window" through which inferences about adaptive processes are made (Hartmann, 1973). One strategy is to expose subjects to a presleep "problem," note differences in psychological change from presleep to postsleep, and correlate these differences in change with differences in dream content which can not be attributable to theoretically extraneous events. Consider the "problem" experimentally induced in the Cohen and Cox (1975) study. Subjects were treated impersonally, given no information that could be used to reduce uncertainty about the meaning of their experience, and isolated after experiencing difficulty or failure on items ostensibly taken from an IQ test that college students find relatively easy. Presleep-affect rating (on bipolar scales such as depressed-happy and anxious-calm), dream content, and postexperimental predebriefing ratings on a questionaire revealed evidence that the subjects were indeed impressed in a personal way with the manipulation. During the night, subjects were awakened during REM and NREM (Stage 2) sleep to report dreaming experience. The following semester two judges who had been experimenters and who were therefore familiar with the presleep situation provided reliable ratings of the presence or absence of experiment/experimenter-relevant preoccupation in the dream protocols. Of the 21 subjects exposed to the negative condition, 10 showed evidence of such preoccupation (incorporators), while 11 did not (nonincorporators). While there was no difference in presleep-affect ratings, the two groups differed markedly on postsleep-affect ratings. The difference between them in affect change was highly significant (p < .01) and accounted for about 35% of the variance. In fact, while 9 of the 10 incorporators showed an increase in positive affect, only 4 of the 10 nonincorporators did so. These findings suggest that the content of dreaming may have mediational properties with respect to an individual's "state of mind."
353
DREAMING
A second analysis utilized another measure, specifically subjects' responses on a five-step scale indicating willingness to participate again as a subject on an unpaid basis. The following results were obtained on the quasi-behavioral estimate of attitude in 7 incorporators and eight non incorporators who were followed up. The first measure (predebriefing) showed no difference in mean ratings of willingness between the incorporator and the nonincorporator subjects; on the average, their mean ratings indicated that they were "not sure" (roughly midpoint on the rating scale). However, on the second rating obtained during an unexpected telephone call 3 to 10 weeks later, all 7 of the incorporators versus only 2 of the 8 non incorporators showed increased willingness to return as subjects (p < .02). These effects (see Figure 4) could not be attributed to experimenter expectation because subject ratings were obtained by naIve interviewers and independent of any knowledge of subject's category. These results
• 0
3.00
2.00 Incorp Nonincorp
f
0.50 w
lJ.. lJ..
-3.00
1.50
a.. x
w
1.25
0
1.00
z a::
0.75
w
0.50
0
0.25
a::
I-
0.00
-0.50
a::
::l I-
0.25
-0.25
1.75
I-
0.75
u
Z
w
::;:
w
1.00
I-
I-
IJ
f IJ
Presleep
en en
0.00
~
-0.25
j
-0.50
i
-2.00
w
r1
Post sleep (predebrief 1
z
~ 1~
Post sleep (predebriefl
If1 Approx. I Month later
FIGURE 4. Changes in affect and attitude toward experimental participation for incorporator and nonincorporator subjects.
354
DAVID
B.
COHEN
suggest that along with a manifest change in postsleep affect, incorporation facilitated a potential change in attitude actualized by the debriefing process. What do these preliminary results suggest about the nature of the dream experience? First, the content of dreaming may have effects on the individual that are distinguishable from the effects of the physiology of sleep. Second, these effects may be considered adaptive in the sense that the individual is in a more positive state of mind for having dreamed in a certain manner. Third, this change in state of mind may be manifest in certain areas (e.g., affect), while latent in other areas (e.g., attitude). As with any preliminary set of data, an immediate requirement is replication and further investigation. A number of questions need to be explored in further research. For example, what kind of dream content is most facilitative of change? Specifically, does it involve "problem" orientation, evidence of concentration rather than simply reflection, evidence of primary process, or certain kinds of progressive development of theme? Is such content partially dependent on the interaction of specific subject variables and specific pre sleep conditions? Which area of psychological functioning (e.g., mood, attitude, or performance) is most affected, or what is more likely, does the strength of the effect depend on a complex interaction of individual, condition, and dream-content factors? Further, are there differential effects of REM versus NREM dreams on psychological change? That is, does a particular kind of dream have a more powerful effect if it occurs in REM rather than if it occurs in NREM sleep, or are such "effective dreams" (dreams that promote change) simply more likely to occur in REM? In addition to REM versus NREM differences in the processing of information, there may be important differences during early and later REM periods. I raise this question in the specific light of preliminary evidence that I recently obtained suggesting a relative increase from early to later REM periods in the prominence or influence of lefthemispheric activity. This evidence is composed of the following: (1) an increase from early to late REM periods in the prominence of lefthemisphere-related dream-content categories (e.g., verbal activity) but not of right-hemisphere-related dream-content (e.g., music); (2) a corresponding increase in the tendency to look to the right; (3) a corresponding tendency for the ratio of left-to-right EEG amplitude (L/
DREAMING
355
R) to diminish; (4) correlations between increase in left- (but not right-) hemisphere-related content, and (a) increase in right looking and (b) decrease in LlR ratio and (5) the failure of subjects with a sinistral bias to fit the observed pattern. These very preliminary results suggest the possibility that information-processing shifts in quality during the night, perhaps even favoring different kinds of problems (recent events early versus remote events later?) with somewhat different kinds of "treatment" emphases (more analogical early versus more analytical later?). While this speculation rests on the slimmest of empirical bases, it does suggest the interesting possibility that both the representative characteristics and the differential effectiveness of certain kinds of dreaming (e.g., representation of a presleep "problem") may, in part, be determined by both type and temporal position of the sleep stage within which that dreaming takes place. Were it confirmed that in most cases problem-oriented dreaming is an effective vehicle for promoting desirable change in the individual, would it be possible to encourage such change by experimental manipulation of dream content through presleep or sleep suggestion? Perhaps it is not too optimistic to hope that the use of experimental strategies such as the one just outlined will bring us closer to a psychophysiological theory of dreaming, that is, a theory that specifies the relation between dream content and factors such as personality, presleep conditions, sleep physiology, and behavioral change. For the present, investigation of the dreaming experience as a mediator of psychological change should provide a useful test of the hypothesis that "the function of the dream is ... important in itself for the processing of experiences and the formation of attitudes, behavior patterns, and personality" (Shapiro, 1967, p. 79). REFERENCES ADELSON, J. Creativity and the dream. Merrill-Palmer Quarterly, 1959,6, 91-97. ANTROBUS, J. 5., ANTROBUS, J. 5., AND SINGER, J. L. Eye movements accompanying daydreaming, visual imagery, and thought suppression. Journal of Abnormal and Social Psychology, 1964, 69, 244-252. ANTROBUS, J. 5., DEMENT, W. C., AND FISHER, C. Patterns of dreaming and dream recall. Journal of Abnormal and Social Psychology, 1964, 69, 341-344. ATINEAVE, F. How do you know? American Psychologist, 1974,29, 493-499. AUSTIN, M. D. Dream recall and the bias of intellectual ability. Nature, 1971, 59-61. BAEKELAND, F. Effects of presleep procedure and cognitive style on dream content. Perceptual and Motor Skills, 1971,32, 63-69.
356
DAVID B. COHEN
BAKAN, D. The duality of human existence. Chicago: Rand McNally, 1966. BARBER, B. Factors underlying individual differences in rate of dream reporting Paper presented to the Association for the Psychophysiological Study of Sleep, Boston, March,1969. BARDWICK, J. Psychology of women: A study of biosocial conflicts. New York: Harper and Row, 1971. BARTLETt, F. C. Remembering. Cambridge, England: Cambridge University Press, 1932. BERGER, R. J. Oculomotor control: A possible function of REM sleep. Psychological Review, 1969,76, 144-164. BERTINI, M. REM Sleep as a psychophysiological "agency" of memory organization. In W. P. KOELLA AND P. LEVIN (Eds.), Sleep: Physiology, biochemistry, psychology, pharmacology, clinical implications. Basel, Switzerland: S. Karger, 1973. BERTINI, M., LEWIS, H. B., AND WITKIN, H. A. Some preliminary observations with an experimental procedure for the study of hypnagogic and related phenomena. In C. TART (Ed.), Altered states of consciousness. Garden City, N.Y.: Doubleday, 1972. BERTINI, M., AND PONTALTI, C. Clinical perspectives of a new technique in dreaming research. Paper presented to the Association of the Psychophysiological Study of Sleep, Bruges, Belgium, June, 1971. BONIME, W. The clinical use of dreams. New York: Basic Books, 1962. BREGER, L. Function of dreams. Journal of Abnormal Psychology Monograph, 1967,72 (No. 5, whole No. 641), 1-28. BREGER, L., HUNTER, J., AND LANE,R. W. The effect of stress on dreams. Psychological Issues, Monograph No. 27, 1971,7, 1-210. BRYSON, D., AND SCHACHER, S. Behavioral analysis of mammalian sleep and learning. Perspectives in Biology and Medicine, Autumn, 1969, 71-79. BYRNE, D. Repression-sensitization as a dimension of personality. In B. A. MAHER (Ed.), Progress in experimental personality research (Vol. 1). New York: Academic Press, 1964. CARLSON, R. Sex differences in ego functioning: Exploratory studies of agency and communion. Journal of Consulting and Clinical Psychology, 1971,37, 267-277. CARRINGTON, P. Dreams and schizophrenia. Archives of General Psychiatry, 1972, 26, 343-350. CARTWRIGHr, R. D. Dreams, reality, and fantasy. In J. FISHER AND L. BREGER (Eds.), The meaning of dreams: Recent insights from the laboratory. Sacramento, Calif.: Bureau of Research, California Department of Mental Hygiene, 1%9. CARTWRIGHr, R. D. Problem solving in REM, NREM, and waking. Paper presented to the Association of the Psychophysiological Study of Sleep, Bruges, Belgium, Juno:, 1971. CARTWRIGHr, R. D. The influence of a conscious wish on dreams: A methodological study of dream meaning and function. Journal of Abnormal Psychology, 1974,83, 387-393. CARTWRIGHr, R. D., BERNICK, N., BOROWnz, G., AND KLING, A. Effect of an erotic movie on sleep and dreams of young men. Archives of General Psychiatry, 1969,20, 262271. CARTWRIGHr, R. D., AND MONROE. L. J. The relation of dreaming and REM sleep: The effects of REM deprivation under two conditions. Journal of Personality and Social
Psychology, 1968,10, 69-74.
CHASE, M. H. (Ed.), The sleeping brain. Los Angeles: Brain Information Service: Brain Research Institute, 1972.
DREAMING
357
CHERNIK, D. A. Effect of REM sleep deprivation on learning and recall by humans. Perceptual and Motor Skills, 1972,34, 282-294. COHEN, D. B. Dream recall and short term memory. Perceptual and Motor Skills, 1971,33, 867-871. COHEN, D. B. Presleep experiences and home dream reporting: An exploratory study. Journal of Consulting and Clinical Psychology, 1972, 38, 122-128. COHEN, D. B. Sex role orientation and dream recall. Journal of Abnormal Psychology, 1973,82, 246-252. COHEN, D. B. Effect of personality and presleep mood on dream recall. Journal of Abnormal Psychology, 1974a, 83, 151-156. COHEN, D. B. Presleep mood and dream recall. Journal of Abnormal Psychology, 1974b, 83, 45-51. COHEN, D. B. Toward a theory of dream recall. Psychological Bulletin, 1974c, 81, 138-154. COHEN, D. B., AND Cox, D. Neuroticism in the sleep laboratory: Implications for representational and adaptive properties of dreaming. journal of Abnormal Psychology, 1975, 84, 91-108. COHEN, D. B., AND MACNEILAGE, P. F. A test of the salience hypothesis of dream recall. Journal of Consulting and Clinical Psychology, 1974,42, 699--703. COHEN, D. B., AND WOLFE, C. Dream recall and repression: Evidence for an alternative hypothesis. Journal of Consulting and Clinical Psychology, 1973,41, 349--355. DALLETI, J. Theories of dream function. Psychological Bulletin, 1973,79,408-416. DEMENT, W. C. The effect of dream deprivation. Science, 1960,131, 1705-1707. DEMENT, W. C. Some must watch while some must sleep. Stanford, Calif.: Stanford Alumni Association, 1972. DEWAN, E. M. The programming (P) hypothesis for REMs. Physical Science Research Papers, No. 388, Air Force Cambridge Research Laboratories, Project 5628, 1969. DOMHOFF, B. Home dreams vs. laboratory dreams: Home dreams are better. In M. KRAMER (Ed.), Dream psychology and the new biology of dreaming. Springfield, Ill.: Thomas, 1969. EMPSON, J. A. c., AND CLARKE, P. R. F. Rapid eye movements and remembering. Nature, 1970,227, 287-288. EPHRON, H. S., AND CARRINGTON, P. Rapid eye movement sleep and cortical homeostasis. Psychological Review, 1966,75, 500-526. ERIKSON, E. E. The dream specimen of psychoanalysis. Journal of the American Psychoanalytic Association, 1954, 2, 5-56. EVANS, F. J., GUSTAFSON, 1. A., O'CONNELL, D. N., ORNE, M. T., AND SHOR, R. E. Verbally induced behavioral responses during sleep. Journal of Nervous and Mental Disease, 1970,150, 171-187. EYSENCK, H. J., AND RACHMAN, S. The causes and cures of neurosis. San Diego: Robert Krapp, 1965. FISHBEIN, W., AND KASTANIOTIS, C. Augmentation of REM sleep after learning. Sleep Research, 1973,2, 94. FISS, H. The need to complete one's dreams. In J. FISHER AND 1. BREGER (Eds.), The
meaning of dreams: Recent insight from the laboratory, Research Symposium No.3. Sacramento, Calif. Bureau of Research, California Department of Mental Hygiene, 1969. FISS, H., AND ELLMAN, S. REM sleep interruption: Experimental shortening of REM period duration. Psychophysiology, 1973,10, 510-516.
