The musician's brain as a model of neuroplasticity

PERSPECTIVESOPINIONThe musician’s brain as a modelof neuroplasticityThomas F. Münte, Eckart Altenmüller and Lutz JänckeStudies of experience-driven neuroplasticityat the behavioural, ensemble, cellular andmolecular levels have shown that thestructure and significance of the elicitingstimulus can determine the neural changesthat result. Studying such effects in humansis difficult, but professional musiciansrepresent an ideal model in which toinvestigate plastic changes in the humanbrain. There are two advantages tostudying plasticity in musicians: thecomplexity of the eliciting stimulus — music— and the extent of their exposure to thisstimulus. Here, we focus on the functionaland anatomical differences that have beendetected in musicians by modernneuroimaging methods.related plasticity, as is its behavioural significance3,4 . Animal research has also revealedneuroplasticity at the molecular, synaptic andmacroscopic structural levels 5,6 . Although animalmodels are useful for studying the cellularand molecular mechanisms of plasticity, thetypical laboratory animal is deprived of normalstimulation and might, therefore, be aspecial case. Moreover, animal models are limitedin the range of stimuli that are used, in thebehavioural manipulations that are associatedwith these stimuli and in the duration oftraining. In addition, it is far from clear howthe mechanisms that govern synaptic plasticityat the cellular level are related to theflexibility of operations seen for large-scaleneuronal networks on the one hand, andcognitive processes on the other.It is therefore important to extend theseinvestigations to the human brain. Significantheadway has been made by studying intermodalplasticity in congenitally blind 7 or deafsubjects 8,9 , or by monitoring the effects oflimb amputations 10 . In this article, however,we are concerned with findings in professionalmusicians that have been describedover the past decade or so. Performing musicat a professional level is arguably among themost complex of human accomplishments. Apianist, for example, has to bimanually coordinatethe production of up to 1,800 notesper minute (FIG. 1). Music, as a sensory stimulus,is highly complex and structured in severaldimensions 11 , so it extends beyond any ofthe stimuli that have been used in animalresearch. Moreover, making music requiresThe size and temporal organization of corticalrepresentations of stimuli are continuallyshaped by experience 1,2 . Animal studies overthe past 20 years have gone a long way towardsexplaining some of the rules of cortical plasticity.For example, it has been shown that trainingto make fine-grained temporal judgementsyields an expansion of the bandwidthor receptive field in both the auditory andsomatosensory modalities, whereas tasks thatrequire fine-grained frequency or spatial tactilediscrimination lead to a decrease in thereceptive-field size of cortical neurons 1,3 . Thiseffect has been explained by Hebbian learningrules, whereby synapses are driven to changeby temporally coherent inputs in a competitiveneural network. Attention to the sensoryinput is very important in driving experience-3 secondsFigure 1 | Example of a hypercomplex musical score. Two three-second segments of the 11thvariation from the 6th Paganini-Etude by Franz Liszt. The depicted segments require the production of1,800 notes per minute. Reproduced by kind permission of Peters Edition Ltd.NATURE REVIEWS | NEUROSCIENCE VOLUME 3 | JUNE 2002 | 473© 2002 Nature Publishing Group

PERSPECTIVESthe integration of sensory and motor information,and precise monitoring of performance.Finally, the study of musicians mightallow us to tease apart the effects of musicaltraining or experience from those of geneticpredisposition.So, the musician’s brain might constitute aperfect model in which to study neuroplasticityin the auditory and motor domains. It can alsobe used to examine the effects of dysfunctionalplasticity, as illustrated by musician’s cramp, aparticular kind of occupational dystonia 12,13 .Functional measures of plasticityIn a seminal study in 1995, Elbert and colleagues14 investigated somatosensory evokedmagnetic fields in string players. Sourceanalysis revealed