The brain basis of piano performance

https://doi.org/10.1016/j.neuropsychologia.2004.11.007Get rights and content

Abstract

Performances of memorized piano compositions unfold via dynamic integrations of motor, perceptual, cognitive, and emotive operations. The functional neuroanatomy of such elaborately skilled achievements was characterized in the present study by using 150-water positron emission tomography to image blindfolded pianists performing a concerto by J.S. Bach. The resulting brain activity was referenced to that for bimanual performance of memorized major scales. Scales and concerto performances both activated primary motor cortex, corresponding somatosensory areas, inferior parietal cortex, supplementary motor area, motor cingulate, bilateral superior and middle temporal cortex, right thalamus, anterior and posterior cerebellum. Regions specifically supporting the concerto performance included superior and middle temporal cortex, planum polare, thalamus, basal ganglia, posterior cerebellum, dorsolateral premotor cortex, right insula, right supplementary motor area, lingual gyrus, and posterior cingulate. Areas specifically implicated in generating and playing scales were posterior cingulate, middle temporal, right middle frontal, and right precuneus cortices, with lesser increases in right hemispheric superior temporal, temporoparietal, fusiform, precuneus, and prefrontal cortices, along with left inferior frontal gyrus. Finally, much greater deactivations were present for playing the concerto than scales. This seems to reflect a deeper attentional focus in which tonically active orienting and evaluative processes, among others, are suspended. This inference is supported by observed deactivations in posterior cingulate, parahippocampus, precuneus, prefrontal, middle temporal, and posterior cerebellar cortices. For each of the foregoing analyses, a distributed set of interacting localized functions is outlined for future test.

Introduction

Musical performance is very likely the domain in which humans produce the most intricate, complex integration of expert perceptual, motor, cognitive, and emotive skills. But although it may be the pinnacle of human central nervous system performance (and what space aliens most covet), its basis in the brain rarely has been investigated. Fortunately, musical ability and cognition appear to yield to fractionation (e.g., Peretz & Coltheart, 2003; Sergent, 1993), and components of musical performance have been studied with neurological, electrophysiological, and neuroimaging methods. These performance components include perception, sight-reading, motor-sensory processes, and attention.

The most deeply studied component is the neural basis of perceptual aspects of musical performance. Researchers have demonstrated, for example, strong associations amongst the strength of neurophysiological responses to pure tones in the musical range (detected via magnetoencephlography, MEG), the volume of anterior-medial Heschl's gyrus from which the responses originate, and musical skill (Schneider et al., 2002). Others have demonstrated enhanced neural representation for the timbre of the instrument in which a musician specializes, as compared to others (Pantev, Roberts, Schultz, Egnelien, & Ross, 2001) and the differences in the neural representation of musical pitch and rhythm between musicians and individuals with very little musical performance experience or training (Evers, Dannert, Rodding, Rotter, & Ringelstein, 1999; Parsons & Thaut, 2001). Studies have also examined the relation between auditory perception and motor behavior. Thus, one MEG study demonstrated that when pianists, but not singers, listen to familiar piano pieces to detect errors, they exhibited involuntary activations in cerebral cortical motor systems (Haueisen & Knosche, 2001).

Fewer functional brain investigations have targeted activities more intimately related to the production aspects of musical performances, which is the focus of the present paper. Various approaches have been used to investigate the sight-reading of musical scores. Several neurological case studies examined musicians’ acquired impairments in sight-reading (Cappelletti, Waley-Cohen, Butterworth, & Kopelman, 2000; Judd, Gardner, & Geschwind, 1983; Marin & Perry, 1992; Sergent, 1993; Stewart & Walsh, 2001). Positron emission tomography (PET) has been used to study pianists sight-reading (Sergent, Zuck, Terriah, & McDonald, 1992), and to study conductors sight-reading a score as they detected errors in its heard performance (Parsons, Hodges, & Fox, 1998). More recently, MEG was used to investigate musicians imagining the musical sounds of a score they sight-read (Schurmann, Raij, Fujiki, & Hari, 2002). These studies and others (e.g., Nakada, Fujii, Suzuki, & Kwee, 1998; Schon, Anton, Roth, & Besson, 2002; Stewart et al., 2003) suggest that a core distributed network of areas in parietal, temporal, and occipital cortices support sight-reading, with other areas in frontal, sub-cortical, and cerebellar areas, being recruited depending on whether the score is merely read, read and imagined to be heard, or read while being performed.

