musical training enhances automatic encoding of melodic ... · of melodic contour and interval...

Musical Training Enhances Automatic Encodingof Melodic Contour and Interval Structure

Takako Fujioka14 Laurel J Trainor3 Bernhard Ross12Ryusuke Kakigi4 and Christo Pantev12

Abstract

amp In music melodic information is thought to be encoded intwo forms a contour code (updown pattern of pitch changes)and an interval code (pitch distances between successivenotes) A recent study recording the mismatch negativity(MMN) evoked by pitch contour and interval deviations insimple melodies demonstrated that people with no formalmusic education process both contour and interval informa-tion in the auditory cortex automatically However it is stillunclear whether musical experience enhances both strategiesof melodic encoding We designed stimuli to examine contourand interval information separately In the contour conditionthere were eight different standard melodies (presented on80 of trials) each consisting of five notes all ascending inpitch and the corresponding deviant melodies (20) werealtered to descending on their final note The interval con-dition used one five-note standard melody transposed to eight

keys from trial to trial and on deviant trials the last note wasraised by one whole tone without changing the pitch contourThere was also a control condition in which a standard tone(9907 Hz) and a deviant tone (11110 Hz) were presented Themagnetic counterpart of the MMN (MMNm) from musiciansand nonmusicians was obtained as the difference between thedipole moment in response to the standard and deviant trialsrecorded by magnetoencephalography Significantly largerMMNm was present in musicians in both contour and intervalconditions than in nonmusicians whereas MMNm in thecontrol condition was similar for both groups The intervalMMNm was larger than the contour MMNm in musicians Nohemispheric difference was found in either group The resultssuggest that musical training enhances the ability to automati-cally register abstract changes in the relative pitch structure ofmelodies amp

INTRODUCTION

Musical ability is an essential component of humannature and there is no known human society past orpresent without music (Wallin Merker amp Brown 2000)Across different societies musical structure has twoaspects involving time (rhythm) and pitch (Deutsch1999) although the particular instantiations differ fromone musical system to another From a perceptual pointof view sequences of tones or melodies have twoaspects a contour and an interval code (Dowling 19781982) The contour representation consists of informa-tion about the up and down pattern of pitch changesregardless of their exact size and is common to bothspeech prosody and musical melody (Patel PeretzTramo amp Labreque 1998) The interval representationconsists of the exact ratio of pitch between successivetones and is specific to music forming the basis fromwhich scales and harmony can emerge Behavioral stud-ies have provided evidence that contour is more funda-

mental than interval in that both infants and musicallyuntrained adults are able to process contour informa-tion but have difficulties encoding the intervals of unfa-miliar melodies (Trehub Trainor amp Unyk 1993 Bartlettamp Dowling 1980 Dowling 1978 Cuddy amp Cohen 1976)The pitch intervals of unfamiliar melodies are betterperceived by trained than untrained musicians (TrainorDesjardins amp Rockel 1999 Peretz amp Babaı 1992)

How is melodic information processed stored andretrieved in the brain and what is the effect of musicaltraining Several behavioral studies support the idea thatplastic changes in the brain caused by years of trainingenable superior musical perception and performanceFor example it has been reported that musicians showa left hemispheric dominance in recognizing melodicsequences while nonmusicians show a right hemi-spheric dominance (Bever amp Chiarello 1974) As wellmusicians demonstrate better performance in the lefthemisphere than the right hemisphere when they areasked to recognize a part of a melodic sequence (Peretzamp Babaı 1992)

Recent neuroimaging investigations have begun toidentify brain structures relevant for music processing(eg Tillmann Janata amp Bharucha 2003 Janata et al

1The Rotman Research Institute Baycrest Centre for GeriatricCare 2University of Munster 3McMaster University 4NationalInstitute for Physiological Sciences

D 2004 Massachusetts Institute of Technology Journal of Cognitive Neuroscience 166 pp 1010ndash1021