that the cortical representationof the digits of the left hand (the fingeringhand) was larger in these musicians thanin controls. In the case of the right hand, inwhich no independent movements of the fingersare required in string players, there wereno differences between musicians and controls.The cortical reorganization of the representationof the fingers was more pronouncedin musicians who had begun their musicaltraining at an early age. A first indication thatextensive musical training can plastically alterreceptive functions was provided by Pantevand colleagues 15 . Equivalent current dipoles,computed from evoked magnetic fields, wereobtained in response to piano tones and topure tones of equal fundamental frequencyand loudness. In musicians, the responses topiano tones (but not to pure tones) were~25% larger than in non-musicians. In astudy of violinists and trumpeters, this effectwas most pronounced for tones from eachmusician’s own type of instrument 16 .Although musical sounds were used inthese early studies, they lacked a main characteristicof music: tones become music onlywhen they are structured. The temporalstructure of music comprises local rhythm,defined by the duration and temporal distanceof tones, and global metrum, whichdetermines whether a piece is, for example, amarch or a waltz. As far as the pitch dimensionis concerned, the interval between twosuccessive notes, the global contour of a musicalpiece and the harmonic structure can bedistinguished. In processing (and enjoying)music, the extraction of structural regularitiesis of paramount importance. At what cognitiveand neural levels are such regularities ofmusic processed? The mismatch negativity(MMN), a frontal negative wave in the eventrelatedpotential (ERP), is a marker of thepre-attentive detection of changes in regularsequences of auditory stimuli 17 . Consider, forFigure 2 | Structural changes in the brains ofmusicians. Some of the brain areas that havebeen found to be enlarged in musicians inmorphometric studies based on structuralmagnetic resonance imaging. Red, primary motorcortex; yellow, planum temporale; orange, anteriorpart of the corpus callosum.example, a series of 1,200-Hz tones that isinterrupted occasionally by a deviant 1,500-Hztone. In this situation, the deviant tones giverise to an MMN. Importantly, the MMNoccurs in the absence of attention to the stimuli,and newer research has indicated that, inaddition to simple physical deviants (such asduration, pitch or intensity), the MMN is alsoelicited by more complex irregularities, suchas changes in sequences of several tones 17 .Professional musicians, unlike non-musicians,show an MMN for tones that aremistimed by as little as 20 ms in a series ofregularly spaced tones 18 . For stimuli that aremistimed by 50 ms, which do produce anMMN in non-musicians, the MMN in musicianswas considerably larger than that of controls18 . Musicians also showed an MMN forslightly impure chords that were presentedamong perfect major chords, whereas, again,non-musicians did not 19 . Recently, an MMNwas found for small changes in the contour oftransposed melodies in musicians who performprimarily without a score 20 . Source localizationstudies have shown that the MMNarises mainly from neurons on the supratemporalplain of the temporal lobe, withfurther contributions from the frontal cortex17,21,22 . These findings indicate that, afteryears of musical training, neuronal populationsin the auditory cortex might be shapedsuch that they automatically detect subtlechanges in auditory stimulus sequences withsimple or higher-order regularities. The parametersthat are needed for the acquisition ofthese skills are unknown, but probably involveinitial attentive processing of the stimuli 20 .In an experiment that required theattentive analysis of chord sequences, the N1component of the ERP, which arises from theprimary auditory cortex, was enhanced inresponse to consonant chords, relative to dissonant(augmented fifth) chords, in musiciansonly 23 . This indicates that the auditorycortex, possibly including the primary areas,in musicians might be tuned to complexharmonic features of sounds 24 .Superiorattention-dependent processing in musicianswas also shown in a study that required