358
DAVID B. COHEN
FOULKES, D. Dreams of the male child: Four case studies. Journal of Child Psychology and Psychiatry, 1967,8, 81-97. FOULKES, D. Personality and dreams. International Psychiatry Clinics, 1970, 7, 1947-1953. FOULKES, D., PIVIK, T., AHRENS, J. B., AND SWANSON, E. M. Effects of "dream deprivation" on dream content: An attempted cross-right replication. Journal of Abnormal Psychology, 1968,73, 403-415. FOULKES, D., AND RECHTSCHAFFEN, A. Presleep determinants of dream content: Effects of two films. Perceptual and Motor Skills, 1964,19, 983-1005. FREUD, S. The interpretation of dreams. New York: Basic Books, 1953. FROMM, E. The forgotton language. New York: Grove Press, 1951. GALIN, D. Implications for psychiatry of left and right cerebral specialization. Archives of General Psychiatry, 1974,31, 572-583. GOODENOUGH, D. R. Some recent studies of dream recall. In H. A. WITKIN AND B. LEWIS (Eds.), Experimental studies of dreaming. New York: Random House, 1967. GOODENOUGH, D. R., WITKIN, H. A., LEWIS, H. B., KOULACK, D., AND COHEN, H. Repression, interference, and field dependence as factors in dream forgetting. Journal of Abnormal Psychology, 1974,83, 32-44. GREENBERG, R., AND PEARLMAN, C. Sleep and dream patterns in a patient in psychoanalysis: An attempt at psychophysiological correlations. Paper presented to the Association for the Psychophysiological Study of Sleep, Lake Minnewaska, New York, May, 1972. GREENBERG, R., AND PEARLMAN, C. A. Cutting the REM nerve: An approach to the adaptive role of REM sleep. Perspectives in Biology and Medicine, Summer, 1974, 513521. HALL, C. S. The meaning of dreams. New York: McGraw-Hill, 1966. HALL, C. S. Normative dream-content studies. In M. KRAMER (Ed.), Dream psychology and the new biology of dreaming. Springfield, Ill.: Thomas, 1969. HALL, C. S., AND VAN DE CASTLE, R. 1. The content analysis of dreams. New York: Appleton-Century-Crofts, 1966. HARTMANN, E. The functions of sleep. New Haven, Conn.: Yale University Press, 1973. HARTMANN, E., AND STERN, W. E. Desynchronized sleep deprivation: Learning deficit and its reversal of increased catecholamines. Physiology and Behavior, 1972,8, 585587. HAURI, P. Evening activity, sleep mentation, and subjective sleep quality. Journal of Abnormal Psychology, 1970,76, 270-275. HAURI, P. White noise and dream reporting. Sleep Research, 1972, 1, 124. HISCOCK, M., AND COHEN, D. B. Visual imagery and dream recall. Journal of Research in Personality, 1973,7.,. 179-188. JOHNSON, 1.,c. Are stages of sleep related to waking behavior? American Scientist, 1973, 61, 326-338. JONES, B. M., AND PARSONS, O. A. Alcohol and consciousness: Getting high and coming down. Psychology Today, January, 1975, 53-58. KLEITMAN, N. Basic rest-activity cycle in relation to sleep and wakefulness. In A. KALES (Ed.), Sleep: Physiology and pathology: A symposium. Philadelphia: Lippincott, 1969. KLINGER, E. Struciure and Function of Fantasy. New York: Wiley-lnterscience, 1971. KLINGER, E. Consequences of commitment to and disengagement from incentives. Psychological Review, 1975,82, 1-25. KOELLA, W. P. AND LEVIN, P. (eds.). Sleep: Physiology, Biochemistry, Psychology, Pharmacology, Clinical Implications. Basel, Switzerland: S. Karger, 1973.
DREAMING
359
KRAMER, M., BALDRIDGE, B. A., WHITMAN, R. M., ORNSTEIN, P. H., AND SMITH, B. A. An exploration of the manifest dream in schizophrenia and depressed patients. Diseases of the Nervous System, 1969,30, 126-130. KRAMER, M., AND ROTH, T. The mood regulating function of sleep. In W. P. KOELLA AND P. LEVIN (Eds.), Sleep: Physiology, biochemistry, psychology, pharmacology, clinical implications. Basel, Switzerland: S. Karger, 1973. KRIPPNER, S., AND HUGHES, W. Dreams and human potential. Journal of Humanistic Psychology, 1970,10, 1-20. LANGFORD, G. W., MEDDIS, R., AND PEARSON, J. D. Awakening latency from sleep for meaningful and nonmeaningful stimuli. Psychophysiology, 1974,11, 1-5. . LEHMANN, D., AND KoUKKou, M. Learning and EEG during sleep in humans. In W. P. KOELLA AND P. LEVIN (Eds.), Sleep: Psysiology, biochemistry, psychology, pharmacology, clinical implications. Basel, Switzerland: S. Karger, 1973. LEHMANN, D., AND KOUKKOU, M. Computer analysis of EEG wakefulness-sleep patterns during learning of novel and familiar sentences. Electroencephalography and Clinical Neurophysiology, 1974,37, 73-84. MOLINARI, S., AND FOULKES, D. Tonic and phasic events during sleep: Psychological correlates and implications. Perceptual and Motor Skills, Monograph Supplement No. 1, 1969,29,343-368. OKUMA, T., SUNAMI, Y., FUKUMA, E., TAKEO, S., AND MOTOIKE, M. Dream content study in chronic schizophrenics and normals by REMP-awakening technique. Folia Psychiatrica et Neurologica Japonica, 1970,24, 151-162. OVERTON, D. A. State-dependent retention of learned responses produced by drugs: Its relevance to sleep learning and recall. In W. P. KOELLA AND P. LEVIN (Eds.), Sleep: Physiology, biochemistry, psychology, pharmacology, clinical implications. Basel, Switzerland: S. Karger, 1973. PEARLMAN, C. A. The adaptive function of dreaming. In E. HARTMAN (Ed.), Sleep and dreaming. Boston: Little, Brown, 1970. PEARLMAN, C. A., AND BECKER, M. Brief posttrial REM sleep deprivation impairs discrimination learning in rats. Physiological Psychology, 1973,1, 373-376. PEARLMAN, C. A., AND GREENBERG, R. Brief REM deprivation impairs consolidation of complex learning in rats. Paper presented to the Association for the Psychophysiological study of sleep, New York, May, 1972. PIVIK, T., AND FOULKES, D. Dream deprivation: Effects on dream content. Science, 1966, 158, 1282-1284. RECHTSCHAFFEN, A. Dream reports and dream experiences. Experimental Neurology, Supplement 4, 1967,4-15. RECHTSCHAFFEN, A. The control of sleep. In W. A. HUNT (Ed.), Human behavior and its control. Cambridge, Mass.: Schenkman Press, 1971. RECHTSCHAFFEN, A. The psychophysiology of mental activity during sleep. In J. McGUIGAN AND R. A. SCHOONOVER (Eds.), The psychophysiology of thinking. New York: Academic Press, 1973. ROEHRS, T., KRAMER, M., LEFTON, W. J., LUTZ, T. E., AND ROTH, T. Mood before and after sleep. Sleep Research, 1973,2, 95. ROFFWARG, H. P., MUZIO, J., AND DEMENT, W. C. The ontogenetic development of the human sleep dream cycle. Science, 1966,152, 604-618. ROTH, T., KRAMER, M., AND ROEHRS, T. The relationship between sleep physiology and mood. Sleep Research, 1973,2, 96. SCHECHTER, N., SCHMEIDLER, G., AND STAAL, M. Dream reports and creative tendencies
360
DAVID
B.
COHEN
in students of the arts, sciences, and engineering. Journal of Consulting Psychology, 1965,29, 415-421. SELIGMAN, M. E. P. On the ger.erality of the laws of learning. Psychological Review, 1970, 77, 406-418. SEYFRIED, B. A., AND HENDRICK, C. When do opposites attract? When they are opposite in sex and sex role attitudes. Journal of Personality and Social Psychology, 1973,25, 1~20.
SHAPIRO, A. Dreaming and the physiology of sleep: A critical review of some empirical data on a proposal for a theoretical model of sleep and dreaming. Experimental Neurology, Supplement 4, 1967,56-81. STARKER, S. Daydreaming styles and nocturnal dreaming. Journal of Abnormal Psychology, 1974,83, 52-55. STOYVA, J., AND KAMIYA, J. Electrophysiological studies of dreaming as the prototype of a new strategy in the study of consciousness. Psychological Review, 1968, 75, 192205. TART, C. T., AND DICK, L. Conscious control of Dreaming: I. The post hypnotic dream. Journal of Abnormal Psychology, 1970,76,304-315. TAUB, J. M. Dream recall and content following various durations of sleep. Psychonomic Science, 1970,18, 82. ULLMAN, M., Dreaming, life style and physiology. Journal of Individual Psychology, 1962, 18, 1&-25. VAN DE CASTLE, R. L. Problems in applying methodology of content analysis. In M. KRAMER (Ed.), Dream psychology and the new biology of dreaming. Springfield, Ill.: Thomas, 1969. VAN DE CASTLE, R. L. The psychology of dreaming. Morristown, N.J.: General Learning Press, 1971. VAUGHN, C. The development and use of an operant technique to provide evidence for visual imagery in the rhesus monkey under sensory deprivation. Doctoral dissertation, University of Pittsburgh, 1964. VOGEL, G. W. REM deprivation: III. Dreaming and Psychosis. Archives of General Psychiatry, 1968,18,312-329. VOGEL, G. W., AND TRAUB, A. C. REM deprivation: I. The effect of schizophrenic patients. Archives of General Psychiatry, 1968,18,287-300. VOGEL, G. W., TRAUB, A. c., BEN-HORIN, P., AND MEYERS, G. M. REM deprivation: II. The effects on depressed patients. Archives of General Psychiatry, 1968,18, 301-311. WEBB, W. B. AND KERSEY, J. Recall of dreams and the probability of stage I-REM sleep, Perceptual and Motor Skills, 1967,24, 627-630. WEISZ, R., AND FOULKES, D. Home and laboratory dreams collected under uniform sampling conditions. Psychophysiology, 1970,6, 58&-596. WITKIN, H. A. Presleep experience and dreams. In J. FISHER AND L. BREGER (Eds.), The meaning of dreams: Recent insights from th.e Laboratory «;,alifornia Mental Health Research Symposium No.3). Sacramento, Calif.: Bureau· of Research, California Department of Mental Hygiene, 1969.
9
Biofeedback and the Twilight States of Consciousness THOMAS
H.
BUDZYNSKI
In recent years the exploration of altered states of consciousness (ASCs) has intrigued scientist and nonscientist alike. These deviations from the normal conscious state have been produced by the ingestion of drugs, hypnosis, meditative techniques, religious practices, severe or prolonged physical or psychological stress, sensory isolation, sensory overload, and so on. All of these conditions can be mapped onto the arousal-level continuum. When this is done we see that altered states of consciousness usually are also altered states of arousal level. It would seem that when the patterning of input to the brain from internal and external stimuli is unusual, out of the ordinary, then the experience may be labeled an altered state. One of the difficulties associated with the study of an altered state is the high variability in the state from moment to moment. For example, the transition zone between wakefulness and sleep is typically a short-lived phenomenon since the individual either passes into deeper sleep or fully awakens. At the other end of the continuum one encounters the dangers of sustained heightened arousal with excessive motor behavior which may take the form of convulsions. Although it is not yet known whether high arousal ASCs may be more useful or productive than those characterized by low arousal, the fact is that the latter certainly are easier to deal with and are potentially less harmful. Therefore the approach to be followed in the rest of this effort will be a THOMAS H. BUDZYNS:\
361
362
THOMAS
H. BUDZYNSKI
consideration of the evidence and potential usefulness of low-arousal states.
1. THE TWILIGHT
STATE
There is in the everyday course of experiential events a relatively small fraction of time spent in a special state. Typically we bracket this state by some 16 hours of waking, reality-oriented activity on one side and perhaps 8 hours of sleep on the other. The waking state generally is characterized by active involvement with the external environment and the conscious processing of information. This processing involves logical thinking, categorization, and compartmentalization, as well as the negation, rejection, acceptance, and distortion of the information. However~ during sleep our contact with the external environment is greatly diminished. Information processing is of a regressive, primitive sort. The conscious properties mentioned above are absent. Internally generated, emergent material is the major informational source. In contrast to the waking and sleep states is the transitory condition wherein one is neither fully awake nor deep asleep. This period has been referred to as the "reverie" state (Koestler, 1964), the "twilight" state (Budzynski, 1973; Budzynski and Peffer, 1973), and, as quoted in Green, Green, and Walters (1971), the "fringe of consciousness" (James, 1950), the "pre-conscious" (Kubie, 1958), the "offconscious" and the "transliminal mind" (Rugg, 1963), and the "transliminal experience" (MacKinnon, 1964). Sleep researchers tend to refer to this zone of consciousness in terms of the hallucinatorylike imagery that is associated with it-hypnagogic if falling asleep and hypnopompic if awakening. Moreover sleep researchers have defined this twilight state in terms of brain-wave patterns and eye movements. Foulkes and Vogel (1965), for example, speak of the drowsy period just preceding Stage 1 sleep as characterized by a slowing of the alpha rhythm (8-12 Hz) accompanied by slow, rolling eye movements (SEMs). As the sleeper passes into Stage 1 sleep, the slowed alpha rhythm begins to break up and is replaced by the even slower, smaller-amplitude theta rhythm (4-7 Hz). As one falls asleep, the duration of this transition from a relaxed, waking alpha pattern to the
BIOFEEDBACK AND THE TWILIGHT STATES OF CONSCIOUSNESS
363
disappearance of the alpha and the appearance of theta is roughly 5-10 minutes.
A. Is a Twilight State the Source of Creative Ideas? During this rather brief period, which will be referred to as the twilight state, people typically report emergent, hallucinatory, dreamlike experiences which are more disjointed and brief than those dreams associated with rapid eye movement (REM) sleep.! These hypnagogic images often resemble static photographic stills which have a vivid, live-in quality. As mentioned by Green et al. (1971), subjects in a study of McKellar and Simpson (1955) reported four main characteristics of hypnagogic images: vividness, independence of conscious control, originality, and changefullness. A number of famous individuals from the fields of science, music, literature, and art have credited the imagery produced during the twilight state for creative solutions or inspiring thoughts, Koestler (1964), in recounting many examples of this phenomenon, concluded that the temporary relinquishing of conscious controls liberates the mind from certain constraints which are necessary to maintain the disciplined routine of thoughts but may become an impediment to the creative leap; at the same time other types of ideation on more primitive levels of mental organization are brought into activity. (p. 169)
The scientist H. W. Magoun (1969) stated that vigilant attention may impede the free-flowing retrieval of previously stored material leading to innovative associations. The most significantly productive of creative ideas, relating disparate information and providing a new insight or synthesis, have often been reported to be generated not in the intensity of strained focused concentration upon the subject or problem, but rather during a contrasting state of mind, as when performing a stereotyped activity, or gazing hypnotically into a glowing fireplace, or even during sleep itself, in dreams. (p. 185)
In their laboratory at the Menninger Foundation, the Greens and Dale Walters have been studying the relationship between twilight states, brain-wave rhythms, and creativity. For example, in one 1
See Stoyva (1973) for an excellent discussion of the conditions contributing to the hallucinatory phenomena characteristic of this state.