2002 Koelsch et al 2002 Ohnishi et al 2001 SatohTakeda Nagata Hatazawa amp Kuzuhara 2001 Halpernamp Zatorre 1999 Platel et al 1997 Zatorre Evans ampMeyer 1994) For example a functional magnetic reso-nance imaging (fMRI) study has shown that listening tomusic produces an enhanced activation in the planumtemporale and the left posterior dorsolateral prefrontalcortex in musicians in comparison to pure tones (Oh-nishi et al 2001) Long-term musical training has bothanatomical and functional consequences For exampleanatomical asymmetries in the auditory cortices of mu-sicians were reported using high-resolution MRI Theleft planum temporale was larger than the right inskilled musicians especially in those with absolute pitch(the ability to recognize and name the pitch of a musicaltone without reference to a comparison tone) (SchlaugJancke Huang amp Steinmetz 1995) Furthermore thespecific anterior subregion of the corpus callosum wasfound to be larger in musicians who commenced musi-cal training before the age of seven (Schlaug JanckeHuang Staiger amp Steinmetz 1995) Functional differ-ences in musicians have also been demonstrated mainlyin the sensorimotor and in the auditory systems Violin-ists practice rapid independent movement of the fingersof their left hand on the fingerboard of their instrumentevery day for many years Somatosensory-evoked mag-netic fields reveal the results of this experience in largercortical representations of the left-hand fingers of musi-cians in comparison to either the right-hand fingers ofmusicians or the fingers of control subjects who neverplayed a violin (Elbert Pantev Wienbruch Rockstroh ampTaub 1995) In contrast behavioral and functional MRIstudies on pianists (Jancke Shah amp Peters 2000 JanckeSchlaug amp Steinmetz 1997) have found less asymmet-rical activity in the motor cortices of musicians than inthose of control subjects since keyboard performancerequires similar motor control skills for both handsFurthermore transcranial magnetic stimulation (TMS)data demonstrated that bimanual motor activities inkeyboard musicians were less inhibited than in normalsubjects (Ridding Brouwer amp Nordstrom 2000) as ispredicted from the observation of an enlarged corpuscallosum in musicians (Schlaug Jancke Huang Staigeret al 1995) Musicians and nonmusicians also differ intheir evoked responses from auditory cortex A magne-toencephalography (MEG) study reveals that musiciansshow larger amplitude auditory evoked N1m responses(peaking at about 100 msec after stimulus onset) forpiano sounds (Pantev et al 1998) and an electroen-cephalography (EEG) study reveals that they showlarger P2 and N1c responses (Shahin Bosnyak Trainoramp Roberts 2003) Furthermore the effect of N1m en-hancement is specific to the instrument of training(Pantev Roberts Schulz Engelien amp Ross 2001) Eventrelated potential (ERP) studies have also demonstrateddifferences between musicians and nonmusicians inlater processing that is likely outside primary or second-

ary sensory areas Besson Faıta amp Requin (1994) andTrainor et al (1999) report enhanced late positive wavesbetween 300 and 600 msec Relations between function-al and anatomical measures are also emerging Schneideret al (2002) combined structural information from MRIand functional information from MEG to reveal a corre-lation between Heschlrsquos gyrus enlargement and en-hanced evoked magnetic field response in the latencyof 19ndash30 msec in musicians In summary the comparisonof brain responses between skilled musicians and naivesubjects can give us new insights into brain plasticityassociated with musical training