thesubjects to select stimuli from one of twoinformation channels that were defined bytheir pitch. In such situations, stimuli that areattended elicit a negativity in the event-relatedpotential, known as the Nd, relative to stimulithat are not the focus of attention 25 . In professionalmusicians, the Nd was more pronounced,indicating superior attentionalselectivity in the pitch dimension 26 .A recurring theme in the animal literatureis the plasticity of sound localization. Interauraldifferences in the level and phase ofsounds are important for sound localization,but the filter functions of the outer ear,known as head-related transfer functions 27 ,also provide crucial information. A conductor,more than any other musician, is likely todepend on spatial localization for successfulperformance — for example, he might needto home in on a certain player. Consistentwith this, professional conductors were foundto be better than pianists and non-musiciansat separating adjacent sound sources, oneof which was task relevant, in the periphery ofthe auditory field. This behavioural selectivitywas paralleled by modulation of the Nd component,which was selective for the attendedsource in conductors, but not in pianists ornon-musicians 28 .Anatomical differencesSince the age of phrenology, neuroscientistshave tried to relate extraordinary skills tochanges in brain anatomy. For example, atthe beginning of the twentieth century,Auerbach reported that the middle and posteriorthirds of the superior temporal gyruswere larger than normal in several postmortemstudies of the brains of famousmusicians 29 . Modern brain-imaging techniques,such as high-resolution magnetic resonanceimaging (MRI), allow us to studyanatomical details in the brains of livinghumans. Studies in which these techniqueshave been used have shown that several brainareas, including the planum temporale, theanterior corpus callosum, the primary handmotor area and the cerebellum, differ in theirstructure and size between musicians andcontrol subjects (FIG. 2). These findings havereopened the debate about whether thesestructural differences are related directly tomusical ability.474 | JUNE 2002 | VOLUME 3 www.nature.com/reviews/neuro© 2002 Nature Publishing Group

PERSPECTIVESwhich is comparable to the oral–aural loop inlanguage processing — is established duringmusical training 48 . At the behavioural level, itis reflected by reports from professionalpianists that their fingers move more-or-lessautomatically when they are listening to pianomusic. Exposing musically naive subjects tocontrolled piano training led to audio-motorcoupling after as little as 20 minutes of training,as shown by topographic analysis of veryslow event-related potentials 48 . After the firsttraining session, there was further activity overmotor areas while subjects listened to simplepiano tunes. Likewise, finger movements on amute keyboard were associated with anincrease in activity over auditory areas. Theeffect could be enhanced and stabilized duringfive weeks of training (FIG. 3). Similar co-activationhas also been found in professionalpianists, who showed magnetoencephalographic(MEG) activity in sensorimotorListening20 years0.9820 days0.9520 minutes0.83Start0.78MovingnegposFigure 3 | Sensorimotor integration inmusicians. Shown are spline-interpolatedisovoltage maps of averaged electroencephalographicrecordings. Subjects eitherlistened to a short piece of piano music (leftcolumn) or performed on a mute piano keyboard(right column). Before training, musically naivesubjects (Start) produced grossly differenttopographic patterns of slow event-relatedpotentials in the two conditions, as indicated bythe maps and a measure of map similarity (scalarproduct). With training for 20 minutes or 20 days,the topographic distributions became increasinglysimilar. In professional pianists with approximately20 years of training, maps in both conditions arevirtually identical. Modified, with permission, fromREF. 