Przetwarzanie informacji muzycznej w mózgu
Transkrypt
Przetwarzanie informacji muzycznej w mózgu
Przetwarzanie informacji muzycznej w mózgu NEUROESTETYKA PIOTR PRZYBYSZ Wykład monograficzny. UAM Poznań 2010 Neuroestetyka muzyki? • Neuroestetyka sztuk wizualnych a neuropsychologia muzyki • Czy prawa i reguły przetwarzania przez mózg obrazu i dźwięku różnią się? • Czy istnieją wspólne zasady estetyczne rządzące poczuciem piękna w muzyce i sztukach wizualnych? • Przekaz muzyczny w sztukach wielomodalnych Dźwięk i słyszenie • Muzyka to złożona sekwencja dźwięków uporządkowanych w czasie pod względem głośności i wysokości • Dźwięk: (1) jako zjawisko fizyczne (=fala dźwiękowa) jest wyrażany przez obiektywne cechy dźwięku: a) częstotliwość, b) natężenie, c) widmo, d) czas trwania, (II) jako zjawisko psychologiczne (=wrażenie dzwiękowe) charakteryzowany jest przez a) wysokość, b) głosność, c) barwę, d) czas trwania BZP[PRiþbc^c[Xf^ĀRX SĄfXþZÙfX]bcad\T]cÙf \dihRi]hRW #7i "7i !7i 5[Tc_XRR^[^ $7i 7i ##7i $7i "7i !7i 5^acT_XP] 7i #7i :^]caPQPb !7i 6^b\ĄbZX "7i 6^bZ^QXTRh &7i Zagadnienie słuchu absolutnego • Słuch absolutny = zdolność do odczytania wysokości dźwięku bez porównywania z dźwiękiem odniesienia. • Własność pozamuzyczna, akustyczna, percepcji słuchowej. • Dotyczy max. 0,1 % populacji (duże rozbieżności danych); częściej spotykany w środowisku osób muzykujących (w szkołach muzycznych ok 15 %); znacznie częściej spotykany wśród nacji posługujących się językami tonalnymi (np. w chińskich szkołach muzycznych do 60 % osób, które zaczęły wcześnie edukację). zattore feature 4/3/05 10:11 am artists on science Page 312 Droga słuchowa (1) Music, the food of neuroscience? Playing, listening to and creating music involves practically every cognitive function. Robert Zatorre explains how music can teach us about speech, brain plasticity and even the origins of emotion. W Figure 1 The processing of sound waves from a musical instrument. After being transduced into neural impulses by the inner ear, information travels through several waystations in the brainstem and midbrain to reach the auditory cortex. The auditory cortex contains distinct subregions that are important for decoding and representing the various aspects of the complex sound. In turn, information from the auditory cortex interacts with many other brain areas, especially the frontal lobe, for memory formation and interpretation. The orbitofrontal region is one of many involved in emotional evaluation. The motor cortex is involved in sensory–motor feedback circuits, and in controlling the movements needed to produce music using an instrument. e tend to consider art and culture from a humanistic or historical perspective rather than a biological one. Yet these products of human cognition must have their origin in the function and structure of the human nervous system. As such, they should be able to yield valuable scientific insights. This line of reasoning is nowhere more evident than in the contemporary interest in the neuroscience of music. Music provides a tool to study numerous aspects of neuroscience, from motor-skill learning to emotion. Indeed, from a psychologist’s point of view,listening to and producing music involves a tantalizing mix of practically every human cognitive function. Even a seemingly simple activity, such as humming a familiar tune, necessitates complex auditory pattern-processing mechanisms, attention, memory storage and retrieval, motor programming, sensory– motor integration, and so forth (Fig. 1). Likewise, the musician does not consider music to be monolithic, but recognizes within it multiple features including melodies, chords, themes, riffs, rhythms and tempos. This complexity — both psychological and musicological — makes music a challenging topic for a scientific research programme. Increasing numbers of investigators are convinced that music can yield valuable information about how the brain NATURE | VOL 434 | 17 MARCH 2005 | www.nature.com/nature 312 ©2005 Nature Publishing Group Droga słuchowa (1I) PRZESTRAJANIE MÓZGU POSZCZEGÓLNE KOMÓRKI w mózgu reagujà optymalnie na okreÊlonà wysokoÊç dêwi´ku, czyli cz´stotliwoÊç (a). Komórki zmieniajà swoje dostrojenie, kiedy zwierz´ si´ uczy, ˝e okreÊlony ton jest dla niego wa˝ny (b). Takie komórkowe przystosowanie modyfikuje „map´ cz´stotliwoÊci” mózgu szczura, by zwi´kszyç obszar kory przetwarzajàcy ton o du˝ym znaczeniu dla zwierz´cia – na przyk∏ad poszerza rejon aktywowany przez dêwi´k o cz´stotliwoÊci 8 kHz, jeÊli ta akurat jest istotna (c). Numer komórki a 2 C D 3 4 5 6 7 E F G A H Reakcja 1 WysokoÊç dêwi´ku b Cz´stotliwoÊç treningowa 140 Najlepsza reakcja Reakcja (liczba wy∏adowaƒ na sekund´) 120 Najlepsza reakcja 100 80 60 Przed treningiem Po treningu 40 20 0 –20 0.1 1 10 100 Cz´stotliwoÊç (kHz) c 32 16 