364
THOMAS
H.
BUDZYNSKI
experiment the electroencephalographic (EEG) patterns of three demonstrably creative individuals (a professor of physics, a psychiatrist, and a psychologist) were recorded as they maintained a reverie or twilight state which they associated with creative thoughts. Each of these subjects was self-trained over a period of 15-30 years in internal scanning techniques. The EEG records showed that two of these individuals produced an unusually high percentage of theta waves. Both reported what they called their customary hypnagogiclike imagery as they did their internal scanning-a looking inward with the exclusion of external stimuli. The third subject showed a slowing of his alpha EEG frequency from 9.5 Hz down to 8.3 Hz and reported that this was a preliminary mind-quieting imageless state in moving toward a deeper state (Green et al., 1971). If the twilight state is indeed associated with the emergence of creative insights, then perhaps creativity can be enhanced in individuals if they are trained to produce or at least prolong twilight or reverie states. Of course there is a related problem here--that of retrieving creative material from the twilight state.
B. Biofeedback and Creativity The Menninger team has developed a psychophysiological technique called autogenic feedback training, through which subjects are taught to control and manipulate various internal states (Green, Green, and Walters, 1970). A combination of self-directed phrases, breathing exercises, and the new technique of biofeedback is used to produce a relaxed, quiet, inward-turned state of mind. The biofeedback involves the sensing, amplification, and information feedback of the alpha and theta brain-wave rhythms. During a practice session the subject hears a tone with a frequency of 900 Hz if alpha waves are present. A tone of 400 Hz is heard if the subject produces theta waves. Elapsed-time indicators provide the total session time as well as beta, alpha, and theta time for the session. During a typical home-training session the subject first does the autogenic relaxation, i.e., the repeats to himself silently phrases that tend to promote a quieting of the autonomic functions as well as the skeletal muscle system. After 5 minutes of this "exercise" the subject
BIOFEEDBACK AND THE TWILIGHT STATES OF CONSCIOUSNESS
365
begins 1 minute of hyperventilation for activation of the central nervous system, followed by 6 minutes of even, equalized breathing. The attention is focused on the breath at the nostrils during this breathing procedure. Finally, the biofeedback unit is turned on and feedback practice begins. Approximately 40 minutes are spent attempting to produce theta rhythms, although both alpha and theta feedback are used. After the session the subject records his subjective experiences in a notebook. There is a second phase to this study which is quite important. After subjects have learned to produce theta states they next practice the verbal reporting of these states-in other words, retrieving the hypnagogic imagery so that it can be brought to full awareness. In order to enhance this awareness during their practice sessions, the subjects use "repeat" alarm clocks as alerting devices to bring them out of reverie in order to report. During the session the subject must press the bar on the top of the clock every 5-7 minutes or the alarm will ring. Each time he depresses the bar, or is alerted by the alarm, he makes a note of his hypnagogic imagery. Thus far the Menninger studies show that subjects can learn to increase the amount of theta frequencies in their EEGs. Moreover, the subjects do seem to be more aware of the hypnagogic imagery associated with this state (Green, Green, and Walters, 1973). The most common type of imagery reported by the subjects is visual, the next most common is auditory, and then somatic (body size, shape, and position changes). Tactile, gustatory, and olfactory sensations are rare.
1. Other Experiences Resulting from Theta Training In addition to an increased ability to report hypnagogic imagery the subjects in the Menninger studies reported a change in their awareness of dreams. Recall of the dreams improved, and the vividness and meaningfulness of the dreams increased as well. Forgotten memories, often of childhood, were recalled during the theta sessions. Perhaps the most important result so far (all data have not been analyzed) is that subjects say that the integrative experiences are associated with extended alpha-theta practice. A few subjects felt fatigued, sluggish, or nervous for the first week or two of training. These effects, fortunately, were only temporary. Later there were
366
THOMAS
H.
BUDZYNSKI
many reports of calmer, more peaceful, more relaxed feelings. Other comments included "greater energy," "clearer thinking," and "better concentration." Green et ai. (1973) also noted that a number of subjects experienced Jungian "archetypal images." Images of tunnels, stairways, caves, and pyramids were seen. A wise old man, a teacher or professor of some sort, would appear to offer advice. A book containing desired knowledge was a recurring image. This sort of imagery was no doubt potentiated to a great extent by the needs and desires of the student subjects. However, not all of the theta experiences were of a positive nature. As Green et al. (1973) concluded, "To the person experiencing it, hypnagogic imagery may seem meaningful or nonsensical, enjoyable or distressing." These researchers also noted that some psychiatrists and neurologists have associated theta only with psychopathology, but that may be because they have observed it only in patients. 2
2. But Is It Really Related to Creativity? Evidently many creative individuals credit the reverie or twilight state and its hypnagogic imagery with the generation of novel ideas. Brain-wave studies of this condition show that it is characterized by slowed EEG frequencies with theta waves (4-7 Hz) predominating. When subjects learn to produce this brain-wave pattern through biofeedback training they also report the emergence of hypnagogic imagery. Further training results in an increased ability to bring such imagery to full awareness. The subjective reports from the latest Menninger study are now being analyzed according to an imageclassification scheme adapted from the Wallach and Kagan study of creativity variables (Green et ai., 1973). Although the reports obtained in earlier studies by Green et ai. were encouraging, their anecdotal nature leaves them somewhat open to criticism. The image-classification scheme to be used in this study should lend a great deal of credence to the phenomenon. 2
Kubie and Margolin (1942) used respiratory-sound feedback with patients to facilitate the recovery of emotionally charged material. While concentrating on the sound of their own breathing the patients would enter a hypnagogic state which was said to enhance free association.
BIOFEEDBACK AND THE TWILIGHT STATES OF CONSCIOUSNESS
367
C. Learning in the Twilight State? If the twilight state can be recognized by the generation of certain brain-wave frequencies and bizarre, disjointed, and potentially creative imagery, it also can be characterized by a loosening of the realityoriented frame of reference and a suspension of the critical cognitive faculties. Moreover evidence from several areas of research, including sleep, sleep-learning, arousal level, sensory deprivation, hypnosis, attitude change, altered states, and psychotherapy, supports the concept of twilight-state hypersuggestibility (Budzynski, 1973). T. X. Barber (1957), for example, showed that subjects were just as suggestible when in a light sleep or in a drowsy condition as when hypnotized. A remark by one of the subjects is significant: "1 was just sleepy enough to believe what you were saying is true. I couldn't oppose what you wanted with anything else" (p. 59). Barber stated that at the therapeutic level it is possible that suggestions could be presented to people while they sleep for purposes of helping overweight people reduce, getting heavy smokers to cut down, and helping timid people gain confidence.
1. Attitude Change A dissertation study at Yale University produced an interesting result that supported Barber's statement. The study (Felipe, 1965) tested the effects of attitude-change information presented via tape recordings to subjects during waking, drowsy, and deep-sleep conditions. A portion of the attitude information concerned interracial dating. Felipe used several pre-post attitudes scales to measure change that may have occurred during the three conditions. Only in the condition where the subjects were presented the message while drowsy did the attitude change reach significance. This findings is consistent with the premise that attitude change is potentiated in a drowsy or light-sleep state because of a lowering of defenses. Changes were negligible in the waking condition, perhaps because these defenses were intact. Little or no effect was seen as a result of presenting the attitude-change material during deeper sleep. Felipe's research demonstrates the relatively uncritical acceptance of information even though the information may differ from the
368
THOMAS
H.
BUDZYNSKI
subject's present belief system. Further support for this notion has been generated by the experimentation on sensory deprivation. Heron (1961) found that a group of sensory-deprived subjects showed significantly greater changes in attitude toward psychic phenomena than did a control group which was not sensory-deprived. Both groups heard a record which argued for belief in various types of psychical phenomena.
2. Complexities of Material and Learner Although the evidence accumulates that people can assimilate information and perhaps show attitude changes as a result of learning in restricted-environment, low-arousal conditions, several important parameters still need to be considered. These parameters are the complexity of the material and the cognitive style of the learner. Suedfeld (1964) presented abstract and complex as well as concrete attitude-change material to subjects who were conceptually simple or conceptually complex individuals. The environment conditions were sensory deprivation and a control condition of nonconfinement. Suedfeld found that the sensory deprivation produced more change in attitude and that conceptually concrete subjects changed more than the abstract-thinking individuals. However, his results also indicated that the sensory deprivation reduced tlre subject's ability to achieve complex differentiations and integrations of incoming information, thus leading to all-or-nothing responses, especially in the more concrete personality types. The all-or-nothing response simply means that an individual either completely accepts or completely rejects the information and the new attitude which is advocated. Suedfeld cited the study by Myers, Murphy, and Smith (1963), in which a one-sided argument presented to initially hostile subjects who were sensory-deprived failed to produce attitude change, probably because the attempt to manipulate their attitude was so obvious as to arouse resistance.
3. Psychotherapy and Sensory Deprivation In a more clinical vein, Gibby and Adams (1961) reported that the self-concepts of sensorially deprived psychiatric patients showed more change in response to a taped message than did those of a nondeprived group. Adams (1965) later reported a case in which a program
BIOFEEDBACK AND THE TWILIGHT STATES OF CONSCIOUSNESS
369
of this sort was used to facilitate the reduction of symptomatology as well as the increase of self-awareness, self-acceptance, and insight in a psychiatric patient. In fact, a rather extensive research program by Adams and his co-workers resulted in these conclusions: 1. Disruptive phenomena rarely occurred during relatively short
periods of mild sensory deprivation and social isolation. 2. Unlike "normal" subjects, psychiatric patients as a group showed substantial positive changes following exposure to those conditions. 3.. The greater the degree of personality disturbance in these patients prior to deprivation, the greater was the improvement observed afterward. 4. A standardized prerecorded tape-message presented to individual patients during exposure to deprivation was more readily accepted and produced more positive changes in selfconcept measures than the same message presented to a control group under non deprivation conditions. 5. The patients showed generalized improvement, enhancement of insight, and increased realistic self-awareness. 4. "Suggestopaedia"
The presentation of information to individuals who are in lowarousal states has been the focus of a great deal of attention ever since the work of Lozanov, a Bulgarian scientist, became known several years ago (see Ostrander and Schroeder, 1970). Lozanov allowed students in his tutorial "suggestopaedia" program to relax in comfortable chairs while focusing their attention on classical music. The learning material was presented along with the music, and the voice intonation of the instructor was synchronized to the changes of tempo and volume of the music. The students were not to concentrate on the learning material but rather on the music. Comprehension is said to be very high even for students with "mental blocks" for the material. It is reported that the learning is both qualitatively and quantitatively different from that occurring in the normal, waking state. Lozanov claimed that the learning is more intuitive and more holistic and is retained longer than the rote, specific, often short-lived learning usually associated with that obtained under the normal, alert conditions. An interesting observation is that after a session of suggesto-
370
THOMAS
H.
BUDZYNSKI
paedic learning most students reported that they learned little or nothing at all, yet they scored high on the comprehension tests. In this country Elizabeth Philipov of Pepperdine University has adapted Lozanov's technique for the instruction of English-speaking students in Cyrillic-alphabet foreign languages. The study was the first application of the suggestopaedic method in a university setting, as well as the first experiment of this sort in the West. The Cyrillic alphabet was mastered in less than 6 hours. In 120 hours of training, the students learning Bulgarian assimilated a volume of 1800 new words and were able to use them in reading, writing, and speaking the language. The oral proficiency of the suggestopaedic students after 120 hours was compared with students having had 360 hours of Russian taught by traditional methods. Judges from the department of linguistics rated the students on oral proficiency. A Mann-Whitney Utest for nonparametric, rank-ordered data showed that even with only one-third of the training time the suggestopaedic students were more proficient than those taught with traditional techniques (Philipov, personal communication, 1975).
5. Sleep Learning (or Not?) Another area of research that bears on the question of the assimilation of material during low-arousal conditions is sleep learning. Felipe's research suggested that the material presented during deep sleep did not seem to be assimilated. This result is in general agreement with the majority of sleep-learning reports from this country and the United Kingdom (see Simon and Emmons, 1955, 1956; Bruce, Evans, Fenwick, and Spencer, 1970). Although the evidence indicates that depth of sleep is inversely related to the assimilation of auditory material, some learning apparently can take place even in the deepest sleep state (Stage 4). Levy, Coolidge, and Staab (1972) presented pairs of Russian and English equivalents for five nights during stages REM (rapid eye movement) and 4 (the deepest sleep stage). Using a multiple-choice recognition task they found that subjects could identify the correct words above chance. Different word-pairs were presented each night so that it could be determined whether subjects were "learning to learn." They were, indeed, since they improved their scores on each succeeding night. In some cases the same word-pairs were presented on two successive nights. The
BIOFEEDBACK AND THE TWILIGHT STATES OF CONSCIOUSNESS
371
recognition scores improved on the second test, indicating that repetition of material across nights was facilitative. It is important to note that in this study a preparatory set to learn and to remember the material was established. All subjects were given a suggestion prior to sleep and before each sleep presentation of the word pairs that they would learn and remember the material. Conditioning effects during the deeper sleep stages have been documented (Lindsley, 1957; Williams, Morlock, and Morlock, 1966); however, the processing of information changes as sleep deepens. A very recent experiment by Lasaga and Lasaga (1973) produced these conclusions: 1. Even during Stages 3 and 4 some perception of verbal stimuli is possible during sleep. 2. There is a progressive blurring of perception from stage 1 and REM to Stages 3 and 4. 3. Some forms of learning seem to be possible during deeper sleep (e.g., association of words), but perceptual distortions make extremely unlikely the assimilation of complex verbal materials. 4. Some subjects reported they heard nothing, yet they did well on recognition tests, suggesting some subliminal perception during sleep. Even if learning during deep sleep is difficult to implement, there is a good deal of evidence that learning can take place quite regularly during lighter sleep stages. Moscu and Vranceanu (1967), for example, presented lists of emotional and nonemotional words to subjects during the first cycle of sleep. Upon awakening, subjects were able to recall 22% of the words and to recognize 59% of them from a list of 60 words. It is interesting to note that the subjects recognized more of the emotional rather than the nonemotional words. Rubin has long been a champion of sleep learning and has defended the Russian tutorial sleep-learning technique known as "hypnopaedia" (1968, 1970). He has reminded us that unlike most other studies the Soviet experiments (and tutorial programs) incorporate repetitive practice over several days or even months and that there is a great emphasis on producing the correct set for learning and retention before the presentation of the material. It has been deter-
372
THOMAS
H.