48 © 2001 New York Academy of Sciences.cortical regions while listening to pianomusic 49 . These neural networks thus seem tobehave in a similar way to the ‘mirror neurons’in the monkey frontal cortex (area F5), whichare active during both the execution of complexmovements and visual observation of thesame movements 50 .Maladaptive plasticityThere is a dark side to the increasing specializationand prolonged training of modernmusicians — namely, loss of control anddegradation of skilled hand movements, adisorder referred to as ‘musicians’ cramp’ orfocal dystonia 12 . The first historical record ofthis condition, from 1830, appears in thediaries of the famous pianist and composerRobert Schumann 51 . As was probably the casefor Schumann, prolonged practice and painsyndromes due to overuse can precipitate dystonia,which is developed by ~1% of professionalmusicians and usually ends theircareers 12 . Neuroimaging studies point to dysfunctional(or maladaptive) neuroplasticity asits cause 13,52 . For example, an MEG study ofmusicians with focal dystonia showed fusionof the digital representations in the somatosensorycortex, reflected in a decreased distancebetween the representations of theindex and little fingers relative to healthy controlmusicians 13 (FIG. 4). These findings arecorroborated by psychophysical measurementsand fMRI investigations in a relateddisorder, writer’s cramp, which showeddecreased temporal and spatial discriminationat the finger tips 53,54 . Observations inmonkeys indicate that rapid, stereotypicalmovements in a learning context can activelydegrade the cortical representations of sensoryinformation that guide fine hand movements55 . This dedifferentiation of sensoryfeedback information has been proposed toform the basis of focal dystonia 47,55 . Indeed,the repeated temporal association of movementpatterns is a characteristic of music —for example, when playing arpeggios or musicalscales. In a further study, symptoms wereprovoked in five dystonic guitarists when theyplayed a modified guitar inside an fMRI scanner56 . Relative to non-dystonic guitarists, theyshowed more activation of the contralateralsensorimotor cortex but less activation of premotorareas, indicating abnormal recruitmentof cortical areas that are involved in thecontrol of complex movements.Musicians as a model?Neuroplasticity allows the brain to adapt toenvironmental factors that cannot be anticipatedby genetic programming. The neuraland behavioural changes that are attributedto plasticity have been observed on differenttimescales, ranging from several minutes tothe whole lifetime of the individual.Different processes are likely to supportplastic changes at the extremes of this timeline.Accordingly, experience-driven neuroplasticityhas been explained by both thede novo growth and improvement of newdendrites, synapses and neurons 5,57 , and thedisinhibition or inhibition of pre-existinglateral connections between neurons bysensory input 58 . The former mechanismentails structural changes at the microscopicand macroscopic levels, whereas the lattercan be achieved by strengthening or inhibitingpre-existing synaptic connections in thespirit of Hebbian learning. Sometimes, evenmore rapid changes of brain responses,occurring in the order of milliseconds, havebeen discussed under the heading of neuroplasticity.However, these are likely to resultfrom the attentional modulation of neuralcircuits, and should be distinguished fromtrue plastic changes 59 .Research into plasticity in musicians is still inits infancy, but, already, many of the findingsfrom animal studies have found parallels instudies of musicians. At one extreme, years ofmusical experience, especially in those musicianswho begin training early on, might lead toan increase in grey and white matter volume inseveral brain regions 31,32,36,39,41 . These anatomicalalterations seem to be confined to a criticalperiod. The fact that, in several studies, a correlationwas found between the extent of theanatomical differences and the age at whichmusical training started strongly argues againstthe possibility that these differences are preexistingand the cause, rather than the result, ofpractising music. The view that these differencesrepresent genuine plastic changes of the brainreceives further support from neuroimagingstudies in other populations. For example, acorrelation between the size of the posteriorhippocampus and years of driving experiencehas been reported in London taxi drivers 60 .Further research using advanced