A variety of research has focused on sensorimotor processes related to performing on musical instruments. MEG studies indicate that the extent of cortical representations for musicians’ digits is related to the degree of skilled performance with those digits, as well as to the age at which the musicians started training on the musical instrument (Elbert, Pantev, Wienbruch, Rockstroh, & Taub, 1995). Such functional differences are complemented by anatomical MRI studies reporting increased size and specific structural differences in musicians, as compared to non-musicians, in areas such as planum temporale, anterior corpus callosum, hand primary motor cortex, anterior-medial Heschl's gyrus, and anterior cerebellum (see Gaser & Schlaug, 2003; Munte, Altenmuller, & Jancke, 2002; Schneider et al., 2002).

Musicians performing rhythmic, sequential finger tapping tasks have been studied with functional neuroimaging. When performing a novel, simple unimanual tapping task, a within-session increase in neural activity in primary motor cortex was detected in musicians, but not in non-musicians, implicating adaptive motor skill processes already acquired or present in musicians (Hund-Georgiadis & von Cramon, 1999). The musicians exhibited at the same time a smaller extent of activation in supplementary motor area (SMA), pre-SMA, and motor cingulate than did non-musicians, implying more efficient motor control processes. Similar activation patterns were observed when musicians perform complex bimanual tapping tasks (Jäncke, Shah, & Peters, 2000), with the exception that less activity is seen in primary motor cortex for musicians than non-musicians. An fMRI study of complex sequences of unimanual finger tapping reported significantly reduced activation for musicians as compared to non-musicians in primary motor and premotor cortices, SMA, and superior parietal cortex (Krings et al., 2000). A metanalysis of neuroimaging studies examined the temporal and sequence ordering involved in the foregoing kinds of tapping tasks (Janata & Grafton, 2003). The results suggested that such tasks elicit responses in sensorimotor cortex, SMA, cerebellum, and premotor cortex. Moreover, with increasing task complexity, other areas appear to be recruited in anterior cingulate, insula, precuneus, intraparietal sulcus, basal ganglia, ventrolateral cortex, and thalamus.

A recent study of attentional states elicited by musical performance suggest that with the use of EEG-based feedback training, musicians can improve the musical quality of their performances (Egner & Gruzelier, 2003). In training sessions prior to musical performance, pianists learned to increase the theta over alpha band amplitudes in their EEG. The enhancements in quality of musical performance appear to be a consequence of a deep relaxed focus of attention, and may not be due to mere reductions in anxiety, as other methods of relaxation training reduced anxiety but did not affect the quality of musical performance.

The brain basis of musical performance per se has been studied in the context of singing or piano playing. One PET investigation studied non-musicians singing simple monotone sequences using a vowel (Perry et al., 1999) and a second fMRI study examined non-musicians overtly or covertly singing a familiar melody without words (Riecker, Ackermann, Wildergruber, Dogil, & Grodd, 2000). A more recent investigation employed PET to examine amateur musicians who performed ‘listen and respond’ tasks in which they either sang back repetitions of novel melodies, sang back harmonizations to accompany novel melodies, or vocalized monotonically in response (Brown, Parsons, Martinez, Hodges, & Fox, in press). Across these three studies, major singing-specific activations were observed in primary and secondary auditory cortices, primary motor cortex, frontal operculum, SMA, insula, posterior cerebellum, and basal ganglia. However, in the last study, melody singing and harmonization, but not monotonic vocalization, activated planum polare (Brodmann area (BA) 38), implicating it as an area supporting higher musical representations.

In an early, concerted effort to study the brain basis of musical performance (Sergent et al., 1992), pianists were scanned with PET as they performed several conditions involving listening to scales, playing scales, sight-reading a score, and sight-reading a score while playing it. The pianists always heard the sounds they produced on the piano, and they sight-read or played an obscure partita by J.S. Bach. When the pianists played this piece, they used only their right hand. As control conditions, the researchers included a task requiring a manual response to indicate the location of a series of single visual dots within quadrants on a screen and a task with fixation point rest. Of most interest at present is the pattern of activations detected for sight-reading when playing (and hearing) the partita, as contrasted with a combination of the sight-reading and listening conditions. This analysis revealed increases in left frontal operculum (BA 44) that were interpreted to support the patterning of motor sequences of the right hand. There was also activity in left parietal cortex (supramarginal gyrus, BA 40), possibly involved in mapping visual and auditory representations of the melody. There were activations in left occipitoparietal sulcus and bilateral superior parietal cortices (BA 7), which might subserve sensorimotor transformations required for visually guided finger positioning.