BUDZYNSKI
mined by the Russians that retention of the material is optimized if the presentation takes place during the first 30-40 minutes of sleep. Rubin (1970) has noted that the common denominator among successful sleep-learning studies is that "superficial sleep" (Stages 1 and 2) is the psychophysiological background for maximum receptivity. 6. Information Processing at Low-Arousal Levels
If an individual can be maintained in a twilight-state, reverie, sleep-onset condition, or Stage 1 or 2 sleep, it appears that he is capable of assimilating information, especially if he has been prepared to learn and remember the material. However, there is little doubt that information processing in such a state is quite different from that in the alert, waking state. Just as internally generated information is processed through the strange illogic of the twilight state into bizarre, disjointed hypnagogic imagery, so too does externally presented material undergo transformation under these conditions. Lasaga and Lasaga (1973) have noted the blurring of perception that occurs with regard to numerals presented during deepening sleep. One of the leading Russian researchers, Svyadoshch (1968), commenting on sleep-learning, stated: Speech assimilated during sleep, in contrast to that assimilated during waking state is not subjected during assimilation to the critical processing, and is experienced on awakening as a thought of which the source remained outside consciousness: to some extent therefore, it is as if it belonged to an alien personality. (p. 112)
If the perception of external information is blurred and distorted by the twilight state, is there a limit to the complexity of information that can be assimilated? Lasaga and Lasaga (1973) concluded from their study that perceptual distortions increase with deepening sleep, thus making the assimilation of complex verbal material less likely. For this reason is would seem to be important to maintain the learner in a drowsy or light-sleep stage or perhaps to vary the complexity of the presented material in accordance with his arousal level; that is, the complexity would decrease as the learner drifted from drowsy to deeper sleep.
BIOFEEDBACK AND THE TWILIGHT STATES OF CONSCIOUSNESS
373
D. The Production of Low Arousal through Biofeedback Can the ordinary individual take advantage of twilight-state learning? That is, can he produce a drowsy state at will and remain "there" without passing into deeper sleep or alerting? Probably not. However, research conducted in our laboratory at the University of Colorado Medical Center over the past 6 years has resulted in the development of training techniques for the implementation of lowarousal states. Through the use of electromyographic (EMG) or muscle-tension feedback, we were able to teach subjects to reach very low levels of skeletal muscle tension over selected muscle sites (Budzynski and Stoyva, 1969, 1973). We found evidence of decreased cortical and autonomic arousal as subjects focused on relaxing their muscles with EMG biofeedback (Budzynski, 1969).
1. Is It Necessary to Decrease EMC Levels in Order for Theta Rhythms to Appear? Apparently theta rhythms are not likely to appear in those individuals with high forehead tension. In a biofeedback study (Sittenfeld, Budzynski, and Stoyva, 1972) we divided subjects into high and low forehead-EMG groups. Half of each of these two groups received eight sessions of training in theta EEG production with theta biofeedback. The other half of each high- and low-EMG group received one session of forearm extensor EMG feedback and three sessions of forehead (frontalis) feedback, followed by four sessions of theta EEG feedback. Those subjects with initially low forehead-tension levels learned to increase their theta amounts whether they received only theta feedback or when they received EMG followed by theta training. Those subjects with initially high forehead-EMG levels also learned to increase theta, but only if they first received EMG training to decrease tension levels. Those subjects with initially high foreheadEMG did not increase their theta production with theta-feedback training alone. It was concluded that one must learn to relax his muscles, particularly those of his face and head, before he can begin to approach the twilight state as evidenced by an increase of 4-7-Hz rhythms. This is consistent with subjective reports from our other studies, which indicated that the appearance of the theta rhythm and
374
THOMAS
H.
BUDZYNSKI
feelings of drowsiness are associated with the deep relaxation of the forehead muscles. Other researchers have employed procedures such as autogenic exercises (Green et al., 1973; Svyadoshch, 1968) and white noise and ganzfeld fields (Bertini, Lewis, and Witkin, 1969) to facilitate the production of a twilight state. Whatever the technique used, the goal appears to be the development and maintenance of a relaxed, drowsy, possibly even light-sleep state during which information is made available to the subject in auditory form. 2. Why Is Such a Twilight State Necessary for Some Kinds of Learning?
The evidence would seem to point to the existence of phenomena known as defense mechanisms, resistances, and mental blocks, which can distort, hinder, or prevent the assimilation of certain types of information presented during the alert, waking state. The research results would also suggest that these phenomena are rendered less effective as arousal level declines. However, as arousal level decreases toward deeper sleep, the ability to assimilate information changes quantitatively. In the deeper stages of sleep very little information can be absorbed and it must be of a concrete rather than an abstract nature. There is, in addition, some scattered evidence that the presentation during the waking state of material which tends to activate these phenomena only strengthens the resistance to the information.
E. A Twilight-State Biofeedback System On the basis of the conclusions and assumptions referred to above, we have developed a biofeedback system which allows the subject to maintain himself in a twilight state while tape-recorded information is presented. The functional logiC of this system is as follows: 1. A relatively constant auditory stimulus masks background
noise and enhances habituation and a lowering of arousal level. Bertini et al. (1969) have demonstrated that white noise
BIOFEEDBACK AND THE TWILIGHT STATES OF CONSCIOUSNESS
375
facilitates this process and the subsequent appearance of theta EEG rhythms. Therefore this biofeedback unit provides a "pink" noise 3 background, the volume of which can be adjusted to mask varying levels of ambient noise. 2. The material should be presented when the subject is in the twilight state and not when he is deep asleep or fully alert. The system allows the subject to hear the tape-recorded material only when he is producing brain waves in a selected frequency band. The band width is adjustable on the upper end to 9, 8, or 7 Hz, i.e., if the subject's brain waves are at or less than the selected cutoff frequency, he does not hear the material but rather the same voice suggesting that he will relax, get drowsy, and learn and remember the material to be presented. Thus the subject at all times hears a voice from one or the other channel of a stereo tape-recorder. 1. What if the Subject Shifts toward Deeper Sleep?
The design of the system is such that the subject's brain-wave frequencies decrease and their amplitudes increase, the volume of the mesage increases. This volume increase in almost all cases will prevent the subject from falling into deeper sleep. Thus the design of the system is such that if the subject so desires, he can easily maintain himself in a twilight state and absorb the learning material while doing so. On the other hand, he can, if he desires, arouse himself and thus tum off the presentation. The selectable bandpass should allow us to determine optimal EEG-frequency parameters for twilight-state learning. It is possible, for example, that material likely to be defended against on some higher level of awareness may have to be presented at EEG frequencies lower than those for less controversial material. The dual-track recorder and its associated processing circuitry should allow us to study two different levels of information complexity. Thus a more concrete message can be presented when lower EEG frequencies are present, and a more complex message could be heard in the event of a more activated brain-wave state. 3
Pink noise is essentially white noise filtered to produce a more pleasant sound.
376
THOMAS
H.
BUDZYNSKI
2. Pilot Research with the Twilight-State System Almost 2 years of refinement have resulted in an electronic means of allowing individuals to maintain themselves in a nondrug, lowarousal ASC. If Felipe's results are to be believed, this state should facilitate attitude change. If Lozanov's results and those of the Russian sleep-learning studies are real phenomena, then such a state may be useful for educational purposes. Adams's work suggests its use in psychotherapy. Although there is the possibility of confounding placebo effects in the reporting of case studies, some reports may be of interest in demonstrating the potential use of this technique in psychotherapy. In one instance a client who had found little success in some 3 years of traditional psychotherapy reported dramatic improvement following five sessions of twilight learning. First, the increase in theta-associated hypnagogic imagery vividly brought out the early roots of an authority problem stemming from an inability to say no to an overbearing father. Subsequently a tape was prepared using the client's own voice. The tape dealt with the relation of his present problem (not being able to say no to unreasonable requests) to the early experience with his father. The tape also reminded him that he had the right and the ability to say no when he wished to do so. Following the training, the client found that he could easily tum down unreasonable requests and in fact felt a deep sense of accomplishment and relief in doing so. A second client was going through the anxiety associated with a divorce. His wife had requested the divorce, yet was threatening suicide and thus driving him back to a drinking problem he formerly had mastered. A tape was designed to aid in altering the irrational thoughts that were contributing to the client's severe guilt feelings. He heard the tape under twilight-learning conditions on three occasions. Subsequently he reported less anxiety and guilt, a decrease in earlymorning insomnia, a feeling of a burden being lifted, and a cessation of drinking behavior. A graduate student had already failed a language exam and found himself so anxious about taking it a second time that he was unable to study at all. A Spanish-English tape with instructions to learn and remember was played back to the student under twilight-learning conditions four times a week for 3 weeks. He easily passed the exam the second time.
BIOFEEDBACK AND THE TWILIGHT STATES OF CONSCIOUSNESS
377
Another student with severe test anxiety responded with strong GSR activity to certain words from a list read to him. A tape was designed by the student and the therapist and recorded in the student's voice. The message suggested, in effect, that the words and the situations they represented would no longer arouse anxiety. He listened to the tape twice in the twilight state. When the word list was presented again, the GSR did not respond to the key words. The student reported a lessening of test anxiety. Educational Material. In order to test for later retention of educational material presented during a twilight state, eight individuals were allowed to hear, for a total of 2 hours of "theta time," a repeated passage about a little-known topic: tide-pool life. Before-after comprehension-testing revealed that only one person had increased his score significantly as a result of the training. However, it was noted that this individual had remarked after training that "it was just like my radio." When questioned about this, he stated that he daily awakened to a clock radio tuned to a news station. He typically would remain in bed for 15 minutes or so after the alarm went off before rising. He further stated that his friends were always amazed at his general knowledge of current events, although he couldn't recall many specifics. On the basis of this one subject's report, we felt that a "learning to learn" may be necessary for the absorption and/or recall of material from this state. Therefore the other seven subjects were brought back for additional training with the same material. Comprehension tests were given before and after each new hour of twilight learning. After the third hour two other subjects improved significantly on their scores. Following the fourth hour two more improved, and in the fifth hour a sixth subject improved his score from an average of 44 to one of 72. Two of the subjects did not improve even after 5 hours of training. In one case the subject oscillated in and out of theta rather rapidly, so that the material was quite disconnected. The second failure managed to slip into deeper sleep stages (Stages 3 and 4) as evidenced by delta rhythms (1-3 Hz), which routinely appaeared in his EEG. This initial study, carried out in association with Kirk Peffer at the University of Colorado Medical Center, does suggest a learning-tolearn factor. However, an alternative explanation is that the subjects were able to increment bits of information by momentary increases in arousal before the system turned off the tape recorder. Although a
378
THOMAS
H.
BUDZYNSKI
possibility, the increments would be rather small since the system shuts off in .5-1.0 seconds when alpha or beta frequencies appear. Nevertheless a second study, now underway, focuses on the presentation of foreign words and English equivalents. Different word lists are presented at each interval of training so that the learning-to-learn effect can be tested more precisely. However, if the yield (the amount of recall or recognition) is small with low-arousal learning compared with fully alert learning, why go to all the trouble? Cooper and Hoskovec (1972), for example, found that highly hypnotizable subjects with a preparatory set to learn while sleeping recalled only 28% of the word pairs presented during Stage 1 REM as compared to alert-learning recall of 92%. Other investigators have demonstrated that registration and acquisition of auditory material can occur during sleep, since the subjects can respond on subsequent nights without repetition of the suggestion, yet without any apparent waking memory of the material (Evans, Gustafson, O'Connell, Orne, and Shor, 1966). Perhaps the problem, then, is one of retrieval in the waking state. If this difficulty can be overcome, then low-arousal learning may be very useful for these reasons: 1. It would appear that the learning is of a holistic, intuitive sort which is not forgotten as easily as is the typical, fully alert rotelearning we usually acquire. 2. The material is learned during a relaxed stale rather than one of tension and anxiety, as is often the case in competitive educational situations. 3. Related to point 2 is the possibility of the absorption of learning material which has formerly been "mentally blocked," i.e., subject material which the student is convinced he cannot learn. 4. Positive attitude-change with informed consent may be effected. The moral implications of this application are, of course, an important consideration. 5. The retrieval of hypnagogic or primary-process information constitutes one aspect of the exploration of the less-conscious mental processes, with the other being the input of information intended to clarify or to alter positively the complex under study. 6. A final, and admittedly highly speculative, possibility is that
BIOFEEDBACK AND THE TWILIGHT STATES OF CONSCIOUSNESS
379
physical processes can be altered more effectively through proper suggestion presented during a twilight state.
3. A Brief Summary of What Has Been Discussed Thus Far The self-maintenance of a sustained twilight state would appear to be useful for the generation of potentially creative material. This material also may be of use to the psychotherapist who is interested in primary-process or unconscious-process information. On the other hand, the twilight state can be a vehicle for the presentation of educational, attitude-change, or self-growth material. Although low-arousal states can be produced by many methods, biofeedback allows a precision and a consistency which are lacking in the other techniques. Even tense individuals can be taught through EMG feedback to relax to the point at which theta brain-wave rhythms begin to appear. Such training may be a necessary precursor to twilight-state learning in some individuals. The precise selection of responses (e.g., certain brain-wave frequencies) to be fed back for purposes of studying the most efficient "learning window" can be achieved with biofeedback. In general, if the physiological correlates of an ASC can be defined, then biofeedback can be used to shape and maintain the ASC.
II.
FUTURE CONSIDERATIONS
The early research with alpha brain-wave feedback was at first thought to be the quick route to Nirvana because meditating Zen monks were said to produce an abundance of alpha. However, many "alpha hours" later it became apparent that getting there with alpha happens to very few. Many individuals have an abundance of alpha in the normal state with eyes closed. Still others show no alpha under any circumstances. Those who can benefit most from alpha feedback training are individuals with roughly 10--40% alpha in their EEG prior to training. With proper feedback training these people can learn to produce a higher percentage of alpha, quite possibly accompanied by feelings associated with a decreasing arousal level. Thus alpha training can be a useful first stage in the shaping of a low-arousal pattern. We
380
THOMAS
H.
BUDZYNSKI
have preferred to use frontalis EMG feedback as a precursor to theta training if a subject has high muscle tension initially. When the subject approaches a twilight state, heralded by the appearance of theta rhythms, he can begin to receive theta feedback. If the goal of the production of a twilight state is the retrieval of hypnagogic material for use in creativity or psychotherapy or self-growth, then the theta feedback could assume anyone of several forms, for example, a tone or a modulated carrier. However, the inputting or presentation of material during a twilight state requires a precise synchronization of theta rhythms and material. Additionally, the feedback system must guard against deeper sleep, which often is a natural result of a twilight state. We have found it useful in the early phase of twilight training to present only music (no lyrics) to the subject. As a second phase we are now experimenting with the presentation of verbal material mixed with background music, in fact, the identical music that was used in the first phase. It appears that the absorption of verbal material is facilitated by the music background. This is consistent with the view that low-arousal learning may be localized in the nondominant hemisphere.