imaging techniques,such as magnetic resonance spectroscopyand diffusion tensor imaging, and theextension of studies beyond the conventionalcross-sectional design, are needed to investigatethe underlying neurophysiological changes.At the other extreme, several minutes oftraining can induce changes in the recruitmentof areas of the motor cortex 44 or establishauditory–sensorimotor coupling 48 .Some of the other findings discussed hereprobably require training in the order ofmonths to several years and, at present, it isunclear what neural processes support thisbehavioural plasticity.476 | JUNE 2002 | VOLUME 3 www.nature.com/reviews/neuro© 2002 Nature Publishing Group

PERSPECTIVESD5D2–D4D1Non-dystonic handThe investigations presented in thisoverview convincingly show the value of themusician’s brain as a model of neuroplasticityand have set the stage for furtherresearch. Some of the questions that need tobe tackled include the following. What arethe training parameters that lead to successfullearning and plasticity? Can these parametersbe exploited in musical education orto enhance learning in other domains? Whatis the role of genes in determining auditoryneuroplasticity in musicians? What is therange of structural regularities that can beextracted from the auditory input in anautomatic, pre-attentive fashion? As makingmusic undoubtedly requires intense selfmonitoring,and error detection and correction,are there any plastic changes in theexecutive brain systems that are responsiblefor performance monitoring?Finally, one has to bear in mind that musiccan elicit powerful emotional reactions.Strong emotional responses to music, leadingAffected digitsD3–D5RightLeftFigure 4 | Fusion of the somatosensory representation of single digits of the hand in a musiciansuffering from focal dystonia. The best-fitting dipoles to explain the evoked magnetic fields aftersensory stimulation of single digits (D1–D5) are shown projected on the individual’s magnetic resonanceimaging scan. Whereas for the non-affected hand, the typical homuncular organization (inset) reveals adistance of ~2.5 cm between the sources for the thumb and the little finger (yellow circle and square onthe right of the brain), the somatosensory representations of the fingers on the dystonic side are blurred,resulting from a fusion of the neural networks that process incoming sensory stimuli from different fingers(red circles). Modified, with permission, from REF. 13 © 1998 Lippincott, Williams and Wilkins.D1D2to shivers down the spine and changes in heartrate, are accompanied by the activation of abrain network that includes the ventral striatum,midbrain, amygdala, orbitofrontal cortexand ventral medial prefrontal cortex — areasthat are thought to be involved in reward,emotion and motivation 61 . Further researchwill show whether activity in these areas is alsodirectly involved in mediating neuroplasticity.Thomas F. Münte is at the Department ofNeuropsychology, Otto-von-Guericke University,Universitätsplatz 2, Gebäude 24,39106 Magdeburg, Germany.Eckart Altenmüller is at the Institute of MusicPhysiology and Musicians’ Medicine, HannoverSchool of Music and Drama, Emmich Platz 1,30175 Hannover, Germany.Lutz Jäncke is at the Department ofNeuropsychology, University of Zürich, ZürichbergStrasse 43, CH-8044 Zürich, Switzerland.Correspondence to T.F.M.e-mail: thomas.muente@med.uni-magdeburg.dedoi:10.1038/nrn8431. Kilgard, M. P. et al. Sensory input directs spatial andtemporal plasticity in primary auditory cortex.J. Neurophysiol. 86, 326–338 (2001).2. Singer, W. Development and plasticity of corticalprocessing architectures. Science 270, 758–764(1995).3. Recanzone, G. H., Merzenich, M. M., Jenkins, W. M.,Grajski, K. A. & Dinse, H. R. Topographic reorganizationof the hand representation in cortical area 3b of owlmonkeys trained in a frequency-discrimination task.J. Neurophysiol. 67, 1031–1056 (1992).4. Ahissar, M. & Hochstein, S. Task difficulty and thespecificity of perceptual learning. Nature 387, 401–406(1997).5. Anderson, B. J. et al. Glial hypertrophy is associated withsynaptogenesis following motor-skill learning, but notwith angiogenesis following exercise. Glia 11, 73–80(1994).6. van Praag, H., Kempermann, G. & Gage, F. H. Neuralconsequences of environmental enrichment. Nature Rev.Neurosci. 