32 16 8 8 4 2 4 Przed treningiem (Êrodkowa cz´stotliwoÊç oktaw w kilohercach) 2 Po treningu go, natomiast przez pozosta∏e 10 lat ˝ycia zachowa∏ zdolnoÊç pisania muzyki. Tak wi´c przypuszczenia o niezale˝nym przetwarzaniu mogà byç s∏uszne, choç nowsze badania prowadzà do bardziej wywa˝onej koncepcji, w której przyk∏ada si´ wi´kszà wag´ do cech wspólnych muzyki i j´zyka – funkcji komunikacyjnej oraz sk∏adni, czyli zbioru regu∏ okreÊlajàcych w∏aÊciwe kombinacje elementów (odpowiednio nut lub s∏ów). Na podstawie wyników badaƒ z wykorzystaniem techniki 54 ÂWIAT NAUKI GRUDZIE¡ 2004 neuroobrazowania Aniruddh D. Patel z Neurosciences Institute w San Diego stwierdzi∏, ˝e w p∏acie czo∏owym istnieje obszar, który zajmuje si´ sk∏adnià zarówno muzyki, jak i j´zyka, natomiast pozosta∏e okolice mózgu odpowiadajà za inne aspekty przetwarzania i j´zyka, i muzyki. Badania obrazowe dajà tak˝e doÊç precyzyjny wglàd w to, jak mózg reaguje na muzyk´. Ich wyniki naj∏atwiej zrozumieç, jeÊli weêmie si´ pod uwag´ sposób przekazywania przez ucho dêwi´ków do mózgu [ramka na poprzedniej stronie]. Podobnie jak inne zmys∏y s∏uch jest zorganizowany hierarchicznie. Sk∏ada si´ z ciàgu neuronalnych stacji przetwarzania bodêca – od ucha a˝ do poziomu najwy˝szego, czyli s∏uchowych obszarów kory mózgu. Przetwarzanie dêwi´ków takich jak te, z których sk∏ada si´ muzyka, zaczyna si´ w uchu wewn´trznym (Êlimaku), które rozk∏ada z∏o˝one dêwi´ki, na przyk∏ad brzmienie skrzypiec, na poszczególne cz´stotliwoÊci. Âlimak przekazuje te informacje dalej jako sekwencje wy∏adowaƒ neuronalnych, które biegnà przez w∏ókna nerwu s∏uchowego nastrojone na ró˝ne cz´stotliwoÊci. Wreszcie sekwencje te docierajà do kory s∏uchowej p∏ata skroniowego. Na okreÊlone cz´stotliwoÊci reagujà ró˝ne komórki. Po∏o˝one obok siebie majà zaz´biajàce si´ krzywe strojenia (profile wra˝liwoÊci na cz´stotliwoÊç). W rezultacie, poniewa˝ sàsiadujàce ze sobà komórki sà nastrojone na podobne cz´stotliwoÊci, na powierzchni kory s∏uchowej tworzy si´ „mapa cz´stotliwoÊci” [ramka z lewej]. Sam odbiór muzyki jest jednak bardziej z∏o˝ony. Muzyka to sekwencja dêwi´ków, a jej percepcja polega na spostrzeganiu relacji pomi´dzy nimi. W przetwarzanie ró˝nych aspektów muzyki w∏àczonych jest wiele obszarów mózgu. Weêmy na przyk∏ad dêwi´k, który ma zarówno okreÊlonà cz´stotliwoÊç, jak i g∏oÊnoÊç. KiedyÊ badacze przypuszczali, ˝e wykrycie danej cz´stotliwoÊci zawsze wywo∏uje takie same reakcje komórek, które sà na nià nastrojone. Jednak w drugiej po∏owie lat osiemdziesiàtych, gdy wspólnie z Thomasem M. McKennà pracowa∏em w moim laboratorium w University of California w Irvine, zakwestionowaliÊmy ten poglàd, zbadawszy kontur melodii, czyli wzór wznoszenia si´ i opadania linii melodycznej, który stanowi podstaw´ ka˝dej melodii. NapisaliÊmy melodie sk∏adajàce si´ z tych samych pi´ciu dêwi´ków, ale ró˝niàce si´ konturem, i rejestrowaliÊmy reakcje pojedynczych neuronów w korze s∏uchowej kotów. Okaza∏o si´, ˝e reakcje komórek ró˝ni∏y si´ w zale˝noÊci od konturu. Zale˝a∏y od umiejscowienia danego dêwi´ku w melodii. Komórki mogà silniej reagowaç na dêwi´k poprzedzany przez inne ni˝ na dêwi´k rozpoczynajàcy sekwencj´. Ponadto reagujà na ten sam dêwi´k inaczej, gdy jest on elementem konturu wznoszàcego si´ (dêwi´ki coraz wy˝sze), a inaczej, kiedy wchodzi w sk∏ad konturu opadajàcego lub mieszanego. Wyniki te pokazujà, ˝e przebieg melodii ma du˝e znaczenie – przetwarzanie s∏uchowe nie polega na prostym odwzorowaniu dêwi´ku jak w telefonie czy zestawie hi-fi. Choç wi´kszoÊç badaƒ koncentruje si´ na melodii, rytm (wzgl´dny czas trwania dêwi´ków i odst´pów mi´dzy nimi), harmonia (wzajemna relacja wysokoÊci co najmniej dwóch równoczeÊnie brzmiàcych tonów) i barwa dêwi´ku (charakterystyczna ró˝nica w brzmieniu tego samego tonu wytwarzanego przez dwa ró˝ne instrumenty) tak˝e sà obiektem naukowych dociekaƒ. Wyniki wielu badaƒ nad rytmem sugerujà, LAURIE GRACE Mapy mózgowe i ich przestrajanie Where Does the Brain “Hear”? Left Hemisphere: Rhythm Right Hemisphere: Pitch and Melody Music is processed in various areas of the brain, which change depending on the focus of the listener and his or her experience. When the brain of an amateur musician processes simple rhythmic relations in a melody, such as the variance in length between certain tones, he utilizes the premotor, or movement-preparation, regions as well as sections of the parietal lobe in the left hemisphere. If the temporal relations among the tones are more complex, premotor and frontal lobe regions in the right hemisphere become active. In both cases, the cerebellum (which is commonly supposed to be involved in movement control) also participates. In contrast, musicians who are discerning between rhythms or meter predominantly employ parts of the frontal and temporal lobes in the right