The ability to encode an acoustical context is reflectedelectrophysiologically in the mismatch negativity (MMN)component of ERPs The MMN and its magnetic coun-terpart the MMNm are elicited within about 100 to200 msec after any discriminable change that occurs in-frequently in a repeatedly presented auditory stimulieven when the stimulus is not attended (Picton et al2000 Naatanen amp Picton 1987 Naatanen 1992) TheMMN reflects processing in neural memory traces bywhich the auditory cortex handles representations ofthe recent acoustic past and its repetitive aspects Thesources of the MMN have been located mainly in thesupratemporal plane (Alho 1995) by dipole modeling(Scherg Vajsar amp Picton 1989) and scalp current den-sity (SCD) maps obtained from EEG (Giard PerrinPernier amp Bouchet 1990) and MEG (Levanen AhonenHari McEvoy amp Sams 1996) The contribution of fron-tal generators of MMN has been also suggested not onlyby SCD (Giard et al 1990) but also by lesion studies(Alain Woods amp Knight 1998 Alho Woods AlgaziKnight amp Naatanen 1994) MMN is elicited not onlyfor changes in single acoustic features but also for morecomplex and abstract features Recent studies show thatMMN can be obtained even by changes in an auditorypattern (Alain Cortese amp Picton 1999 Alain Woods ampOgawa 1994) and a change from ascending to descend-ing tone pairs (Saarinen Paavilainen Schroger Terva-niemi amp Naatanen 1992) Furthermore MMN is affectedby experience such as phoneme categorization in aparticular language (Phillips et al 2000 Cheour et al1998 Naatanen et al 1997) Even after a short time ofintense listening training the increase in MMN ampli-tude parallels the increased discrimination performancewith tone frequency discrimination (Menning Robertsamp Pantev 2000) or with foreign phoneme categories(Menning Imaizumi Zwitserlood amp Pantev 2002)These studies indicate that the conscious process of dis-crimination and the unconscious process of extractingchanges in memory traces interact and that these twoprocesses are affected by training

The goal of the present study is to investigate therelationship between long-term musical training andautomatic melodic processing Changes in both pitchcontour (Trainor McDonald amp Alain 2002 TervaniemiRytkonen Schroger Ilmoniemi amp Naatanen 2001 Paa-

Fujioka et al 1011

vilainen Jaramillo amp Naatanen 1998 Tervaniemi Mauryamp Naatanen 1994 Saarinen et al 1992) and pitch in-terval (Trainor et al 2002) can evoke MMN responsesfrom nonmusicians even when the pitch level arechanged from trial to trial However no study to datehas assessed musical training effects separately in con-tour and interval encoding We investigated the MMNmresponses to contour and interval changes in bothmusicians and nonmusicians The stimuli were designedto clearly separate contour and the interval encodingFactors in the musical context other than contourand interval such as out-of-key changes familiarity ofmelody and the range of pitch leaps were carefullycontrolled A control condition examining frequencydeviations to single tones was also included

RESULTS

Clear auditory evoked magnetic fields (AEFs) were ob-tained from both musicians and nonmusicians in all sti-mulus conditions The grand-averaged dipole momentwaveforms for both groups are shown in Figure 1 P1mndashN1mndashP2m responses from the first to the fourth notewere observed in both melodic conditions and in bothgroups Those waveforms showed similar slow baselineshifts over the duration of the stimulus in both groupsDespite the smaller signal- to-noise ratio in the deviants(smaller number of trials) than standards highly repro-ducible response patterns were obtained in both melodicconditions After the onset of the fifth note musiciansshowed clear MMNm responses in both hemispheres forboth contour and interval conditions In contrast non-musicians showed unclear responses in both contour andinterval conditions In the single tone control conditionon the other hand both groups showed clear MMNmresponses to the frequency change of the stimulus

The magnified MMNm waveforms after onset of thedeviation are shown in Figure 2 with the mean of 95confidence limits calculated using bootstrap resamplingfrom every data point (shown as horizontal lines aroundzero from 50 to 250 msec after onset of deviation) Thisallows us to identify the MMNm response significantlydifferent from zero as the parts lying outside the limitsAccording to the analysis musicians showed highlysignificant MMNm responses for the deviation of melo-dies at 100 to 200 msec after onset In contrast theMMNm responses in nonmusicians reached significanceonly in the right hemisphere for the interval conditionFor both groups the MMN evoked by single tonefrequency deviation was highly significant between 95and 200 msec after stimulus onset

The magnitude of MMNm for the two melodic con-ditions in the musician group was compared using thesame analysis method As shown in Figure 3 musiciansshowed significantly larger MMNm in the interval thanin the contour condition The data in nonmusicians

was not compared since the contour MMNm was notpresent significantly above baseline levels