Aspects of this early research have been pursued further in a recent study that used fMRI to compare the right hand performance while sightreading the score of a Bartok piano piece to its imagined simulation (Meister et al., 2004). In the baseline control the pianists read score with a single note repeated. Comparing the actual performance to control, the authors observed activations in primarily left sensorimotor areas (BA 2–4), left SMA (BA 6), bilateral precuneus (BA 7), bilateral inferior parietal (BA 40), left occipital (BA 37), left BA 5 (parietal), left posterior cerebellum (VI), midline anterior cerebellum (V), and left thalamus. Several of these activations confirm those in the early PET study (Sergent et al., 1992). Comparing the imagined simulation to control, they observed bilateral superior premotor (BA 6), left frontal (BA 9), bilateral parietal (BA 40 and 7), bilateral occipital (BA 18 and 19), and left posterior cerebellum (VI). A direct comparison between performed and imagined performance revealed performance-specific activations in left sensorimotor (BA 4, 2 and 3), left SMA (BA 6), bilateral inferior parietal (BA 40), right anterior cerebellum (III), and left posterior cerebellum (VI). Activation specific to imagined performance was limited to left occipital (BA 19). These findings are in accord with prior research, such as that by Jeannerod, 1994, Jeannerod, 1997, which demonstrated that motor and sensory imagery involves psychological and neural processes similar to those for real motor and sensory experiences (on auditory and musical imagery, see Halpern & Zatorre, 1999; Janata, 2001; Reiser, 1992).

Imagined musical performance was also examined in another recent study (Langheim, Callicott, Mattay, Duyn, & Weinberger, 2002). This study combined fMRI data from imagined performance for different instruments (cello, violin, piano) and different memorized compositions (various Vivaldi or Bach pieces) in order to localize common, music-specific areas. Overall, the imagined musical performance (compared to rest) engaged right SMA, right superior premotor cortex (BA 6), right superior parietal lobule (BA 7), right inferior frontal gyrus (BA 47/45), left thalamus, left basal ganglia (caudate), and bilateral posterior cerebellum (VI). This hemodynamic pattern was distinct from that for passive listening to musical pieces and for a self-paced bimanual, finger-tapping task (both compared to rest). Thus, the results were taken to suggest that the foregoing areas are involved in representing information for performing music. It is notable that neither of the two studies just described of imagined musical performance observed activations in temporal cortical regions that support auditory and musical information.

In sum, apart from brain areas for sensory-motor, attention, and executive control processes, three brain regions have been identified so far that appear to be important for higher-level information processing aspects of music performance. One area is the frontal operculum, which can be activated (left, right, or bilaterally) by sight-read piano performance, by music singing, and by imagined string and piano performance of memorized music. This area is often interpreted to be involved in sequence production and imitation learning. A second area is the planum polare, which can be activated (either right or bilaterally) by singing and by sight-read piano performance, but not apparently by imagined musical singing or imagined string or piano playing. This area appears to represent musical representations of a higher order than, for example, is present in more posterior superior temporal cortices (BA 22). A third region of interest in higher-level music is in rostromedial prefrontal cortex, which responds to dissonance and consonance, and to changes in tonality (Blood, Zatorre, Bermudes, & Evans, 1999; Janata et al., 2002; Peretz, Blood, Penhune, & Zatorre, 2001); however, this area has not yet been studied in the context of musical performance.

The goal of the present study was to focus directly on musical performance as such, in order to complement and clarify the foregoing findings. PET was employed to delineate brain areas subserving bimanual piano performance of memorized music. This provides new information relative to prior studies since by recording brain activity when both hands were equally and concurrently producing music, we examined neural systems when both cerebral hemispheres were fully involved in a performance of a natural kind for musicians. This approach goes beyond the Sergent et al. (1992) and Meister et al. (2004) studies in which only the right hand was used to perform the music, a design that left unclear which particular right hemispheric areas may be involved in music performance as such.

In addition, by asking pianists to perform a memorized composition, our design eliminated musical score reading from scanned task performance. In this respect, we examined brain activation during a more purely musical performance. Sight-reading a score during performance adds a considerable cognitive load, one unrelated directly to music performance and cognition per se. Indeed, there is a common belief amongst musicians that a fully memorized piece, one performed without score reading, engenders a distinctly deeper understanding of the composition and more satisfying realization of the piece in performance (Aiello, 2001; Chaffin & Imreh, 2002; Chaffin, Imreh, & Crawford, 2002; Mach, 1998). This belief is congruent with a significant role of a deep focus of attention in the quality of musical performance, as discussed earlier.