hemisphere. Rhythmic relations display a similar picture: people who are not musically trained process in the left side, whereas experienced musicians generally do so in the right. When a musical layperson compares different pitches, the right posterior frontal lobe and right upper temporal lobe convolution are active. The tones are stored for future use and comparison in the auditory working memory located in the temporal region. The middle and lower areas of the temporal lobe are also active when processing more complex musical structures or structures being stored in memory for a longer period. In contrast, professional musicians show increased activity in the left hemisphere when they are differentiating among pitches or perceiving chords. When the listener is focusing on whole melodies rather than individual tones or chords, entirely different sections of the brain become active: in addition to the primary and secondary auditory cortices, the auditory associative regions in the upper temporal lobe are at work. In this case, the activity is once again concentrated in the right hemisphere. Muzyka w mózgu Zasady przetwarzania muzyki: a) rozproszoność/asymetria półkulowa (ograniczona), b) hierarchiczność, c) modularność. Left Right Parietal lobe Rhythm Rhythm Frontal lobe Melodies Frontal lobe m yth Temporal Rh lobe Occipital lobe Cerebellum Rhythm Occipital lobe Complex musical structures Despite the gaps, scientists are piecing together a general understanding of where the brain “hears” music. We know, for example, that both sides, or hemispheres, of the brain are involved, though asymmetrically. For a long time, it was common to believe in a distinct division between the left brain’s processing of language (the side that also handles reasoning tasks) and the right brain’s processing of music (the half that contains emotional and spatial information). Many medical textbooks included this simplified theory until the 1980s. In recent years, however, researchers have established that injuries to either side can impair musical abilities. This happens not only in the case of damage to the auditory areas in the temporal lobe but also when associated regions of the frontal lobe and the pari- 28 Auditory cortex Parietal lobe Cerebellum M elo d ies Pitch comparison Temporal lobe Auditory working memory etal regions are affected. (If the Heschl’s gyrus is destroyed on both sides, incidentally, total deafness does not occur. Instead the ability to distinguish between various sounds is severely impaired. A patient with this condition would not be able to understand language or perceive music at all.) Early stages of music perception, such as pitch (a note’s frequency) and volume, occur in the primary and secondary auditory cortices in both hemispheres. The secondary auditory areas, which lie in a half-circle formation around the primary auditory cortex, process more complex music patterns of harmony, melody and rhythm (the duration of a series of notes). Adjoining tertiary auditory areas are thought to integrate these patterns into an overall perception of music. Farther for- COPYRIGHT 2003 SCIENTIFIC AMERICAN, INC. SCIENTIFIC AMERICAN MIND THOMAS BRAUN G&G •Z © 2003 Nature Publishing Group http://www.nature.com/natureneuroscience REVIEW Figure 1 A modular model of music processing. Each box represents a processing component, and arrows represent pathways of information flow or communication between processing components. A neurological anomaly may either damage a processing component (box) or interfere with the flow of information between two boxes. All components whose domains appear to be specific to music are in green; others are in blue. There are three neurally individuated components in italics—rhythm analysis, meter analysis and emotion expression analysis—whose specificity to music is currently unknown. They are represented here in blue, but future work may provide evidence for representing them in green. Acoustic input Acoustic analysis Temporal organization Pitch organization Tonal encoding Interval analysis Contour analysis Rhythm analysis Meter analysis Acoustic-tophonological conversation foot (thus a direct connection to tapping in Emotion Musical Phonological Fig. 1; see accompanying review36 in this expression lexicon lexicon issue). Both the melodic and temporal pathanalysis ways send their respective outputs to either the ‘musical lexicon’ or the ‘emotion expression analysis’ component. The musical lexicon Associative Vocal plan memories is a representational system that contains all formation the representations of the specific musical phrases to which one has been exposed during one’s lifetime. The same system also keeps a Speaking Singing Tapping record of any new incoming musical input. Accordingly, successful recognition of a