A. Is Twilight Learning Minor-Hemisphere Learning? Recent research findings from the area of brain lateralization (Galin, 1974) and the beautifully integrated discussion of consciousness by Ornstein (1972) provide an exciting context in which to place twilight learning. Consider that many of the descriptions of the recall and recognition of material presented during low-arousal states are identical to those attributed to minor-hemisphere learning. In both cases concrete rather than abstract information is processed more effectively. Moreover when the information is coded in music or rhythm, it is more easily absorbed, as is also said to be the case with minor-hemisphere learning. Finally, there is the observation that following a twilight or low-arousal learning-session, subjects very often cannot verbalize what they have supposedly learned, yet they do well on recognition-comprehension tests. And in split-brain studies minor-hemisphere learning is often best tested with nonverbal recognition-techniques simply because the minor hemisphere has little
BIOFEEDBACK AND THE TWILIGHT STATES OF CONSCIOUSNESS
381
control over verbal-output functions. 4 That is the domain of the major hemisphere. Perhaps the decrease in critical, analytical, logicallinear-functioning that occurs with a lowering arousal level is the gradual, functional disabling of the major hemisphere. We might also speculate that the minor hemisphere is not as quickly disabled as is the major with decreasing arousal. Thus at the point where the major hemisphere ceases to process effectively incoming (or outgoing) information, the minor hemisphere can still do so in its own holistic manner. At a still lower point on the arousal continuum (deeper sleep stages), even the minor hemisphere's efficiency drops off sharply. However, it is the region between these two points, or what we might call the "window," in which the major hemisphere would appear to relinquish control, thus allowing the uncritical absorption by the minor hemisphere of external material or the emergence of internally generated material. But there is still a communication problem here.
B. A Language for the Minor Hemisphere If the verbal hemisphere is functionally disabled, how is the material to be presented (to a primarily nonverbal hemisphere), or retrieved, in the case of hypnagogic stuff? The accumulating evidence indicates that the minor hemisphere may be able to absorb more verbal material than was previously thought, although the extent of learning may have to be tested by nonverbal means. Even so, it seems to be a fact that if the verbal material is relatively more concrete than abstract and if it is coded in music or poetry, it will be more readily absorbed. Therefore what we hope to evolve is a minor-hemisphere language which can be used in twilight-learning situations. Since the minor hemisphere is specialized for visual-spatial processing, it is very possible that this language will include visual forms.
C. Retrieval Difficulties Retrieval of emergent, ongoing material generated during a twilight state is a troublesome dilemma. If the subject is awakened and 4
See Schwartz (1974) for a description of research relating to lateralization testing in intact humans.
382
THOMAS
H.
BUDZYNSKI
asked to report, he may have difficulty slipping back into the state. We have found that if subjects are asked to give only one word or a short phrase to describe the imagery, they can readily return to the twilight state. Later, upon awakening, subjects usually can freeassociate and flesh out the image or images. Given the ability to produce and maintain an altered state which optimizes creative associations, the problem for the future will be the development of efficient retrieval techniques.
D. Cognitive Balance There is a growing feeling that perhaps our technologically oriented Western culture has selectively favored one cognitive mode over the other (Fischer and Rhead, 1974; Ingrasci and Kimura-Bucholtz, 1974). Thus the emphasis on the major, dominant, or "Aristotelian" hemisphere, with its analytical, rational, sequence-perceiving processes. Largely ignored in our culture has been the minor, nondominant, or "Platonic" hemisphere, with its nonverbal, synthesisoriented, intuitive functions. As Fischer and Rhead have noted, the problem lies in the implicit belief that the rational and unlimited technological conquest of nature will necessarily lead to a better life. The corollary to such a belief is that any non technological approach to the good life is reactionary, irrational, and primitive. However, the overemphasis on the scientific, rational belief-system has, in fact, resulted in a rebound effect in the form of a swing toward the minorhemisphere mode of thinking. This is evidenced by the growing interest in Eastern meditation, the occult, astrology, and parapsychology and a declining interest in organized science and religion. Perhaps men and women in our culture can benefit by training in the use of minor-hemisphere processes. Perhaps feelings, emotions, intuitions, and creative ideas can be brought to awareness and then integrated into a rational approach. Perhaps, too, biofeedback may be one of the ways for a technological culture to acquire this balance. REFERENCES
A case utilizing sensory deprivation procedures. In L. P. ULLMAN AND L. (Eds.), Case studies in behavior modification. New York: Holt, Rinehart & Winston, 1965, pp. 164-170.
ADAMS,
H. B.
KRASNER
BIOFEEDBACK AND THE TWILIGHT STATES OF CONSCIOUSNESS
383
BARBER, T. X. Experiments in hypnosis. Scientific American, 1957,196, 54--61. BERTINI, M., LEWIS, H. B., AND WITKIN, H. A. Some preliminary observations with an experimental procedure for the study of hypnagogic and related phenomena. In c. T. TART (Ed.), Altered states of consciousness. New York: Wiley, 1969, pp. 93-115. BRUCE, D. J., EVANS, C. R., FENWICK, P. B., AND SPENCER, V. Effect of presenting novel verbal material during slow-wave sleep. Nature, 1970,225, 873, BUDZYNSKI, T. Feedback-induced muscle relaxation and activation level. Unpublished doctoral dissertation. University of Colorado, 1969. BUDZYNSKI, T. Some applications of biofeedback-produced twilight states. Fields within Fields . .. within Fields, 1972,5, 105-114. Republished in D. SHAPIRO, T. X. BARBER, L. V. 01 CARA, J. KAMIYA, N. E. MILLER, AND J. STOYVA (Eds.), Biofeedback and selfcontrol, 1972. Chicago: AIdine-Atherton, 1973. BUDZYNSKI, T., AND PEFFER, K. Twilight-state learning: A biofeedback approach to creativity and attitude change. Paper presented at the Transformations of Consciousness Conference sponsored by the R. M. Bucke Memorial Society and the Department of Psychiatry, McGill University, Montreal, Canada, 1973. BUDZYNSKI, T., AND STOYVA, J. An instrument for producing deep muscle relaxation by means of analog information feedback. Journal of Applied Behavior Analysis, 1969,2, 231-237. BUDZYNSKI, T., AND STOYVA, J. A biofeedback technique for teaching voluntary relaxation of the masseter. Journal of Dental Research, 1973,52, 116-119. COOPER, L. M., AND HOSKOVEC, J. Hypnotic suggestions for learning during Stage 1 REM sleep. American Journal of Clinical Hypnosis, 1972,15, 102-111. EVANS, F. J., GUSTAFSON, L. A., O'CONNELL, D. N., ORNE, M. T., AND SHOR, R. E. Response during sleep with intervening waking amnesia. Science, 1966, 152, 666667. FELIPE, A. Attitude change during interrupted sleep. Unpublished doctoral dissertation. Yale University, 1965. FISCHER, R., AND RHEAD, J. The logical and the intuititive. Main Currents, 1974,31,50-
54.
FOULKES, D., AND VOGEL, G. Mental activity at sleep-onset. Journal of Abnormal Psychology, 1965,70, 231-243. GALIN, D. Implications for psychiatry of left and right cerebral specialization: A neurophysiological context for unconconscious processes. Archives of General Psychiatry, 1974,31, 572-583. GIBBY, R. G., AND ADAMS, H. B. Receptiveness of psychiatric patients to verbal communication. Archives of General Psychiatry, 1961,5, 366-370. GREEN, E., GREEN, A., AND WALTERS, D. Voluntary control of internal states: Psychological and physiological. Journal of Transpersonal Psychology, 1970,1, 1-26. GREEN, A., GREEN, E., AND WALTERS, D. Psychophysiological training for creativity. Paper presented at the meeting of the American Psychological Association, Washington, D.C., 1971. GREEN, A., GREEN, E., AND WALTERS, D. Brainwave training, imagery, creativity, and integrative experiences. Unpublished manuscript, 1973. HERON, W. Cognitive and physiological effects of perceptual isolation. In P. SOLOMON, P. KUBZANSKY, P. LEIDERMAN, J. MENDELSON, R. TRUMBULL, AND D. WEXLER (Eds.), Sensory "deprivation. Cambridge, Mass.: Harvard University Press, 1961. INGRASCI, R., AND KIMURA-BUCHOL"IZ, M. Altering consciousness: A new dimension to public health. Fields within Fields . .. within Fields, 1974, 13, 38-48. JAMES, W. The principles of psychology (Vol. 1). New York: Dover Publications, 1950. KOESTLER, A. The act of creation. New York: MacMillan, 1964.
384
THOMAS
H.
BUDZYNSKI
KUBIE, L. Neurotic distortions of the creative process. Lawrence: University of Kansas Press, 1958. KUBIE, L., AND MARGOLIN, S. A physiological method for the induction of states of partial sleep, and securing free association and early memories of such states. Transactions of the American Neurological Associations, 1942, pp. 136-139. LASAGA, J., AND LASAGA, A. Sleep learning and progressive blurring of perception during sleep. Perceptual and Motor Skills, 1973,37, 51-62. LEVY, C. M., COOLIDGE, F., AND STAAB, L. Paired-associate learning during EEG defined sleep: A preliminary study. Australian Journal of Psychology, 1972,24, 219-225. LINDSLEY, O. R. Operant behavior during sleep as a measure of depth of sleep. Science, 1957, 126, 129~1291. MAcKINNON, D. W. Creativity and transliminal experiences. Address given at the American Psychological Association Convention, Los Angeles, 1964. MAGOUN, H. Advances in brain research with implications for learning. In J. KAGAN (Ed.), On the biology of learning. New York: Harcourt, Brace & World, 1969, pp. 169190. McKELLAR, P., AND SIMPSON, L. Types of synaesthesia. Journal of Mental Science, 1955, 101, 141-147. Moscu, K., AND VRANCEANU, M. Quelques resultats concernant I'action differentielle des mots affectogenes et nonaffectogenes pendant Ie sommeil nature!. Paper presented at the International Psychosomatic Week, Rome, September 11-12, 1967. In M. BERTINI (Ed.), Psicofisciologia del Sonna e del Sogno. Milan: Editrice Vita e Pensiero, 1970. MYERS, T., MURPHY, D., AND SMITH, S. The effect of sensory deprivation and social isolation on self-exposure to propaganda and attitude change. American Psychologist, 1963,18, 440 (abstract). ORNSTEIN, R. The psychology of consciousness. San Francisco: Freeman, 1972. OSTRANDER, S., AND SCHROEDER, L. Psychic discoveries behind the Iron Curtain. Englewood Cliffs, N.J.: Prentice Hall, 1970. PHILIPOV, E. Personal communication, 1975. RUBIN, F. (Ed.), Current research in hypnopaedia. London: MacDonald, 1968. RUBIN, F. Leaming and sleep. Nature, 1970,226, 447. RUGG, H. Imagination. New York: Harper & Row, 1963. SCHWARTZ, G. E. Hemispheric asymmetry and emotion: Bilateral EEG and lateral eye movements. Paper presented at the American Psychological Association Convention, New Orleans, 1974. SIMON, c., AND EMMONS, W. Learning during sleep? Psychological Bulletin, 1955, 52, 328-342. SIMON, c., AND EMMONS, W. Responses to material presented during various levels of sleep. Journal of Experimental Psychology, 1956,51, 89-97. SITTENFELD, P., BUDZYNSKI, T., AND STOYVA, J. Feedback control of the EEG theta rhythm. Paper presented at the American Psychological Association Convention, Honolulu, 1972. STOYVA, J. Biofeedback techniques and the conditions for halludnatory activity. In F. J. MCGUIGAN AND R. A. SCHOONOVER (Eds.), The psychophysiology of thinking. New York: Academic Press, 1973, pp. 387-405. SUEDFELD, P. Attitude manipulation in restricted environments. Journal of Abnormal Psychology, 1964,68, 242-247. SVYADOSHCH, A. The assimilation and memorization of speech during natural sleep. In
BIOFEEDBACK AND THE TWILIGHT STATES OF CONSCIOUSNESS
385
F. RUBIN (Ed.), Current research in hypnopaedia. London: MacDonald, 1968, pp. 91117. SVYADOSHCH, A. Perception and memory of speech during natural sleep. In Sleep learning in the U.S.S.R., ATD report 68-91-108-6, Aerospace Technology Division, Library of Congress, February 7, 1969. WILLIAMS, H. L., MORLOCK, M. c., AND MORLOCK, J. V. Instrumental behavior during sleep. Psychophysiology, 1966,2, 208-216.