1, 191–198 (2000).7. Röder, B. et al. Improved auditory spatial tuning in blindhumans. Nature 400, 162–166 (1999).8. Bavelier, D. et al. Visual attention to the periphery isenhanced in congenitally deaf individuals. J. Neurosci.20, RC93 (2000).9. Bavelier, D. & Neville, H. J. Cross-modal plasticity: whereand how? Nature Rev. Neurosci. 3, 443–452 (2002).10. Flor, H. et al. Phantom-limb pain as a perceptual correlateof cortical reorganization following arm amputation.Nature 375, 482–484 (1995).11. Schuppert, M., Münte, T. F., Wieringa, B. M. &Altenmüller, E. Receptive amusia: evidence for crosshemisphericneural networks underlying musicprocessing strategies. Brain 123, 546–559 (2000).12. Lim, V. K., Altenmüller, E. & Bradshaw, J. L. Focal dystonia:current theories. Hum. Mov. Sci. 20, 875–914 (2001).13. Elbert, T. et al. Alteration of digital representations insomatosensory cortex in focal hand dystonia.Neuroreport 9, 3571–3575 (1998).14. Elbert, T., Pantev, C., Wienbruch, C., Rockstroh, B. &Taub, E. Increased cortical representation of the fingers ofthe left hand in string players. Science 270, 305–307(1995).15. Pantev, C. et al. Increased auditory cortical representationin musicians. Nature 392, 811–814 (1998).16. Pantev, C., Roberts, L. E., Schulz, M., Engelien, A. &Ross, B. Timbre-specific enhancement of auditorycortical representations in musicians. Neuroreport 12,169–174 (2001).17. Picton, T. W., Alain, C., Otten, L., Ritter, W. & Achim, A.Mismatch negativity: different water in the same river.Audiol. Neurootol. 5, 111–139 (2000).18. Russeler, J., Altenmüller, E., Nager, W., Kohlmetz, C. &Münte, T. F. Event-related brain potentials to soundomissions differ in musicians and non-musicians.Neurosci. Lett. 308, 33–36 (2001).19. Koelsch, S., Schroger, E. & Tervaniemi, M. Superior preattentiveauditory processing in musicians. Neuroreport10, 1309–1313 (1999).20. Tervaniemi, M., Rytkonen, M., Schroger, E., Ilmoniemi,R. J. & Näätänen, R. Superior formation of corticalmemory traces for melodic patterns in musicians.Learn. Mem. 8, 295–300 (2001).21. Tiitinen, H. et al. Tonotopic auditory cortex and themagnetoencephalographic (MEG) equivalent of themismatch negativity. Psychophysiology 30, 537–540(1993).22. Tervaniemi, M. et al. Functional specialization of thehuman auditory cortex in processing phonetic andmusical sounds: a magnetoencephalographic (MEG)study. Neuroimage 9, 330–336 (1999).23. Regnault, P., Bigand, E. & Besson, M. Different brainmechanisms mediate sensitivity to sensory consonanceand harmonic context: evidence from auditory eventrelatedbrain potentials. J. Cogn. Neurosci. 13,241–255 (2001).24. Tramo, M. J., Cariani, P. A., Delgutte, B. & Braida, L. D.Neurobiological foundations for the theory of harmony inwestern tonal music. Ann. NY Acad. Sci. 930, 92–116(2001).25. Hillyard, S. A., Teder-Salejarvi, W. A. & Münte, T. F.Temporal dynamics of early perceptual processing.Curr. Opin. Neurobiol. 8, 202–210 (1998).26. Münte, T. F., Nager, W., Rosenthal, O., Johannes, S. &Altenmüller, E. in Integrated Human Brain Science (ed.Nakada, T.) 389–398 (Elsevier, Amsterdam, 2000).27. Musicant, A. D., Chan, J. C. & Hind, J. E. Directiondependentspectral properties of cat external ear: newdata and cross-species comparisons. J. Acoust. Soc.Am. 87, 757–781 (1990).NATURE REVIEWS | NEUROSCIENCE VOLUME 3 | JUNE 2002 | 477© 2002 Nature Publishing Group

PERSPECTIVES28. Münte, T. F., Kohlmetz, C., Nager, W. & Altenmüller, E.Neuroperception. Superior auditory spatial tuning inconductors. Nature 409, 580 (2001).29. Meyer, A. in Music and the Brain (eds Macdonald, C. &Henson, R. A.) 255–281 (Heinemann Medical Books,London, 1977).30. Jäncke, L., Schlaug, G., Huang, Y. & Steinmetz, H.Asymmetry of the planum parietale. Neuroreport 5,1161–1163 (1994).31. Schlaug, G., Jäncke, L., Huang, Y. & Steinmetz, H.In vivo evidence of structural brain asymmetry inmusicians. Science 267, 699–701 (1995).32. Keenan, J. P., Thangaraj, V., Halpern, A. R. & Schlaug, G.Absolute pitch and planum temporale. Neuroimage 14,1402–1408 (2001).33. Zatorre, R. J., Perry, D. W., Beckett, C. A., Westbury, C.F. & Evans, A. C. Functional anatomy of musicalprocessing in