familIvelisse Robles iar tune depends on a selection procedureIsabelle that Peretz & Max Colthear (2003) Modularity of music processing, Nature Review Neuroscience, vol.6, iss. 7. takes place in the musical lexicon. The output of the musical lexicon three new output modules again stems from the study of neurological can feed two different components, depending on task requirements. patients: singing performance in aphasic patients38 and tapping abiliIf the goal is to sing a song like “Happy Birthday,” the corresponding ties in adults suffering from congenital amusia39. Thus, our proposed melody, represented in the musical lexicon, will be paired with its asso- modular architecture for processing music provides a plausible frameciated lyrics that are stored in the phonological lexicon and will be work for further investigating the neural mechanisms of music protightly integrated and planned in a way that is suitable for vocal pro- cessing. duction. If the task requires retrieving nonmusical information about a musical selection, such as naming the tune or retrieving a related ACKNOWLEDGMENTS Based on research supported by grants from the Natural Sciences and Engineering experience from memory, the associated knowledge stored in the ‘asso- Research Council of Canada and the Canadian Institutes of Health Research to I.P. ciative memories’ component will be invoked. We thank C. Palmer and T. Griffiths for insightful comments made on an earlier In parallel with memory processes, but independently, the percep- draft. tual modules will feed their outputs into an emotion expression analyReceived 18 March; accepted 21 April 2003 sis component, allowing the listener to recognize and experience the Published online 25 June 2003; doi:10.1038/nn1083 37 emotion expressed by the music . This emotional pathway also contributes to recognition via the musical lexicon. Emotion expression 1. Handel, S. Listening: an Introduction to the Perception of Auditory Events (MIT press, Cambridge, Massachusetts, 1989). analysis is a pivotal processing component because music has the power to elicit strong emotional responses. It takes as input emotion- 2. Bregman, A. Auditory Scene Analysis. The Perceptual Organization of Sound. (MIT press, London, 1990). specific musical features, such as mode (e.g. major or minor) and 3. Zatorre, R. & Peretz, I. (eds.) The Biological Foundations of Music (NY Acad. Sci., New York, 2001). tempo (e.g. slow or fast) as computed by the melodic and temporal Fodor, J. The Modularity of Mind (MIT press, Cambridge, Massachusetts, 1983). pathways, respectively. What is currently unclear is to what extent this 4. 5. Fodor, J. The Mind Doesn’t Work That Way (MIT press, Cambridge, Massachusetts, emotion expression analysis component is specific to music as 2001). opposed to being involved in more general kinds of emotional pro- 6. Coltheart, M. Modularity and cognition. Trends Cogn. Sci. 3, 115–120 (1999). 7. Gardner, H. Musical intelligence. in Frames of Mind (ed. Gardner, H.) 31–58 (Basic cessing. A patient who could recognize pieces of music but could not Books, New York, 1983). respond emotionally to them, while being able to respond emotionally 8. Jackendoff, R. Consciousness and the Computational Mind (MIT Press, Cambridge, Massachusetts, 1987). to other media, would be informative here. 9. Peretz, I. & Morais, J. Music and modularity. Contemporary Music Rev. 4, 277–291 In sum, we propose a modular functional architecture for music (1989). processing that comprises several component modules. Our model 10. Peretz, I. et al. Functional dissociations following bilateral lesions of auditory cortex. 117, 1283–1302 (1994). (Fig. 1) also describes the pathways of information flow among these 11. Brain Peretz, I., Belleville, S. & Fontaine, S. Dissociations entre musique et langage après component modules. The characterization of each box and arrow repatteinte cérébrale: un nouveau cas d’amusie sans aphasie. Can. J. Exp. Psychol. 51, 354–368 (1997). resented in the model has been provided by the detailed study of 12. Griffiths, T.D. et al. Spatial and temporal auditory processing deficits following right brain-damaged patients with selective impairments or preservations hemisphere infarction: a psychophysical study. Brain 120, 785–794 (1997). of particular musical abilities (for review, see ref. 24). The inclusion of 13. Wilson, S.J. & Pressing, J. Neuropsychological assessment and the modeling of Przypadek Maurice Ravela • Po 1927 utracił on zdolność powtarzania ze słuchu melodii, zapisywania (agrafia) i rozumienia (afazja) i czytania (alexia) zapisanej muzyki (uszkodzenia skroniowo-ciemieniowe lewej półkuli), poruszania się w skoordynowany sposób (apraksja). • Zachował zdolności w zakresie rozpoznawania melodii (np. swoich utworów), wychwytywania