Our design referenced the brain activity during the piano performance of a musical composition by J.S. Bach to that during the two-handed performance of scales. The Bach and scales performances required movements of approximately comparable frequency and complexity from each hand. This is a more comparable control contrast than in the Langheim et al. (2002) study for (imagined) memorized musical performance. In this design, real perceived musical sounds and similar executed motor behavior are compared across tasks of varying musical structure to isolate the neural substrates of musical performance. In the Langheim et al. study, imagined sounds and movements were compared to real ones, and the imagined motor behaviors were very different than the control motor behavior (e.g., playing a cello piece versus finger tapping). Nonetheless, it was recognized that studying a natural kind of musical performance, as compared to scales, entailed a number of factors varying to influence brain activity, apart for musicality. Thus, there were intrinsic differences in fingering complexity, independence of hands and melodic lines, complexity of memorized information to recall, emotional content, and attentional demands. The outline of interactions observed here amongst these factors sets the stage for more detailed, parametrically controlled studies of high-level performance skills.

Section snippets

Participants

After giving informed consent, eight professional musicians (five females and three males) volunteered to participate in this study. All volunteers were right handed (Oldfield, 1971) and ranged from 27 to 54 years of age. Each individual had from 14 to 20 years of training in piano performance, in addition to 10–18 years of training and education on other instruments (either horn, voice, or string) and on other aspects of music (composition and education).

Stimuli and tasks

Prior to the scanning session, the

Results

The performance of scales, when contrasted with rest, activated an array of motor, somatosensory, and auditory functional areas, among others (Table 1 and Fig. 3). Responses were observed in bilateral primary motor cortex for hand (BA 4), left SMA, bilateral predominantly left insula (BA 13), bilateral dorsolateral premotor cortex (BA 6), right motor cingulate, right pulvinar, left lateral globus pallidus, and right red nucleus. There were also increases in bilateral superior temporal gyrus (BA

Discussion

These data offer an initial glimpse of brain areas engaged when an expert pianist plays a memorized musical composition. Psychological study of the process of memorizing the composition performed in this study has characterized the nature of the extensive preparation for performance, particularly the role of conceptual, auditory, and motor memory processes (Chaffin & Imreh, 2002; Chaffin et al., 2002). Our discussion first considers the functional brain data common to the performance of scales

Acknowledgments

We are grateful to Michael Martinez for expert assistance with data analysis and figures, and to an anonymous reviewer for very helpful suggestions and comments.

References (118)

  • I.G. Meister et al.

    Playing piano in the mind—An fMRI study on music imagery and performance in pianists

    Cognitive Brain Research

    (2004)
  • R. Oldfield

    The assessment and analysis of handedness: The Edinburgh inventory

    Neuropsychologia

    (1971)
  • P. Redgrave et al.

    The basal ganglia: A vertebrate solution to the selection problem?

    Neuroscience

    (1999)
  • S. Samson et al.

    Melodic and harmonic discrimination following unilateral cerebral excision

    Brain and Cognition

    (1988)
  • M. Satoh et al.

    Activated brain regions in musicians during an ensemble: A PET study

    Cognitive Brain Research

    (2001)
  • J.D. Schmahmann et al.

    Three-dimensional MRI atlas of the human cerebellum in proportional stereotaxic space

    NeuroImage

    (1999)
  • R.J. Schneider et al.

    A modality-specific somato-sensory area within the insula of the rhesus monkey

    Brain Research

    (1993)
  • M. Schurmann et al.

    Mind's ear in a musician: Where and when in the brain

    NeuroImage

    (2002)
  • R. Aiello

    Playing piano by heart: From behavior to cognition

  • M. Bentovoglio et al.

    Limbic thalamus: Structure, intrinsic organization, and connections

  • D.L. Bihan et al.

    Activation of human primary visual cortex during visual recall: A magnetic resonance imaging study

    Proceedings of the National Academy of Sciences

    (1993)
  • A.J. Blood et al.

    Intensely pleasurable responses to music correlate with activity in brain regions implicated in reward and emotion

    Proceedings of the National Academy of Science

    (2001)
  • A.J. Blood et al.

    Emotional responses to pleasant and unpleasant music correlate with activity in paralimbic brain regions

    Nature Neuroscience

    (1999)
  • J.M. Bower

    Control of sensory data acquisition

  • J.M. Bower et al.