listeners with absolute pitch and relativepitch. Proc. Natl Acad. Sci. USA 95, 3172–3177 (1998).34. Amunts, K. et al. Hand skills covary with the size of motorcortex: a macrostructural adaptation. Hum. Brain Mapp.5, 206–215 (1997).35. Jäncke, L., Schlaug, G. & Steinmetz, H. Hand skillasymmetry in professional musicians. Brain Cogn. 34,424–432 (1997).36. Schlaug, G., Jäncke, L., Huang, Y., Staiger, J. F. &Steinmetz, H. Increased corpus callosum size inmusicians. Neuropsychologia 33, 1047–1055 (1995).37. Aboitiz, F., Scheibel, A. B., Fisher, R. S. & Zaidel, E. Fibercomposition of the human corpus callosum. Brain Res.598, 143–153 (1992).38. Ridding, M. C., Brouwer, B. & Nordstrom, M. A. Reducedinterhemispheric inhibition in musicians. Exp. Brain Res.133, 249–253 (2000).39. Schlaug, G. The brain of musicians. A model forfunctional and structural adaptation. Ann. NY Acad. Sci.930, 281–299 (2001).40. Ashburner, J. & Friston, K. J. Voxel-based morphometry— the methods. Neuroimage 11, 805–821 (2002).41. Gaser, C. & Schlaug, G. Brain structures differ betweenmusicians and non-musicians. Neuroimage 13, 1168(2001).42. Karni, A. et al. The acquisition of skilled motorperformance: fast and slow experience-driven changes inprimary motor cortex. Proc. Natl Acad. Sci. USA 95,861–868 (1998).43. Karni, A. et al. Functional MRI evidence for adult motorcortex plasticity during motor skill learning. Nature 377,155–158 (1995).44. Hund-Georgiadis, M. & von Cramon, D. Y. Motor-learningrelatedchanges in piano players and non-musiciansrevealed by functional magnetic-resonance signals.Exp. Brain Res. 125, 417–425 (1999).45. Jäncke, L., Shah, N. J. & Peters, M. Cortical activationsin primary and secondary motor areas for complexbimanual movements in professional pianists. Brain Res.Cogn. Brain Res. 10, 177–183 (2000).46. Krings, T. et al. Cortical activation patterns duringcomplex motor tasks in piano players and controlsubjects. A functional magnetic resonance imagingstudy. Neurosci. Lett. 278, 189–193 (2000).47. Pascual-Leone, A. The brain that plays music and ischanged by it. Ann. NY Acad. Sci. 930, 315–329 (2001).48. Bangert, M., Haeusler, U. & Altenmüller, E. On practice:how the brain connects piano keys and piano sounds.Ann. NY Acad. Sci. 930, 425–428 (2001).49. Haueisen, J. & Knosche, T. R. Involuntary motor activity inpianists evoked by music perception. J. Cogn. Neurosci.13, 786–792 (2001).50. Umilta, M. A. et al. I know what you are doing. Aneurophysiological study. Neuron 31, 155–165 (2001).51. Schumann, R. Tagebücher, Band 1 (Stroemfeld RoterStern, Basel, 1971).52. Byl, N. N., McKenzie, A. & Nagarajan, S. S. Differences insomatosensory hand organization in a healthy flutist anda flutist with focal hand dystonia: a case report. J. HandTher. 13, 302–309 (2000).53. Sanger, T. D., Tarsy, D. & Pascual-Leone, A. Abnormalitiesof spatial and temporal sensory discrimination in writer’scramp. Mov. Disord. 16, 94–99 (2001).54. Sanger, T. D., Pascual-Leone, A., Tarsy, D. & Schlaug, G.Nonlinear sensory cortex response to simultaneous tactilestimuli in writer’s cramp. Mov. Disord. 17, 105–111 (2002).55. Byl, N. N., Merzenich, M. M. & Jenkins, W. M. A primategenesis model of focal dystonia and repetitive straininjury. I. Learning-induced dedifferentiation of therepresentation of the hand in the primary somatosensorycortex in adult monkeys. Neurology 47, 508–520 (1996).56. Pujol, J. et al. Brain cortical activation during guitarinducedhand dystonia studied by functional MRI.Neuroimage 12, 257–267 (2000).57. Polat, U. & Sagi, D. Spatial interactions in human vision:from near to far via experience-dependent cascades ofconnections. Proc. Natl Acad. Sci. USA 91, 1206–1209(1994).58. Jacobs, K. M. & Donoghue, J. P. Reshaping the corticalmotor map by unmasking latent intracorticalconnections. Science 251, 944–947 (1991).59. Noppeney, U., Waberski, T. D., Gobbele, R. & Buchner,H. Spatial attention modulates the corticalsomatosensory representation of the digits in humans.Neuroreport 10, 3137–3141 (1999).60. Maguire, E. A. et al. Navigation-related structural changein the hippocampi of taxi drivers. Proc. Natl Acad. Sci.USA 97, 4398–4403 (2000).61. Blood, A. J. & Zatorre, R. J. Intensely pleasurableresponses to music correlate with activity in brain regionsimplicated in reward and emotion. Proc. Natl Acad. Sci.USA 98, 11818–11823 (2001).62. Profita, J. & Bidder, T. G. Perfect pitch. Am. J. Med.Genet. 