błędów. • • Amuzja Przypadki odmienne: W. Shebalin - afazja bez amuzji articles Mózgi muzyków i nie muzyków d e b © 2002 Nature Publishing Group http://neurosci.nature.com © 2002 Nature Publishing Group http://neurosci.nature.com a articles Fig. 1. The auditory stimulus, evoked magnetic fields and cortical anatomy. (a) Stimulus waveform. A modulation frequency of 26–37 Hz was superimposed on sinusoidal tones with carrier frequencies of 100–5,600 Hz to measure the responses to tone onset and to each modulation cycle. (b) Typical averaged response at an MEG sensor over the right auditory cortex shows middle latency onset components P30m and P50m, long latency components N100m and the sustained field (SF). The responses to the modulation cycles appear superimposed on the SF. (c) Typical early N19m-P30m response of the PAC after deconvolution of the modulated signals. (d, e) Source model with one equivalent dipole in each hemisphere depicted in sagittal and transversal T1-weighted MRI images. The transversal section is parallel to the supratemporal plane. The source activity is modeled with dipoles drawn in the left and right hemispheres. (f) Three-dimensional (3D) gray matter surface reconstruction between of the right The FTS defines the activity, anterioramHG boundary and volume and musical aptitude. (a) The N19m-P30m Fig. 4. Correlations earlyHG. neurophysiological source gray matter the most posterior HS defines themean posterior boundary. The first transdipole moment was strongly correlated with the gray matter volume of amHG. Values were averaged over the right and left hemispheres. verseraw HGscore is sometimes by(AMMA the SI, atest) shallow sulcuscorrelated which does notboth the N19m-P30m dipole moment (b) and the gray (b, c) The tonal of musicaldivided aptitude was highly with matter volume of amHG (c).full length. extend over its Tekst and amateur musicians showed an intermediate gray matter Influence of external variables volume (189–798 mm3). The total volume of HG, including We found no influence of the covariates sex, age or head size signal that awas 115variance ± 18% and larger over white and average gray matter, showed larger could notall frequencies on either theinearly dipole amplitudes or on the gray matter volthegroups right(non-musicians, hemisphere (F1,955–4,694 P <3; 0.0001) 87of±amHG. 17% To exclude influences of attention during MEG 1,22 = 73.4,mm separate the ume profes- and 3; hemisphere larger in the left (F1,22 = 43.6, 0.0001).recording Non-musiand of the frequency modulation in the stimulus, we sionals, 2,629–6,297 mm amateurs, 2,151–7,603 mmP3).<The f cians, by contrast, didshowed not have significantly larger dipole carriedampliout two additional control sessions in a subgroup of 24 asymmetry measures (Methods) only one significant stochastically effect: the tudes total volume HG than was 14% larger the right hemiin the of right in the leftinhemisphere (5 ± 9%, F1,11 = 3.6, selected subjects. While watching a video, subjects sphere of professional musicians (right, 3,986 ± 305 mm3; left, were n.s). In amateur musicians, dipole amplitudes 19detected ± 14% deviant tones of a different frequency (1,100 Hz instead the 3,468 ± 263 mm3;in!HG 0.14 ±than 0.04;they F1,11were = 11.7, < 0.01). larger the=right inPthe left hemisphere (Fof = standard 500 Hz) and indicated them by button c 1,12attention press in the experiment. We found no significant effect 17.8, P < 0.01). This was significant at three frequencies (100 Hz, of attention on the primary N19m-P30m component. Within Correlation with musical aptitude 500 Hz and 1,100 Hz, P < 0.05). noise limits, the N19m-P30m signals for the onset of pure sinuFor all three groups, there was a high correlation between the a pronounced difference between lateagreed with the signals deconvoluted from the soidaland tones N19m-P30m There signal was amplitude and musical aptitude as mea-the early auditory cortical Whereas the primary early N19m-P30m commodulated tones23,40. sured by the AMMA tonal responses. test (Fig. 4b). Both the source activity and much the tonal score musical aptitude plex was larger inofmusicians, the latecomN100m component, pletely separated the evoked professional musicians the non- tones, DISCUSSION which was by the onset offrom the sinusoidal showed musicians.similar The amateur musicians showed an intermediate Here(Fig. we found amplitudes in all three groups over all frequencies 3b). a large difference in the early neurophysiological range of musical aptitude and dipole amplitudes that overactivity of the auditory cortex in musicians versus non-musicians, For all groups, N100m largest around 1,000 Hz. frequency andP, ranged from 78% Hz,HG, P < 0.0001) to 144% Schneider Scherg M, (1,100 Dosch Specht HJ, Gutschalk A,the Rupp A.was Morphology of Heschl's reflects lapped with the two other groups. Within groups, the correlausing simplegyrus tonal stimuli. In addition, we found strong correla(5,600 Hz, P < 0.0001). tion was significant for non-musicians (r = 0.55, P < 0.05), but tions of this activity with the gray matter volume of amHG and We next averaged the peak-to-peak N19m-P30m Morphology Heschl’s gyrus notmusicians. for amateurs (r =Nat 0.19, of n.s.) or professionals (r = 0.05, n.s.). with musical aptitude. Using partial correlations, we showed that enhanced activation in the auditory cortexdipole of Neurosci. 