On Practice: How the Brain Connects Piano Keys and Piano Sounds MARC BANGERT, UDO HAEUSLER, a AND ECKART ALTENMÜLLER Institute for Music Physiology Physiology and Performing Performing Arts Medicine, State Academy of Music and Drama, Hanover, Germany aGerman Primate Center, Goettingen, Germany
KEYWORDS: Piano practice; Brain, joint activation of auditory and sensorimotor areas of; Brain, representaiton representaiton of ear and hand in
INTRODUCTION
This review of our unpublished data demonstrates how practice promotes a joint mental representation of ear and hand, which is a phenomenon familiar to most musicians: silent dexterity drills produce “audible tones inside the head,” and, vice versa, sounding music “flashes right into the fingers.” By dissociating auditory and motor features of piano playing in the experimental paradigm, we tested the effects of short-term (20 minutes), long-term (5 weeks), and lifetime ( >15 years, in professional musicians) piano practice.
MATERIAL AND METHODS
Twenty-four right-handed subjects, two groups of nine novices each (“map group” and “no-map group,” see below) and an expert performer group of six professional pianists, were examined. A single session consisted of (1) an EEG recording during a set of 60 purely auditory tasks (passive listening to piano melodies) and 60 purely motor tasks (pressing keys on a silent piano keyboard with the right hand) in randomized order, (2) an adaptive and interactive practicing procedure (duration of 20 minutes) in which the novices had to replay acoustically presented melodies (the complexity level level of subsequent tasks was adjusted according to the online analysis of the player’s player’s performance), and (3) another EEG recording identical to the first one, again on the silent keyboard. The map group then worked with the training system twice a week over a period of five weeks with EEG monitoring during duri ng the first, sixth, and eleventh (last) session. The no-map group attended the same program, but the key-to-note assignment was Address for correspondence: Dr. Marc Bangert, Institute for Music Physiology and Performing Arts Medicine, State Academy of Music and Drama, Plathnerstrasse 35, 30175 Hanover, Germany. Voice: +49-511-3100-552; fax: +49-511-3100-557.
[email protected]
425
426
ANNALS NEW YORK ACADEMY OF SCIENCES
. n o i t p a c r o f e g a p g n i w o l l o f e e S . 1
E R U G I F
BANGERT et al.: PIANO PRACTICING AND THE BRAIN
427
shuffled after each single trial. In other words, they had the same chance to acquire timing and dynamics skills as the map learners, but no chance to learn any connection between fingers and pitches. Analysis of the 32-channel EEG1 included the topographical distribution of taskrelated DC-EEG changes, task-related interelectrode coherence.2 Based on DC-EEG data, the following regions of interest (ROI) were predefined for statistical analysis: frontotemporal areas (left and right) and motor/premotor areas of the left hemisphere. The investigated frequency bands comprised alpha (8–13 Hz), low beta (13 – 21 Hz), gamma (35 –45 Hz), and delta (0.1 –4 Hz) frequencies. A vectorial description of the potential distributions allowed the dot product to be used as a measure of pattern similarity. Additionally, the subjects ’ performance based on MIDI data (note onset times and key velocities) was examined. In a subsequent study, classical conditioning of the eyeblink reflex by means of an airpuff to one of five notes was used to determine if an intermodal transfer of the stimulus (silent piano keys instead of audible pitches) is possible in highly trained pianists. During the procedure the subjects (15 nonpianists, 16 pianists) were distracted by verbal semantic and working memory tasks mentally.
RESULTS
The five-week training schedule as well as a single training session induced common activation of auditory and motor areas in the DC-EEG (F IG . 1). In the auditory task the additional activation of premotor and motor areas as well as parietal cortex and frontolateral areas was observed, whereas in the silent motor task the additional activation of the temporal lobes, active inhibition of the sensorimotor field of the unused left hand, and again activity in the right lateral frontal lobe were established with increasing practice time. Vector analysis of cortical pattern similarities indicated that in the novices the activation patterns for the physically different tasks in “passive listening” and “silent motion” were different in the first place but became more similar in the course of practice. A comparison betwee n the two training groups (map vs. no map) produced a right frontotemporal region ( p < 0.01), possibly reflecting an additional skill that only the map group had been able to acquire. The task-related coherence, in the predefined ROIs serving as a measure of functional coupling of mental ears and fingers, yielded (specific to the map group) increased coupling for the passive auditory and the silent motor task, after one session constrained to specific bands and after five weeks broadband coupling of all ROIs.
FIGURE 1. Differences in cortical task-related DC-potential patterns of the map group compared to the normal prestudy activation (e.g., motor cortex when moving), which is displayed on the left . The three panels on the right show exclusively the additional activity due to training. Each pattern was acquired 2–3 seconds after stimulus/movement onset and was averaged across 1,000 ms and 60 trials in 9 subjects. Dark = negative potential (depolarization/activation) or increased activity; light = decreased activity or inhibition.
428
ANNALS NEW YORK ACADEMY OF SCIENCES
Behavior al Data
Under ordinary practicing conditions (i.e., with auditory feedback), the no-map group performed much better after five weeks with respect to timing and dosage of fine finger forces. A possible reason for this effect may be the additional interfe rence for the map group originating from processing the pitch (melody) feedback. Under conditions of missing feedback, the two groups performed similarly. Five weeks of practice yielded improvements of finger velocity, velocity variances, and timing. In the blink reflex experiment, after successful conditioning, the subjects were asked to voluntarily press keys on a silent keyboard. As expected, nonmusicians showed no specific reaction. Professional pianists, however, exhibited a markedly intensified eyeblink when pressing the corresponding key.
CONCLUSIONS
Piano playing provides a model for a motor system under auditory control with high demands on quick and precise coordination. After practice, cortical auditory and sensorimotor hand areas are jointly activated for purely auditory as well as for mute motor tasks. The topographical distributions induced by each of the involved modalities are similar. The degree of similarity correlates to the accumulated practice time. In addition, a right dorsolateral prefrontal area engages in this corepresentation in beginners and experts , but not in beginners who could not establish a mental map of the keyboard. This map area is connected to motor fields functionally, as indicated by increasing coherences during practice. Interestingly, the strong corepresentation is established during the first minutes of training, 3 is consolidated within weeks, and may provide the basis for further skill achievement. After years of practice, the described sensorimotor corepresentation is automatized to a high degree and can even be activated preattentively.4
REFERENCES
1. JASPER, H.H. 1958. The ten-twenty electrode system of the international federation. Electroencephal. Clin. Neurophysiol. 10: 371 –375. 2. ANDRES, F.G. et al. 1999. Functional coupling of human cortical sensorimotor areas during bimanual skill acquisition. Brain 122: 855 –870. 3. CLASSEN , J. et al. 1998. Rapid plasticity of human cortical movement representation induced by practice. J. Neurophysiol. 79: 1117–1123. 4. KOELSCH, S., E. S CHROGER & M. TERVANIEMI. 1999. Superior pre-attentive auditory processing in musicians. Neuroreport 10: 1309–1313.