29, 763–771 (1988).63. Baharloo, S., Johnston, P. A., Service, S. K., Gitschier, J.& Freimer, N. B. Absolute pitch: an approach foridentification of genetic and nongenetic components.Am. J. Hum. Genet. 62, 224–231 (1998).64. Revesz, G. Introduction to the Psychology of Music(Longmans Green, London, 1953).65. Baharloo, S., Service, S. K., Risch, N., Gitschier, J. &Freimer, N. B. Familial aggregation of absolute pitch.Am. J. Hum. Genet. 67, 755–758 (2000).OPINION66. Gregersen, P. K., Kowalsky, E., Kohn, N. & Marvin, E. W.Absolute pitch: prevalence, ethnic variation, andestimation of the genetic component. Am. J. Hum.Genet. 65, 911–913 (1999).67. Ohnishi, T. et al. Functional anatomy of musical perceptionin musicians. Cereb. Cortex 11, 754–760 (2001).68. Crummer, G. C., Walton, J. P., Wayman, J. W., Hantz,E. C. & Frisina, R. D. Neural processing of musical timbreby musicians, nonmusicians, and musicians possessingabsolute pitch. J. Acoust. Soc. Am. 95, 2720–2727(1994).69. Klein, M., Coles, M. G. H. & Donchin, E. People withabsolute pitch process tones without producing a P300.Science 233, 1306–1309 (1984).Online linksFURTHER INFORMATIONEncyclopedia of Life Sciences: http://www.els.net/auditory processing | brain imaging: observing ongoing neuralactivity | cortical plasticity: use-dependent remodelling |magnetic resonance imaging | topographic maps in the brainMIT Encyclopedia of Cognitive Sciences:http://cognet.mit.edu/MITECS/auditory plasticity | electrophysiology, electric and magneticevoked fields | magnetic resonance imaging | neural plasticity |positron emission tomographyAccess to this interactive links box is free online.Complex-trait genetics: emergence ofmultivariate strategiesTamara J. Phillips and John K. BelknapComplex traits, including many diseaserelatedtraits, are influenced by multiplegenes. Bivariate approaches that associateone gene with one trait are yielding tomultivariate methods to synthesize theeffects of multiple genes, integrate resultsacross independent studies, and aid in theidentification of coordinated pathways andinteractions between loci.The extraordinary success of the molecularrevolution in transforming modern biologyhas generated one important problem — howdo we synthesize the wealth of molecular datato gain insight into ‘higher-order’ processesthat exist at the levels of pathways, organ systemsand whole organisms? This question hasled to the rise of newer, synthetic researchstrategies to complement those based on dissectionand analysis. This trend is evident inthe genetic analysis of complex central nervoussystem traits — the subject of this article.Complex (or quantitative) traits are thoseinfluenced by multiple loci (genes), each ofwhich is known as a QUANTITATIVE TRAIT LOCUS(QTL). Approaches to the study of complextraits are of two general kinds, each withstrengths and weaknesses 1 . Gene-driven methodsfocus on a particular gene and seek todetermine the phenotypes that are influencedby that gene. The study of knockout animals isa prime example. By contrast, trait-drivenmethods focus on a specific trait (phenotype)and seek to discover the underlying genes.QTL mapping, genome-wide mutagenesisscreens and MICROARRAY expression analysis areexamples of this approach. Because of theirability to focus simultaneously on many genes,trait-driven methods lend themselves morereadily to multivariate genetic approachesthan do gene-driven methods. However, someexamples of the latter are also amenable tomultivariate approaches (such as the use ofdouble knockouts), as we discuss below.Arguably, the first important advances inthe study of complex-trait genetics were madeby gene-driven knockout and transgenicstrategies, which have been used to identifyhundreds of genes that probably influencespecific complex traits. Much of the progressthat has been made so far has been based onbivariate approaches — one locus is shown toinfluence one trait, or a series of bivariateexperiments are used to link one locus to478 | JUNE 2002 | VOLUME 3 www.nature.com/reviews/neuro© 2002 Nature Publishing Group