2002 Jul;5(7):688-94. amplitudes for each group (Fig. 3a). Amateur musicians large neurophysiological difference betweenthemusicians In general,The professional musicians had high AMMA scores, high gray matter volume of amHG was the key parameter influsignal amplitudes and large gray matter of amHG. early evoked response of the auditory cortex. The showed an intermediate average increase of 37 ± 11% over nonand non-musicians at the levelvolumes of the PAC coincidedencing with athe large When analysis was restricted to amateur and professional musilarger gray matter volume in professional musicians was most musicians (F1,23 = 7.8, P < 0.05). The difference cians, the correlation was significant (r = 0.52, P < 0.01). pronounced for amHG (130% greater than in non-musicians) between amateur musicians and non-musicians Similarly, the gray matter volume of amHG was highly corand dropped to 37% more volume than in non-musicians when was significant only in the low frequency range related with musical aptitude (Fig. 4c). Within groups, the corthe whole HG was used for anatomical reference. Together with (<1,000 Hz, P < 0.05). There was a frequency ! relation was significant for non-musicians (r = 0.71, P < 0.001) evidence from previous EEG20 and MEG21–23 studies that localgroup interaction (F 10,165 = 2.9, P < 0.01) that and amateurs (r = 0.56, P < 0.05), but not for professionals ized the origin of the primary auditory-evoked N19-P30 source ranged from 77% (100 Hz, P < 0.01) to 14% (r = 0.40, n.s.). When all amateurs and professionals were comactivity to amHG, our findings provide evidence for the augbined, however, the correlation was highly significant (r = 0.70, mentation of PAC gray matter in musicians. (2,500 Hz, nonsignificant (n.s.)). P < 0.0001). This correlation was smaller when considering the This functional–anatomical interpretation is consistent with In professional musicians, dipole amplitudes gray matter volume of aHG in its full lateral extent (r = 0.44, P the microanatomical24–29 finding that amHG comprises most were significantly larger in the right than in the left < 0.01) and was nonsignificant when the whole gray matter of the primary granular core field. However, the macroanatomhemisphere at all frequencies. On average, the volume of HG was calculated (r = 0.26, n.s.). No correlation ically defined amHG is only an approximate measure of the N19m-P30m signal was 21 ± 9% larger in the right was found between musical aptitude and white matter volumes location and extent of PAC, because there is considerable indihemisphere (F1,11 = 47.3, P < 0.0001). Compared of HG. vidual variability27–31. Non-primary cortical fields are most to non-musicians, professional musicians had an Under the assumption that anatomical size determined the likely to be found near the lateral and posterior edges of signal strength, a partial correlation was calculated to eliminate amHG2,27–29. Thus, the larger volume of gray matter in musithe influence of amHG gray matter volume of on the correlacians may comprise PAC as well as surrounding belt areas. The tion between N19m-P30m amplitude and AMMA score. This strong functional–anatomical correspondence at the level of Fig. 2. Auditory evoked N19m-P30m signals and 3D partial correlation was only r = –0.04 (n.s.), indicating that amHG is probably related to the stimulation with sinusoidal gray matter surface reconstructions of HG for all subanatomical size was the key parameter. tones. Whereas functional MRI41,42 and PET43 studies have shown jects aligned in the same order. Both the neurophysiological and the anatomical data show a large increase in professional musicians and a smaller increase in amanature neuroscience • volume 5 no 7 • july 2002 691 teur musicians. Left, dipole strength of the primary cortical response at 500 Hz. Source activities of the right (thick lines) and left (thin lines) hemispheres are superimposed. Right, highlighted areas show the amHG for each subject, aligned in the same order as the primary evoked responses. nature neuroscience • volume 5 no 7 • july 2002 689 served as subyears (range, 7 to 17for years) of 11.7 their instruments a mean period for our jectsyears study. served to nonmusicians as sub(range, 7Six 17 years)served as controls The mean for (15). age both jects for ourstudy.Six nonmusiciansserved Evidence has accumulatedover the past deafferentationof an entireforelimbin ma- groups 24 ? (15). 