    Rethinking the lesser brain

    Scientific American

    (2003)
  • S. Brown et al.

    Passive music listening spontaneously engages limbic and paralimbic systems

    NeuroReport

    (2004)
  • Brown, S., Parsons, L. M., Martinez, M. J., Hodges, D., & Fox, P. T. (in press). The song system of the human brain....
  • M. Cappelletti et al.

    A selective loss of the ability to read and write music

    Neurocase

    (2000)
  • S.T. Carmichael et al.

    Limbic connections of the orbital and medial prefrontal cortex in macaque monkeys

    Journal of Comparative Neurology

    (1995)
  • A.F. Carpenter et al.

    Motor control encoding of serial order in a context-recall task

    Science

    (1999)
  • R. Chaffin et al.

    Practicing perfection: Piano performance as expert memory

    Psychological Science

    (2002)
  • R. Chaffin et al.

    Practicing perfection: Memory and piano performance

    (2002)
  • D.J. Crammond et al.

    Differential relation of discharge in primary motor cortex and premotor cortex to movement versus actively maintained postures during a reaching task

    Experimental Brain Research

    (1996)
  • M. Csikszentimihalyi et al.

    Optimal experience: Psychological studies of flow in consciousness

    (1988)
  • R.B. D’Agostino et al.

    A suggestion for using powerful and informative tests of normality

    American Statistician

    (1990)
  • J.-R. Duhamel et al.

    Ventral intraparietal area of the macaque: Congruent visual and somatic response properties

    Journal of Neurophysiology

    (1998)
  • T. Egner et al.

    Ecological validity of neurofeedback: Modulation of slow wave EEG enhances musical performance

    NeuroReport

    (2003)
  • T. Elbert et al.

    Increased cortical representation of the fingers of the left hand in string players

    Science

    (1995)
  • S. Evers et al.

    The cerebral hemodynamics of music perception: A trancranial Doppler sonography study

    Brain

    (1999)
  • I. Faillenot et al.

    Visual pathways for object-oriented action and object recognition, functional anatomy with PET

    Cerebral Cortex

    (1997)
  • J.A. Fiez et al.

    Impaired nonmotor learning and error-detection associated with cerebellar damage—A single case study

    Brain

    (1992)
  • P.T. Fox et al.

    Noninvasive functional brain mapping by change-distribution analysis of averaged PET images of H215O tissue activity

    Journal of Nuclear Medicine

    (1989)
  • P.T. Fox et al.

    Enhanced detection of focal brain responses using inter-subject averaging and change-distribution analysis of subtracted PET images

    Journal of Cerebral Blood Flow and Metabolism

    (1988)
  • K.J. Friston et al.

    Comparing functional (PET) images: The assessment of significant change

    Journal of Cerebral Blood Flow and Metabolism

    (1991)
  • C. Gaser et al.

    Brain structures differ between musicians and non-musicians

    Journal of Neuroscience

    (2003)
  • P.M. Grasby et al.

    Functional mapping of brain areas implicated in auditory-verbal memory function

    Brain

    (1993)
  • M.S.A. Graziano et al.

    The cortical control of movement revisited

    Neuron

    (2002)
  • T.D. Griffiths et al.

    Analysis of temporal structure in sound by the human brain

    Nature Neuroscience

    (1998)
  • T.D. Griffiths et al.

    A common neural substrate for the analysis of pitch and duration pattern in segmented sound?

    NeuroReport

    (1999)
  • R.W. Guillery

    Branching thalamic afferents link action and perception

    Journal of Neurophysiology

    (2003)
  • Cited by (92)

    • Perceptual and motor laterality effects in pianists during music sight-reading

      2015, Neuropsychologia
      Citation Excerpt :

      In particular the superior parietal lobe seemed to be involved in music reading but not in verbal performance. In another PET study, Parsons et al. (2005) uncovered that musical performance involves an extended set of brain areas depending upon the requested grade of difficulty (e.g. performance of scales or concertos). In a cortical stimulation study on musicians with neurological lesions, Roux et al. (2007) showed that brain areas supporting music reading can be occasionally distinct from language reading areas and are located mainly in the “dominant” parietal lobe (supramarginal gyrus or intraparietal fissure) and in all of the frontal gyri.

    • The Musical Brain: What Students, Teachers, and Performers Need to Know

      2023, The Musical Brain: What Students, Teachers, and Performers Need to Know
    View all citing articles on Scopus
    View full text