Author Index
Abe, M., 107 Adam, G., 26 Adametz, J. H., 39 Adams, H. B., 368, 376 Adamson, 1.,80 Adelson, J., 334, 335 Adey, W. R., 26 Ahn, H., 17, 20 Ahrens, J. B., 350 Aiken, 1. R., 120 Akesson, C. A., 123 Akpinar, 5., 182, 191, 203 Albright, G. A., 103 Aldridge, V. J., 16 Allport, G., 245 Anch, A. M., 108, 109 Anderson, N. 5., 92 Anderson, P., 110 Anderson, W. 1., 181 Antrobus, J., 316 Antrobus, J. 5., 316 Arkin, A. M., 156 Armington, J. c., 105 Arthur, A., 294 Ashby, W. R., 75, 86 Ashern, B., 226 Asratyan, E. A., 15 Attneave, F., 331 Austin, M. D., 321 Baekeland, F., 333 Bagchi, B. K., 106 Bagshaw, M. H., 60, 65, 67, 68, 71, 92 Bailey, C. J., 66 Bailey, P., 55, 80 Bakan, D., 335 Baker, J., 156, 232
Baldridge, B. A., 334 Bancroft, J. H., 281 Bandura, A., 237 Banford, S. A., 181 Barber, B., 318 Barber, T. X., 179, 181, 183,186,187,250,367 Bardwich, J., 337 Barker, W., 186 Barlow, J. 5., 15 Barolin, G. 5., 192 Bartlet, F. c., 331 Bartlett, F., 7, 19, 20, 24, 25,28 Bartolucci, G., 109 Bateman, D. E., 298 Beagley, H. A., 107 Beck, A., 223, 229, 230, 232, 242, 253 Beck, c., 109 Beck, E., 112, 114, 116 Beck, E. c., 192 Becker, M., 348 Beebe-Center, J. G., 83 Beecher, H. K., 156, 160 Begleiter, H., 15 Beier, E. G., 192 Bern,S., 228, 271 Bendfeldt, F., 156 Ben-Horin, P., 349 Benjamin, F. B., 103 Benzies, 5., 68 Berger, R. J., 347 Bergin, A., 226, 237 Bergman, J. 5.,292 Berlyne, D. E., 147 Berman, A. J., 77 Bernheim, H., 176
Bernick, N., 314 Bernstein, N., 77, 79 Berstein, D. A., 274, 277, 281 Bertini, M., 327, 333, 374 Besson, J. M., 167 Best, A., 223 Beverly, K. l., 106 Black, A. H., 270 Blackwood, R., 228 Blanchard, E. B., 292 Blitz, B., 102 Blum, G. 5., 181 Blum, J. 5.,60 Bonime, W., 330, 347 Bonvallet, M., 111 Bootzin, R. R., 283 Borge, G. F., 104, 111 Borkovec, T. D., 275, 277, 278, 279, 280, 281, 282, 283,285,286, 288, 293, 294, 295, 296 Borowitz, G., 314 Boulogouris, J., 226, 294 Boutilier, J., 226 Bowers, c., 107 Boxerman, 1. A., 108 Braff, D., 229 Bragdon, H. R., 122 Braid, J., 175 Brandsma, J. M., 156 Breger, 1., 329, 333, 347 Brendler, S. J., 55 Brenman, M., 148, 181, 183 Brennan, E. P., 183 Brentano, F., 52 Bresnitz, 5., 294
387
388 Brillovin, L., 86 Broadbent, D., 153 Broadhurst, A., 103 Brobeck, J. R., 75 Brook,251 Bruce, D. J., 370 Bryan, M. E., 124 Bryce, D. P., 107 Buchsbaum, M., 104, 105, 106, 108, 111, 114, 116, 117, 118, 120, 121, 122, 125,128 Buchwald, J. 5.,26 Bucy, P., 54, 59 Budzynski, T., 362, 367, 373 Bunker, J. P., 155 Bures, J., 27 Buresova,O., Burgwin, 5., 186 Butier, R. A., 107, 124 Butter, C. M., 90 Byrne, D., 292, 341 Callaway, E., 104 Calma, I., 110 Calverley, D. 5., 180, 181, 183, 186 Cameron, R., 223, 226, 234,235,242,252,254 Cannon, W., 54, 83, 157 Caputo, J. A., 295 Carlson, R., 335 Carnegie, D., 251 Carpenter, R. 5., 13, 15 Carpenter, W., 105 Carranza, M. B., 7 Carrington, P., 334, 347 Cartwright, R. 0.,314, 329,333,334,347,350, 351 Casey, K. L., 156, 158 Cautela, J., 270 Cavonius, C. R., 124 Cepeda, G. V., 7 Chapman, R. M., 122, 124 Charcot, J. M., 175 Chase, M. H., 332, 345, 349 Cheek, D., 155
AUTHOR INDEX
Chernick, D. A., 349 Chertok, L., 186 Chesler, P., 7 Chesney, G. L., 23 Chow, K. L., 39, 60,91 Christiansen, R. L., 117 Clark, W. c., 118 Clarke, P. R. P., 349 Cleckley, H. M., 156 Ciemes, S. R., 143 Ciouston, R. A., 73 Ciynes, M., 106, 192 Coe, W. c., 148 Cohen, D. B., 314, 316, 317,318,319,322,324, 329,333,334,335,337, 338, 343, 351, 352 Collins, W., 102 Cook, L., 109 Coolidge, P., 370 Cooper, L. M., 155, 181, 183,190,196,198,378 Cooper, R., 16 Copeman, c., 233 Coppock, H. W., 68 Coming, W. c., 3 Coue, E.; 251, 252 Cowan, W. M., 80 Cox, D., 314, 316, 318, 329,333,334, 338, 351, 352 Cox, V. c., 75 Craighead, W. E., 277 Cristian, C. N., 26 Crow, H. H., 16 Cruikshank, R. M., 105 Dallet, J., 333, 347 Davis, G., 55 Davis, H., 107 Davis, L. W., 176 Davison, G. c., 274 Dawson, W. E., 123 Debecker, J., 109 Decenteceo, E., 230 de Charms, R., 237 Deci, E., 237 Deinstbier, R. A., 287 Delgado, J. M., 55 Dell, P., 111
Dember, W., 242 Dement, W. c., 316, 346, 347,349 Denny, D., 228 Desmendt, J. E., 109 Deutsh, D., 153 Deutsch, J. A., 153 De Voe, R. G., 105 Dewan, E. M., 349 Diamond, M. J., 181, 182, 204, 205, 207 Di Loreto, A., 232 Dinnerstein, A. J., 102 Dinnerstein, D., 124 Disterhoft, J. P., 26 Doane, B., 26 Dollard, J., 226 Domhoff, B., 328 Donchin, E., 105 Donner, L., 226 Dru, 0.,40 Dumenko, V. N., 12 Dunsmore, R. H., 58 Dustman, R. E. 112, 114, 116,192 Dyal, J. A., 3 Dykman, R. A., 270 Dynes, J. B., 186 D'Zurilla, T., 232, 242 Eason, R. G., 120 Eccles, J. c., 86, 110 Edmonston, W. E., 186 Ekman, G., 123 Ellen, P., 12 Ellis, A., 223, 230, 231232, 239, 242, 270 Elkonine, D., 249 Ellman,S., 350 Emmons, W., 370 Empson, J. A. c., 349 Emurian, c., 209 Engel, J., 26 Engeland, W., 123 Engstrom, D. R., 182, 190, 191, 195, 198, 201, 203, 208, 210 Ephron, H. 5.,347 Epstein, J. A. 55 Erikson, E. E., 329
389
AUTHOR INDEX
Ervin, F. R., 167 Evans, C. R., 370 Evans, F. J., 186, 190, 191, 192, 193, 203, 204, 343, 378 Evarts, E. V., 79 Eysenck, H. J., 269, 286, 294,295 Faw, V., 183 Fedoravicus, A., 231 Felipe, A., 367 Fenwick, P. B., 370 Fibiger, H. c., 73 Finkenzeller, P., 15 Fishbein, W., 348 Fischer, R., 382 Fisher, c., 316 Fiss, H., 319, 329, 350 Fitts, P. M., 92 Flannery, R., 270 Fleishman, D. J., 295 Ford, W. L., 186 Foster, R., 102 Foulkes, D., 316, 317, 328, 333, 334, 350, 362 Fox, S. 5., 26 Frank, G. 5.,155 Frank, J., 223, 240 Franzen, 0., 110 Freedman, J., 228 Freeman, W., 158 Freud,S., 128,313,331, 347 Freyberg, J., 233 Friedlander, J. W., 176 Friedman, H. J., 280 Fromme, E., 330, 347 Frommer, C. P., 91 Fruhstorfer, H., 15 Fry, A., 105 Fuhrer, M., 185 Fukuma, E., 335 Fulton, J. F., 55 Gagne, R., 248, 249 Galambos, R., 12, 15,91 Galanter, E. H., 51, 52, 84, 86, 147 Galbraith, G. c., 190, 191, 295
GaIim, D., 320, 321, 380 Gantt, W. H., 270 Garner, W. R., 91, 92 Gastaut, H., 12 Gath, D. H., 281 Gawpp, L. A., 283 Gazzaniga, M., 155 Gelder, M. G., 281 Genest, M., 254 Gibbey, R. G., 368 Gill, M. M., 148, 181, 183 Gillin, J. c., 107 Gilmore, B., 231 Glascow, R. E., 283, 285, 286 Glass, L. B., 179 Glivenko, E. V., 12 Goff, W. R., 109 Goldfried, M., 230, 232, 235,242 Goldstein, A., 166, 167 Goldstein, A. P., 280, 281 Goldstein, R., 110 Goodenough, D. R., 314, 317, 324, 337, 351 Goodman, J., 118, 223, 227, 228, 233 Goodman, W. 5., 107 Goodwin, F., 106, 111 Gottlieb, A. E., 80 Gough, P., 183 Graham, J. T., 107 Gray, J. A., 125, 126 Green, A., 362, 363, 364, 365 Green, A. M., 209 Green, D., 233 Green, E., 362, 364, 365, 366,374 Green, E. E.. 209, 210 Greenberg, R., 348, 349 Griffin, R. B., 104, 112 Grindberg-Zylberbaum, J., 7, 21 Guilbaud, G., 107 Guirao, M., 123 Gur, R. c., 155 Gur, R. E., 155 Gustafson, L. A., 343, 378
Hagbarth, K. E., 110 Halas, E. 5., 26 Hall, C. 5.,328,331,332 Hall, R. A., 104, 112 Halliday, A. M., 192 Hallsten, L., 123 Hamasaki, D. I., 106 Hance, A. J., 15 Hanel, J., 233 Hanley, E., 270 Hart, B., 156 Hart, J. T., 182, 189, 191, 195 Harter, M. R., 120 Hartig, M., 228 Hartley, L. R., 120 Hartman, E., 324, 327, 347, 348, 352 Hartman, H., 148 Hartnett, J., 191 Harvey, E. N., 186 Hastey, J. M., 156 Hauri, P., 327, 333 Hearst, E., 67 Hebb, D.O., 147 Heckhausen, H., 233 Helvey, W. M., 103 Hendrick, c., 337 Hendrickson, J. L., 80 Henkin, R. I., 117, 125 Henry, G. B., 107 Heron, W., 368 Hetherington, M., 228 Hilgard, E. R., 103, 144, 147, 160, 161, 176, 177, 178,179,180,181,183 Hilgard, J. R., 184 Hillman, P., 109 Hillyard, S. A., 15 Hilz, R., 124 Hirano, T., 26 Hirsch, R., 26 Hirsh, S. K., 107 Hobart, G., 186 Holland, T., 102 Holzman, P. 5., 102, 103 Homme, L., 226 Hooten, T. F., 26 Hopkins, H. K., 104, 112 Hori, Y., 26
390
Hoskovec, J., 378 Hosman, B., 123 Hugelin, A., 111 Hughes, W., 346, 347 Hull, C. L., 147 Hull, R. c., 105, 106 Hunter, J., 329 Husband, R. W., 176 Husser!, E., 52 Ikuta, T., 109 Ill'yanok, V. A., 106 Ingram, G., 226 Ingrasci, R., 382 Itil, T. M., 182, 191, 203 Jaeger, M., 232 Jahoda, G. 245 James, W., 53, 54, 56, 57, 73, 74, 81, 82, 88, 94, 144,148,149,156,362 Jameson, D. H., 156 Janet, P., 138, 139, 140, 144,148 Janis, 1., 235, 249 Jasper, H. H., 26 Johansson, G., 79 John, E. R., 7, 10, 11, 12, 13,14,15,16,17,19,20, 21,24,27,28,32,33,35, 38 Johnson, D. L., 180 Johnson, H., 183, 250 Johnson, H. J., 292 Johnson, J., 233 Johnson, L. c., 349 Johnston, D. W., 281 Johnston, V. L., 23 Jones, B. M., 321 Jones, F. N., 123 Jones, M. c., 262 Jones, R. W., 105 Kaada, B. R., 55 Kahana, B., 228 Kahneman, D., 89, 91, 92, 93, 153 Kajaia, O. A., 107 Kakolewski, J. W., 75 Kaloupek, D. G., 277, 281
AUTHOR INDEX
Kamikawa, K., 26 Kamiya, J., 189, 195, 210, 327 Kanfer, F., 225, 228 Karst, T., 232 Kastaniotis, c., 348 Katkin, E., 250 Kazdin, A., 235 Kejdel, W. D., 107, 124 Kelley, G., 223 Kent, R. N., 283 Kerr, D. 1. B., 110 Kersey, J., 319 Keunishuili, Z., 107 Kewman, D. G., 209 Khechinahuili, S. N., 107, 120 Kiesler, D., 242 Kifer, R., 233 Killam, E. K., 13 Killam, K. F., 11, 12, 13, 15 Kimble, D., 68, 71 Kimble, G., 228 Kimura-Bucholtz, M., 382 Kitajima, H., 106 Kleinman, D., 20, 28, 32, 33,35 Kleitman, N., 322 Kling, A., 314 Klinganman, R. L., 109 Klinger, E., 324, 343, 344, 345 Klinke, R., 15 Kluver, H., 59, 67 Knight, J. J., 107 Knipst,1. N., 12 Knispel, J. D., 111 Knox, V. J., 144, 160, 161, 162 Koella, W. P., 11, 343 Koestler, A., 362, 363 Kohler, W., 102, 124 Kohn, M., 106, 192 Kollar, A., 107 Kooi, K. A., 106 Kopell, B., 122 Komblith, c., 26 Korol'Kova, T. A., 12 Koukkou, M., 320
Koulack, D., 314 Kramarz, P., 186 Kramer, E., 183 Kramer, M., 334, 335, 351, 352 Krasnegor, N. A., 107 Kremen, 1., 292 Krippner, 5., 346, 347 Krishnamorti, 5., 105 Kropf!, W. J., 105 Kruger, L., 80, 377 Kubie, L., 362, 366 Kunnapas, T., 131 Kuznetsova, G. D., 12 Lacey, J. 1.,298 Lader, M. H., 269, 282 Landau, S. G., 105, 106, 120,127 Landes, J., 176 Lane, R., 329 Lang, P. J., 267 Lange, c., 54, 73, 81, 82 Langer, E., 235 Langford, G. W., 343 Larson, D., 102, 103 Lasagna, A., 371, 372 Lasagna, J., 371, 372 Lashley, K. 5., 90 Lauer, L. W., 183, 186 Lavine, R. A., 118, 122 Lazarus, A., 239 Lazarus, R. 5., 294 Lehman, D., 320 Leiman, A. L., 13, 15, 17, 20,26 Lennox, M. A., 58 Levin, P., 343 Levine, M., 233 Levinson, B. W., 155 Levis, D. J., 278, 282, 294 Levy, C. M., 370 Levy, R., 109 Lewin, K., 147 Lewis, E. G., 114, 116 Lewis, H. B., 314, 333, 374 Lewis, M., 233 Lezard, F., 103 Liberson, W. T., 12
391
AUTHOR INDEX
Lichtenstein, M., 120 Lii!beault, A. A., 176 Liebman, J., 107 Liebovitz, M. P., 189, 190 Liebskind, J. c., 167 Lifshitz, K., 106, 192 Lim, H., 92 Lindman, R., 123 Lindsay, P. H., 91 Lindsley, D. B., 105, 188 Lindsley, D. F., 13, 15 Lindsley, O. R., 371 Lefton, W. J., 351 Lipowski, A. J., 127 Livanov, M. N., 11, 12 Livingston, R. B., 55 Ljungberg, L., 123 Lofft, W., 249 Loomis, A. L., 186 London, P., 181, 182, 183, 184, 185, 187, 188, 189, 190, 195, 198 Lowenthal, M., 102 Luce, R. D., 123 Ludwig, A. M., 127, 156 Lundberg, A., 109 Luria, A., 227, 249 Lutz, T. E., 351 Lykken, D. T., 116, 117, 127 Maas, J. W., 104 MacDonald, H., 144, 191, 192, 203, 208, 209 Macinoe, I., 127 MacKay, D. M., 84 MacKinnon, D. W., 362 MacLean, P., 58, 80 MacNeilage, P. F., 317, 318,319 Madell, J. R., 110 Madow, L. W., 183 Magoun, H., 363 Magoun, H. W., 39 Mahoney, M., 223, 226, 232, 234, 237, 242 Majkowski, J., 13 Malis, L. I., 80 Mandler, G., 83, 263, 290, 292