3 years. as was TheBefore meanour for both controls ageinvestwo decades that indicates that alterations caquedeafferentation monkeys (8) of and upperforelimb 1 extremity the a diary forinvesEvidence musicians has accumulated kept Before over the past in ma-tigation, an entire was groups 24 ? 3 years. our in afferent canthat induce plastic reorgaamputation humans(8) (9-11). twoinput decades week, recording amount of time pracindicates that caque in monkeys and upper extremity alterations the the a diary for 1 musicians tigation, kept ? 8.4 of in afferent canthe induce plastic reorga-In amputation in humans nizational (9-11). changesinput within in studies ticedweek, recording timeper pracaddition,it has been shown adult mam9.8amount hours per day (meanthe ? amount nizational changes within in studies addition, it has been shown adult mamticed maliancentral perhad dayestimated (mean 9.8the 8.4 hours with owlIn monkeys that a prolonged nervous system increase (1).the Changes week), and ofper maliancentral withstimulation owl monkeys nervous systemsensory (1). Changes had estimated theprevious amount of in the relation between peripheral of tactile of tothat thea prolonged distal padincrease timeweek), spent and practicing duringthe in their the relation peripheral oftwo tactile sensory stimulation to the pad ofmonth timeand the previous spent practicing fieldsand centralbetween representations have one or in adistal phalanges results greatly 8.8 hours year (10.8 ?during per fieldsand their representations one cortical in a to or two representation phalangesresults greatly month and year (10.8 ? 8.8 hours per been observed for increased thecentral specific somatosensory (2), have week). been for the somatosensory increased representation week). the experimental session, sovisual (1, 3, observed 4), and auditory systems(5), (2), ofcortical that portion the fingers (12, 13).specific Evi- to During visual (1, 3, 4), and auditory systems of (5), that the portion fingers (12, Evi13). the experimental During session, and comparablechanges also have been dence has also been reportedthat suggests matosensory was delivered stimulation tosoandmotor comparable changes have dence has also been was reportedthat stimulation delivered suggests matosensory found for systems (6). Inalso an increased cortical of the the first in separate many of been representation to the to digit and, runs, found for motor systemsof (6).afferent In many index increased of thefifththe first in separate of an representation to the and, runs, these experiments, the removal finger usedincortical reading by blindBraille of digit either hand. condigit Stimulation thesea experiments, theresulted removal afferent index fifth of either hand. condigitsuperficial Stimulation input from corticalregion inofan readers (14).fingerusedin readingby blindBraille of light sisted pressure appliedby from a corticalregion inputby readers(14). sisted light superficial pressure applied "invasion" a neighboring arearesulted whose in an Violinists and other string playerspro- means of aofpneumatic with stimulator the by "invasion" by a neighboring area whose Violinists and other string means of a with players prostimulator pneumatic innervationremainedintact. For example, vide a good model for the study of the use of standard,nonpainfulstimulationin-the innervation remainedintact. Forbeexample, a good model for input the study of thetensity use(9, of standard, stimulation nonpainful the cortical a digit region representing effectsvide of differential afferent to the data (Fig. 16, 17). The 1) indi-inthe cortical a region representing beeffects of differential afferent to the The data digit input (9,center 16, 17). (Fig. 1) indifore amputationin owl monkeyscould be tw,osides of the brain in humans.During catetensity that the of cortical responsivity fore in amputation owl in could be tw,o sides of the brain cate that the center of cortical monkeys humans. During responsivity activatedafter amputationby tactile stim- theirpracticeor performance,the secondto for tactile stimulationof the digits of the after amputation their of the digits practice thehand secondtoleft hand for tactile stimulation of the of an intact ulationactivated the fifth in musicians adjacent The stimtoperformance, was shifted as comfingerby (7).tactile digits (D2or D5) of the left ulationof an the fifth digits (D2 to D5) of the left hand left hand was shifted in musiciansas comadjacent finger (7). The of the orderof changes noted wereintact a few are continuously engagedin fingeringthe paredto that in controls,while at the same changes noted were of the orderof a few are continuouslyengagedin fingeringthe paredto that in controls,while at the same millimeters.Moreextensiveplasticchanges strings, a task that involves considerable time the strength of response increased. millimeters.Moreextensiveplasticchanges strings, a task that involves considerable time the strength of response increased. have recently been observed afterthe abomanual dexterity and and enhanced TheThe shift was towardthe midsensory topographic have recently been observed after the abomanual dexterity enhanced sensory topographicshift was towardthe midlition oflition input of theof the At the stimulation. the thumb thethe of of time,time, surface plane, which, along offrom fromportions inputlarger Atsame the same largerportions stimulation. the thumbsagittal surface sagittal plane, which, along with of body-for example, the neck the althe is toward thethe grasps instrument and, somatosensory postcentral gyrus, region body-for example, with somatosensory graspsthe neck of the instrumentand, al- the postcentralgyrus, is toward region of the individual. Tekst A Dl A D~~~~~~~~~~~~~~~~~~ E25* E25* ~20 ~20 _ _ _ _ 15 15 o 10 o 10 :5- __ |Age StringString players players 0 |Age 3E C3 C 2.5 - 2.5 2.5 2.5 ~2- ~2- :5~0.5 *e~~~~~~~~~~~~~C *e~~~~~~~~~~~~~C O 1* O o 1* o __ U StringU players Stringplayers O~Controls O~Controls E 30- B 30- B D5 DlD~~~~~~~~~~~~~~~~~~ D5 c 15-ic 15-i 0 - 1- - 1 - * Controls Controls ~0.5 Controls Controls 510 1015 1520 20 05 at inception of musical practice at inception of musical practice c D O c O 0 String players String players * - D 0 00.5 0.51 11.5 1.52 22.5 2.5 Distance, to left D5,hand Distance, left hand Dl toDlD5, (cm)(cm) 1. (A) Equivalent Fig. current dipoles by stimulation of the thumb Dl. dipole controls; theisshift is larger D5 than forThe The dipole moment is also larger Fig. 1. (A) Equivalent current dipoles elicited elicited by stimulation of the thumb (D1) (D1) controls; the shift larger for D5for than for Dl. moment is also larger and fifth finger (D5) of hand MRI the left hand are superimposed onto an of hand black (magnetic formusicians' the musicians' D5,indicated as indicated by greater the greater magnitude the and finger (D5) ofWienbruch MRI the left are superimposed onto black (magnetic for the D5, as by magnitude the Elbert T,fifth Pantev C, C, Rockstroh B, an Taub E. Increased cortical representation of the the fingers of theofleft in resonance reconstruction of the cerebral of a control, (B)magnitude The magnitude the dipole moment a function of the resonance imaging)imaging) reconstruction of the cerebral cortex cortex of a control, who who arrow.arrow. (B) The of theofdipole moment as aasfunction of the age age of of was Science, selected to provide inception of musical practice; players are indicated by filled circles, the305-7. interpretation selected to provide anatomical for the for inception of musical practice; stringstring players are indicated by filled circles, interpretation of the of the stringwas players 1995anatomical Octlandmarks 13; landmarks 270(5234): MEG-based localization. The arrows represent the location and orientation of moment for indicontrol subjects by hatched circles. the larger dipole MEG-based localization. The arrows represent the location and orientation of moment for indicontrol subjects by hatched circles. NoteNote the larger dipole ECDfor vector of the twoaveraged digits'averaged viduals beginning musical practice before the of age12. of (C) 12.Scatterplot (C) Scatterplot of the musicians the ECDthe vector eachfor of each the two digits' across across viduals beginning musical practice before the age of the musicians (black)(black) and controls Theof length of the arrows Euclidean distances (in centimeters) between the cortical representations represents the magnimean magniand controls (yellow).(yellow). The length the arrows Euclidean distances (in centimeters) between the cortical representations of of represents the mean tudedipole of themoment DlD5. andThis D5.distance This distance formusicians' the musicians' left hands greater dipole moment two in each group. digits The average tude of the Dl and for the left hands was was greater thanthan thatthat for the for twothe eachingroup. digits The average of Dl D5 are andshifted Dl are medially shifted medially for theplayers string players compared to in controls, butdifference this difference not statistically significant. locationslocations of D5 and for the string compared to in controls, but this is notisstatistically significant. * VOL. SCIENCE 13 OCTOBER * VOL. SCIENCE 270 *27013*OCTOBER 1995 1995 305305 Tekst Schlaug G, Jäncke L, Huang Y, Staiger JF, Steinmetz H. Increased corpus callosum size in musicians. Neuropsychologia. 1995 Aug; 33(8):1047-55.