Mandler, J. M., 292 Marco, J., 107 Marcus, J. J., 123 Margolin, S., 366 Mark, R. F., 109 Mark, V. H., 167 Marks, I., 226, 294 Marlatt, G., 248 Marmor, J., 235 Marset, P., 226, 294 Marshall, G., 208 Marshall, J. F., 74 Marshall, W., 226 Maslach, c., 208 Mason, A. A., 192 Mast, T., 107 Mathews, A. M., 269, 270, 274, 280, 281, 282 May, J., 250 Mayer, D. J., 167 McCulloch, W. S., 55 McDevitt, R. A., 185 McDonald, P. J., 192 McFall, R. M., 270 McFarland, D. J., 75 McGuinness, D., 71, 74, 81,88, 89 McGuire, M., 233 McIllwain, J. T., 26 McKellar, P., 363 McKinney, J. A., 248, 249 McLardy, T., 70 McLeod, P., 292 Meddis, R., 343 Meichenbaum, D., 223, 226, 227, 228, 230, 231, 232, 233, 234, 235, 237, 240, 242, 250, 252, 270, 273 Melei, J., 183 Melzack, R., 110, 127, 156, 158, 160, 240 Mettler, A., 70 Meyers, G. M., 349 Millard, D. W., 103 Miller, G. A., 51, 52, 66, 84,86, 89, 93, 147 Miller, J. G., 127, 294 Miller, N., 226 Milstein, V., 187
Mirsky, A. F., 61, 62 Mischel, W., 228, 246 Mishkin, 60, 92 Mihelstaedt, H., 84 Molinari, S., 316, 317 Molino, J., 119 Monahan, J., 228 Monroe, L. J., 350, 351 Montagu, J. D., 106 Moore, E. J., 107 Moore, R. K., 176, 183, 186 Moray, N., 153 Moreira, N. M., 124 Morgades, P. P., 15, 17, 19,27,28 Morgan, A. H., 103, 122, 144, 160, 161, 180, 183, 191, 192, 203, 208 Morgan, C. T., 1 Morlock, J. V., 311 Morlock, M. c., 371 Morrell, F., 10, 15, 26 Morrell, L., 15 Morrissette, J. R., 71 Moruzzi, G., 39 Moscu, K., 371 Motoike, M., 335 Mountcastle, V. B., 110 Mowrer, O. H., 262, 263, 269,271 Munter, P.O., 287 Murphy, D., 106, 111, 368 Murray, E., 250 Murray, H., 228 Murtaugh, T., 104, 111 Mushin, J., 109 Muzio, J., 347 Myers, T., 368 Nathan, P. W., 158 Nebylitsyn, V. D., 125 Neisser, V., 86 Nelson, R., 283 Newell, A., 149 Nisbett, R. E., 287 Nomikos, M. S., 294 Norman, D. A., 153 Norton, J. c., 112 Norton, T. T., 91
AUTHOR INDEX
392 Novaco, R., 235 Nowliss, D., 190, 191 O'Brien, G. T., 277 O'Brien, J. H., 26 O'Conne ll, D. N., 179, 181, 343, 378 Offenloc h, K., 110 Ogura, H., 12, 26 Okuma, T., 335 Olds, J., 17, 26 Olds, M. E., 17, 26 O'Leary, D., 228 Oliveras , J. L., 167 Opton, E. M., 294 Orne, M. T., 179, 181, 183,202, 204,343, 378 Ornstein , P. H., 334 Ornstein , R., 52, 380 Osborn, A G., 155 Osborne , R. T., 116 Oscarsso n, 0., 109 Ostrande r, S., 369 Overton, D. A., 109, 154, 322 Palkes, H., 228 Parke, R., 228 Parsons, O. A., 321 Paskewit z, D. A., 202 Pattee. H. H., 87 Patterson , G. R., 271 Paul, G. L.,266, 274, 281 Pavlov,1 . P., 125, 126, 128,262 Peale, N. V., 251 Pearlman , C. A., 348, 349 Pearson, J. D., 343 Peffer, K., 362, 377 Perlmutt er, L., 228 Perry, A., 111 Peters, J., 103, 125 Peterson , D., 225 Petrie, A., 101, 102, 104, 105, 117, 123, 125, 127, 128 Pfefferba um, A., 104, 106, 108, 117, 122 Philipov , E., 370 Phillips, A. G., 73
Phillips, E., 233 Piaget, J., 147 Picton, T. W., 15, 107 Piper, W. E., 280 Pitkin, W., 245 Pivik, T., 350 Platonov , K., 250 Platt, D., 102, 103 Platt, J., 233 Platz, P., 15 Polgrin, D., 226 Poliakov , K. L., 11, 12 Pollak, c., 107 Pontalti, c., 327 Poppen, R., 92 Porter, R. W., 26 Powell, T. P. S., 80 Premack , D., 226 Pribram, K. H., 51, 52, 55, 58, 60, 61, 64, 65, 67, 68, 71, 73, 74, 76, 77, 80, 81, 84,86,88 ,89,90,9 1,92, 147 Price, J. L., 80 Prince, M., 138, 141 Pruvot, P., 12 Prytulak , S., 103 Ramos, A., 13, 27, 28 Randall, W., 39 Rankin, N. 0.,294 Rapin, 1., 107 Rappapo rt, M., 104, 112 Ray, A. A., 282, 283, 285, 286 Rayner, R., 262 Reason, J. T., 123, 124 Rechtsch affen, A, 327, 332, 333, 334, 349 Redmon d, D. E., 104 Regan, D., 106 Rescorla, R. A., 262, 269 Reyher, J. A., 181, 182 Rhead, J., 382 Rhead, J. c., 190, 191 Ricci, G., 26 Riccio, D., 294 Richards on, A, 236 Richards on, F., 235 Richey, E. T., 106
Ridberg, E., 228 Rietveld, W. J., 105 Riggs, L. A., 15 Rinzler, S. H., 159 Ripps. H., 105 Roberts, A. H., 209 Roberts, W. W., 75 Robinson , R., 77 Rochman , G., 185 Rodin, J., 287 Roehrs, T., 351 Roffwarg , H. P., 347 Rohrbau gh, M., 294 Rose, D. E., 107 Rosenha n, 183, 185 Rosner, B. S., 109 Ross, L., 287 Rosvold, H. E., 61, 62 Roth, T., 351, 352 Roth, W. T., 120 Rothman , H. H., 107 Rubin, F., 371, 372 Ruch, T. c., 7 Ruchkin , D. S., 12, 17, 20 Rugg, H., 362 Ruhm, H. B., 107 Rule, S. J., 123 Rusinov, V. S., 15 Ryan, E. c., 102 Sachdev, K., 187 Sachs, E., 13, 17, 20, 55 Sachs, L., 226 Sachs, L. B., 181 Sachs, M., 105 Sakhuili na, G. T., 15 Sales, S. M., 123, 127 Saltz, E., 233 Samson, H. H., 106 Sanders, R. S., 182 Sarbin, T., 235 Sarbin, T. R., 148, 176, 183 Saslow, G., 225 Schachte r, S., 83, 240, 250, 263, 269, 287 Schaefler , K., 185 Schafer, R., 183 Schechte r, G., 121, 122 Schechte r, N., 335
393
AUTHOR INDEX
Schimmel, H., 107 Schneider, M., 228, 233 Schmeidler, G., 335 Schooler, c., 103 Schramm, S., 26 Schreiber, F. R., 156 Schroeder, L., 369 Schubot, E., 181 Schulman, R. E., 183 Schwartz, E., 13, 27, 28 Schwartz, G., 250, 381 Schwartz, M., 105, 109 Schwartzbaum, J. S., 65, 66,67,71 Schweitzer, P. K., 109 Sears, T. A., 110 Segal, M., 26 Seyfried, B. A., 337 Shagass, c., 104, 105, 106, 109, 112, 114, 115 Shapiro, A., 355 Shapiro, D., 236 Shapiro, J., 182 Shaw, P. M., 281 Sheatz, G. c., 12, 15 Shenkin, H. A., 70 Shevach, B. J., 141 Shevets, G. c., 12 Shimokochi, M., 19, 20, 28 Shipley, T., 105 Shipman, W. G., 280 Shor, R. E., 179, 181, 184, 343,378 Shure, M., 232 Sidis, B., 141 Siegel, J., 104, 111 Sifneos, P., 233 Silverman, J., 103, 104, 105, 125, 127, 128 Simkins, L., 248 Simon, c., 370 Simon, H. A., 89, 149 Simpson, L., 363 Singer, J., 238, 240, 253, 355 Singerman, K., 287 Sittenfeld, P., 373 Skinner, B. F., 226, 262 Slama, K., 275
Smith, B. A., 334 Smith, S., 368 Soderberg, G., 123 Soloman, P., 102 Soloman, R. L., 262, 269 Soskis, D. A., 104, 106, 112,114 Sparks, D. L., 26 Spencer, V., 370 Sperry, R., 155 Spielberger, C. D., 263, 266,267 Spilker, B., 104 Spinelli, D. N., 90 Spivack, G., 232, 23 Spreng, M., 107, 124 Staab, L., 370 Staal, M., 335 Stampfl, T. G., 278, 282 Stark, L. H., 112 Starker, S., 335 Steffy, R., 223 Steinberg, N. N., 7 Steiner, I., 237 Steiner, J., 109 Steinmark, S. W., 281 Stellar, E., 1 Stephens, S. D. G., 124 Stephenson, C. W., 186 Stephenson, W., 298 Stem, W. E., 348 Sternbach, R. A., 160, 25( Stevens, J. R., 187 Stevens, S. S., 123 Stevenson, J. A., 66 Stewart, M., 228 Stoller, R. J., 156 Stone, N. M., 277 Stone, W. c., 251, 285, 294,296 Storms, M. D., 287 Stoyva, J., 327, 363, 373 Straumanis, J. J., 109 Strauss, J., 105 Strupp, H., 244 Stukat, K. G., 183, 186, 187 Suedfeld, P., 368 Suinn, R., 235 Sunami, Y., 335
Sushinsky, L. W., 283 Sutcliffe, J., 248 Sutton, S., 15 Suzuki, T., 107 Svorad, D., 190, 191 Svyadoshch, A., 372, 374 Swanson, E. M., 350 Sweeney, D. R., 102 Sweet, W. H., 80, 158 Taguchi, K., 107 Takeo, S., 335 Talbott, A. D., 299 Tart, C. T., 181, 182 Taub, E., 209 Taub, J. M., 318 Teas, D., 107 Teitelbaum, P., 66, 74 Tellegen, A., 116, 127 Tepas, D. I., 105, 108, 109 Teplov, B. M., 125 Thigpen, C. H., 156 Thorkelson, K., 116 Thompson, R. F., 10 Throop, W. F., 123, 127 Tizard, J., 127 Tolman, E. c., 147 Tomkins, S. S., 183 Tordior, W. E. M., 105 Tori, c., 235 Toth, M., 156 Tourk, L. M., 107 Traub, A. c., 349 Travell, J., 159 Travis, R. P., 26 Trier, c., 235 Treisman, A. M., 153 Trexler, L., 232 True, R. M., 186 Tubbs, W. E., 92 Tueting, P., 15 Turk, D., 235, 240 Tursky, B., 117 Twentyman, C. T., 270 Ulet!, G. A., 182, 191, 203, 207 Ullmann, L. P., 292, 347 Ungerstedt, V., 73
394
Uttal, W. R., 109 Uviller, E. T., 292 Vale, T. c., 7 Vassilevsky, N. N., 26 Valenstein, E. S., 75 Valins, S., 282, 283, 285, 286 Van de Castle, R. L., 328, 329,332 Vander Tweel, C. H., 106 Van Lehn, R., 298 Vaughn, c., 344 Vaughn, H. G., 105, 106 Venables, P. H., 127 Verduyn Lunel, H. F. E., 106 Villegas, J., 12 Vogel, G., 362 Vogel, G. W., 349 von Bonin, G., 55 von Foerster, H., 86 von Hartman, 138 Vranceanu, M., 371 Vygotsky, L., 227, 249 Waggoner, R., 106 Walker, J. P., 40 Wall, P. D., 55, 109, 110, 127, 158,240 Wall, R. L., 277, 285 Walter, W. G., 16
AUTHOR INDEX
Walters, D., 362, 364, 365 Wanschura, R. G., 123 Ward, A. A., 55 Watson, J. B., 262 Watson, P. D., 117 Watts, J. W., 158 Webb, W. B., 319 Weinberg, H., 16 Weinberg, L., 230 Weiskrantz, L., 60, 64 Weisz, R., 328 Weitzenhoffer, A. M., 176, 177, 178, 179, 181, 183, 186 Weitzenhoffer, G. B., 183 Welch, G., 188 White, R. W., 141 White, T. C., 120 Whitman, R. M., 334 Whittle, P., 15 Wicke, J. D., 105, 188 Wickram, I., 182 Wickramasekera, L, 182 Wilbur, C. B., 156 Wilcox, W. W., 183 Wilkins, W., 274, 281 Williams, H. L., 371 Willows, A. 0., 3 Wilson, G. T., 274, 283 Wilson, M., 92 Wilson, W. A., 71 Winer, B., 198
Winters, R. W., 106 Wiseman, R. J., 181 Witkin, H. A., 314, 321, 329,333,334,337,351, 374 Wittner, W., 122 Wogan, M., 280 Wolfe, G., 314, 321 Wolfer, J., 235 Wolk, I., 102 Wolpe, J., 239 Woody, C. D., 26 Woolsey, T. A., 80 Wooten, B. R., 106 Worrell, L., 235 Wundt, W., 1, 3, 81, 88 Yakolev, P. L., 167 Yeager, C. L., 186 Yoshii, N., 12, 13, 26, 107 Young, L. D., 292 Young, M. L., 106 Zapporozhets, A., 249 Zerlin, S., 107 Zimbardo, P., 208, 250, 287 Zubeck, J. P., 125, 188, 189 Zubin, J., 16 Zuckerman, M., 104, 111, 127
Subject Index
Activation, 72-73, 90, 92-93 Agency, 335-337 Alpha rhythm, 362-365 changes in relationship to age, 188 and hypnosis, 186, 203, 215, 216 skin temperature control, 208-215 stability of, 194, 195 and susceptibility, 189-191, 193, 195-203, 207-215 training, 195-203, 379 Altered states of consciousness, 361 Amnesia, 141 and dissociation, 142 versus repression, 142-143 Anxiety (see also Fear) and arousal (see Arousal) and cognition, 279-282 individual differences, 267, 268, 272-274, 276,289,298-305 intervention strategies, 272-276 as operant learning, 262 as Pavlovian conditioning, 261 snake phobia, 277, 278, 283, 293, 295 social, 277, 295, 296 speech,285 stimulus conditions of, 264-266 Anxiety reaction, 264, 266-268, 277 maintenance of, 268, 272-273, 276 cognitive interpretation, 269-271 images, 270 overt behavior, 269-270 physiological reaction, 269 self-verbalizations, 270 reduction of, 268, 272-276, 306-307 Archetypal images, 366 Arousal, 72-73, 90, 92-93, 120, 250, 361 average evoked response amplitude, 120, 122
Arousal-Continued anxiety, 261, 264, 266, 269, 271, 272, 277, 279-288,292-293 individual differences, 289 low arousal state (see also Twilight state), 362 alpha training, 379 biofeedback, 373-379 electromyographic biofeedback, 373-374 information processing, 372 production, 373-374, 379-380 sleep learning, 370--372 perceived arousal, 290--292 attention focusing, 286, 296-297 autonomic perception questionnaire, 290--292 incubation phenomenon, 294-295 overt behavioral activity, 295-296 subject characteristics, 292-294 Attention, 90--93, 249-250 average evoked response amplitude, 120, 122 divided attention theories, 153 Attention span, 88 Attitude changes sensory deprivation, 368-369 twilight state, 368-369 Audioanalgesia, 118, 122 Augmenting, augmenters, 101-104, 107, 110 arousal, 111, 120 habituation, 106 interstimulus interval, 106 nervous system strength, 126 noise tolerance, 119 pain tolerance, 118, 125 power functions, 123-124 visual, 105-106 Autogenic relaxation, 364 Automatic thoughts, 226
395
396
Automatisms, 139 Autonomic awareness, 276 Autonomic perception, 282, 289, 294 patterns of, 297-305 questionnaire, 290-293, 295, 296, 298-305 Average evoked response(s) (AERs), 14, 10>104 arousal, 120, 122 attention, 120, 122 auditory, 107 augmenting/reducing, 105--107 differential generalization, 17-19 electroretinogram, 105 exogenous and endogenous processes, 25--26 loud noise, 122 meaning, 21 mental arithmetic, 120 noise tolerance, 119 power functions, 12>-124 pupillary diameter, 106, 112 readout or emitted potentials, 16 somatosensory, 109 stimulus intensity, 105--106 tone frequencies, 107 visual, 105 Average evoked response amplitude/intensity slope, 105, 111, 114, 124 arousal, 120, 122, 123 attention, 120, 122, 123 chromosome differences, 117 nervous system strength, 126 noise tolerance, 119 P100, 105, 106, 110 sex differences, 117 twins, 112, 115 Avoidance behavior, 262, 266, 272, 277, 282283,285,295 cognitive, 278-280 Barber Suggestibility Scale, 179-180 Behaviorism, 1 Biofeedback, 8>-84, 382 creativity and, 364-365 electromyographic, 373 theta rhythm, 364-365, 37>-375, 377, 379, 380 twilight state, 364-365, 37>-380 Brain stimulation, electrical, 31-38 Buchsbaum-Silverman stimulus-intensity control model, 125 California Personality Inventory, 335
SUBJECT INDEX
Capacity central,88 spare, 90, 91 Cell assemblies, 147 Cerebrallateralization dreaming, 320-321 hypnotizability, 155--156 twilight learning, 380-381 Children's Hypnotic Susceptibility Scale, 179 Coconscious, 138 Cognition(s), 6, 224-225, 237, 279 as behavior, 225--226 of change, 245--246 conceptualization of therapy, 239-242 as coping skills, 234-236 as defense mechanisms, 236 as final common pathways, 238-239 as irrational beliefs, 230-232 as irrational thinking, 229-230 as problem-solving ability, 232-234 as response, 226 Cognitive avoidance behavior, 279-280 Cognitive balance, 383 Cognitive-behavior modification, 223 Cognitive control system, 137, 145 central control structure, 148-152, 16>-165 conflict resolution, 150-151 multiple structures, 146--147 pain experience, 16>-165 pain reduction, 16>-165 Cognitive expectancy, 280-282 Cognitive modeling, 227-228, 231, 233, 235 Cognitive process, 6 Cognitive restructuring, 231, 232, 244, 251, 269 Cognitive structure, 147 Communion, 335--337 Competency, 89, 92-93 Conditioning operant, 262 Pavlovian, 261 Consciousness, 1, 2, 3, 52, 137, 138, 142, 144, 152 Constraint, internal versus external, 91 Contrast effects, 103 Control theory, 84 Coping skills training, 234, 235 Creative, creativity biofeedback, 364--366 dream content, 334-335 theta rhythm, 364-366 twilight state, 36>-365
397
SUBJECT INDEX
Daydreaming, 234, 335 Defense mechanisms cognitions, 236 twilight state, 374 Diagnostic scale of susceptibility, 179 Dissociation anesthetics, 154, 155 and cerebrallateralization, 155 divided attention, 153 fugues, 156 hypnosis, 142-144 multiple personalities, 156 pain, 144, 157-168 between pain and suffering, 160--168 recoverable amnesia, 153, 154 within sleep, 156 Dream(s), dreaming, 313 affect and dream content, 338-340 presleep mood, 337-341 sensitizers/repressors, 338 animal content of, 329 clinical interpretations of, 330--332 compensation hypothesis, 333, 334 continuity hypothesis, 333--335, 337 disorganization of, 325, 326 ego function during, 325, 326 experience, 313 functions of, 324, 325, 330, 345--355 separate from sleep, 350--354 interference, 321, 324 interpretation versus analysis of, 328, 329 laboratory versus home, 328 menstrual cycle and, 345 in non-human animals, 344, 345 relationship to REM sleep, 349--351, 354, 355 validity of reporting, 327, 328 variation throughout sleep, 354--355 Dream content, 324 creativity, 325, 334 daydreamers, 335 masculinityifemininity, 335--337 neuroticism, 338-343 physiological activity, 332 presleep conditions, 333--341 schizophrenics, 333, 334 Dream recall, 324 dream affect, 314--316 dream discontinuity, 325 dream report, 313 dream salience, 317-320
Dream recall-Continued electroencephalogram, 320 emotionality, 316 presleep stress, 314--318 REM sleep, 316--322, 326, 347 repression, 313--317 right hemisphere, 320--321 sensitization, 314--316 sleep duration, 322-323 Dream salience dream recall, 317-320 presleep stress, 318 REM sleep, 317-320 right hemisphere, 320--321 EEG (see Electroencephalogram) Effort, 73, 80--81, 90, 91, 93 Electroencephalogram alpha (see Alpha rhythm) changes in synchrony, 10 changes with age, 187 changes with deprivation, 188 exogenous and endogenous components, 13 hypnotic state, 186, 216, 217 versus simulated, 203--204 versus sleeping, 185 versus waking, 185, 203, 207 hypnotic susceptibility, 187-188, 192 alpha rhythm, 189--191, 193 evoked potentials, 192 feedback, 195--203 low versus high susceptibility, 207-215 irradiation and consolidation, 11 stability of, 193--195 tracer technique, 11-12 Electromyographic biofeedback, 373 Electroretinogram, 105 Emotion, 52-53, 58, 59, 66, 69, 72, 73, 74--75, 81--83, 84 Engram, 24 Fear, 262, 266--267, 269--271, 273--274, 277, 280,281,286,294--295 Feedback, 84--87 Feedforward, 84--87 Fourier analysis, 79 Fugues, 141, 143 Gate control theory, 110 Habit-family hierarchy, 147
398 Habituation, 69-73 Harvard Group Scale of Hypnotic Susceptibility, 179 Hyperneuron, 43--45 Hypnogogic imagery, 363--366, 372, 376, 380 Hypnopaedia, 371 Hypnosis, hypnotic state, 137, 173--174, 216217 autonomic functioning, 185 children, 183--184 definition of, 174 electroencephalogram, 185-187, 203--207 simulated versus hypnotic state, 203--204 sleeping versus hypnotic state, 185 waking versus hypnotic state, 185,203, 207 observing consciousness, 145 pain, 144, 160-168 pain/suffering reduction, 163--168 posthypnotic amnesia, 142-143 suffering, 160-168 Hypnotic analgesia, 144, 161-166 medial thalamus, 167 Hypnotic susceptibility, 216-217 age and development, 183--184, 187-188 definition of, 175 electroencephalogram (see also Electroencephalogram, hypnotic susceptibility), 187-189, 192, 215-217 individual differences, 182-185 modification of, 181-182, 204-207, 216 motivation, 184-185, 216 scales of, 176-177, 179 skin temperature control, 208-215 stability of, 180
SUBJECT INDEX
Kinesthetic figural aftereffects task, 102-104 Kliiver-Bucy syndrome, 59 Limbic mediobasal motor cortex, 54-66 and habituation, 69-73 lesions of, 60 effects on feeding, 64-66 effects on fighting, 61-63 effects on fleeing, 64 effects on transfer of training, 67 Mental blocks, in twilight state, 374 Mental practice, 236 Modeling (see Cognitive modeling) Morphine, and medial thalamus, 167-168 Motivation, 52-53, 58, 59, 66, 69, 72, 73, 74-75, 81-83, 84 Motive, 53 Mowrer's theory of learning, 262-263 Multiple personalities, 141, 143 three personalities of Leonie, 139-141 Naloxone, and hypnotic analgesia, 167-168 Neodissociation, 141-142 and divided attention theories, 153 Nervous system strength, 126 Neuroticism and dreams, 329, 332, 333,338342 Observational learning, 7-8
Pain gate control theory of, 158, 165 motivational-emotional component, 157 phantom limb, 160 reduction and the medial thalamus, 166167 Images, imagery, 147, 223--224, 227-228, 229230,233--236,238,244,262-265,270, referral type, 159 279-280 sensory information component, 157-159 Implosive therapy, 278 versus suffering, 157, 160-168 Incompatible thoughts, 226, 230, 238, 245 Pain tolerance, 102 Information, 86-87 average evoked response, 117-118 error, 86-87 distraction, 127 redundancy, 87 reducing, 117-118 Intent, 53 Perceptions, 3, 4 Intentionality, 52 Personality (see Dream content) Internal dialogue, 224-225, 237,244,245,246, Petrie procedure (see Kinesthetic figural 248,249 aftereffects task), 102, 103, 123 Phantom limb, 160 James-Lange theory of emotion, 53--54, 57-58, Physiological psychology, 1 73--74, 81-83 Physiological reactors, 277, 279-288
399
SUBJECT INDEX
Posthypnotic amnesia, 142 versus repression, 142-143 Precentral motor cortex, 76-80 Primary process, 142 Problem representation, 346 Problem solving cognitions, 232-234 dreams, 346 training, 232 Psychological magnitude, 123
Recall in twilight state, 378 Recoverable amnesia, 154 Reducing, reducers, 101-104, 110, 126 arousal, 111, 120, 122-123 attention, 120, 122-123 auditory, 107-108 genetic factors, 115-117 habituation, 106 interstimulus interval, 106 nervous system strength, 126 noise tolerance, 119 pain tolerance, 117, 125 power functions, 123--124 single unit recording, 109 sleep, 108 somatosensory, 109 visual, 105-106 Redundancy, 87, 91-92 REM sleep, 363, 378 deprivation effects, 348-349, 350--351 distinction from NREM sleep, 347-349 dream recall, 316--322 functions of, 346--350 in monkeys, 344 neuroticism, 332-333 presleep conditions, 332 rebound, 350 sleep learning, 370--371 Repression amnesia, 142-143 con tentless dream reporting, 321 dream recall, 313-317 Repressors and dream affect, 338 Resistances and twilight state, 374 Response specification, 175 Retrieval-difficulties in twilight state, 364, 381-382 Roles, 148
Satiation, 102 Schizophrenics, 105, 127 dream content, 334-335 REM deprivation, 349 Secondary process, 142 Self, 1, 5 Self-awareness, 1, 5, 6 Self-consciousness, 52 Self-instruction, 227-228, 229, 251-253 training, 223, 227-228 Self-observation, 243 Self-verbalizations, 270, 273 Sensations, 3 Sensitizers and dream affect, 338 Sensory deprivation, 368 Silverman's procedure (see Kinesthetic figural aftereffects task), 103 Sleep (see also REM sleep) sleep state, 362 stage 1, 371-372, 378 stage 2, 372 stage 3, 371, 377 stage 4, 370--371, 377 Sleep learning, 370--372, 378 Snake phobia, 277, 278, 283, 293, 295 Social anxiety, 277, 295-296 Somnambulism, 141-142, 156 Speech anxiety, 285 Stanford Hypnotic Susceptibility Scale, 177179 State-dependent learning, 154-155 Stevens' procedure, 123 Stress-inoculation training, 235, 273 Strychnine neuronography, 55, 58 Subconscious, 138 Subjective experience, 3, 4 Subordinate ego-structures, 148 Suggestopaedia, 369 Susceptibility Scales, 176--181 Barber Suggestibility Scale, 179--180 Children's Hypnotic Susceptibility Scale, 179 Diagnostic scale of susceptibility, 179 Harvard Group Scale of Hypnotic Susceptibility, 179 Stanford Hypnotic Susceptibility Scale, 177-179 Systematic desensitization, 275, 278, 286 Thalamus, medial hypnotic analgesia, 167
400 morphine, 167-168 pain, 166-167 Theta states, 365, 373, 375 Theta training, 373-375, 380 Twilight state, 362-364 attitude change, 367-368 biofeedback, 373-379 creativity, 363-366
SUBJECT INDEX
hypersuggestibility, 367 learning, 367-372 production, 373-380 retrieval, 381-382 theta rhythm, 364-366 Unconscious, 138 Unit responses, 26-30