The MMNm peaks were slightly later in the intervalcondition compared to the contour condition as dis-played in Figure 2 However individual response var-ied widely in morphology (ie single or double peak)This prevented an unambiguous identification of theMMNm peak latency in single subjects which is requiredfor statistics on the latency differences between the stim-ulus conditions

The amplitudes of MMNm calculated around the peaklatency of the grand-averaged waveforms are shown inTable 1 The MMNm was significantly larger in musiciansthan nonmusicians according to the analysis of variance(ANOVA) F(122) = 12787 p lt 01 MMNm was alsosignificantly different across conditions (contour inter-val control) F(244) = 16552 p lt 0001 and there wasa significant interaction between group and conditionF(244) = 402 p lt 05 In musicians the MMNm wassignificantly larger in the control and the interval than inthe contour condition ( p lt 05) The larger amplitudein the interval condition compared to the contourcondition can also be seen in Figure 3 In nonmusiciansthe MMNm amplitude was larger in control than bothcontour ( plt 0001) and interval conditions ( plt 01) asalso clearly seen in the waveforms depicted in Figure 2The factor hemisphere was not significant in any con-ditions for either group nor did hemisphere interactwith any other factors

The results of the behavioral tests are reported inTable 2 The performance of musicians was significantlybetter than that of nonmusicians according to theANOVA F(122) = 23865 p lt 0001 Musicians ex-hibited good performance in both contour (9650)and interval (9583) conditions and they did not showsignificant differences between tasks Nonmusicians per-formed at 630 in the interval task which was abovethe chance levels of 50 t(11) = 3564 p lt01 How-ever their performance of 8617 in the contour taskwas significantly better t(11) = 5946 p lt 0001 than inthe interval task

DISCUSSION

In the present study musicians showed significantlylarger MMNm responses than nonmusicians to devia-tions in melodic contour and interval structure whereasboth groups showed similar MMNm responses to fre-quency deviation in a single pure tone As well bothgroups tended to show larger MMNm responses tointerval than to contour changes The results stronglysupport the hypothesis that musical experience leads tospecific changes in the neural mechanisms for process-ing of abstract but not sensory melodic informationThe contour and interval processing must be performedat the level of melodic patterns because the overall pitchlevels shifts from trial to trial Because MMNm for simple

1012 Journal of Cognitive Neuroscience Volume 16 Number 6

frequency deviation did not differ between the twogroups whereas MMNm to contour and interval changesdid we can conclude that musical training mainly affectsthe pitch contour and interval relations between tonesrather than the encoding of single tones These resultsextend a number of recent studies showing that MMNreflects not only the encoding of simple sensory fea-tures but also the encoding of abstract rules and

patterns in an auditory context (Trainor et al 2002Alain Achim amp Woods 1999 Alain Cortese et al 1999Paavilainen et al 1998 Alain et al 1994 Saarinen et al1992) by showing differential effects of training onMMNm responses to different types of MMN The resultsare also in line with observations of more pronouncedMMN responses in musicians to the deviation of music-related stimuli such as harmonic chord progressions

Figure 1 Grand averages of

the source space waveforms

from both musicians and

nonmusicians for eachcondition (contour interval

and frequencymdashsingle tone)

in both left and right

hemispheres The y-axis showsthe dipole moment (positive

upward) and the x-axis shows

the time related to thestimulus onset The standard

averaged data the deviant

averaged data and the

difference between the deviantand the standard (MMNm

response) are shown The

deviation of stimuli of both

contour and interval conditionsoccur at the fifth note Closed

arrows indicated the clearly

identifiable MMNm responseswhereas open arrows indicated

the vague responses

Fujioka et al 1013

(Koelsch Schroger amp Tervaniemi 1999) pitch sequen-ces within a tonal structure (Brattico Naatanen ampTervaniemi 2002) and complex temporal patterns oftones (Tervaniemi Ilvonen Karma Alho amp Naatanen1997) The results also extend those of Tervaniemi et al(2001) by showing that the enhancement in musicians isnot restricted to contour processing but also applies tointerval processing without contour change

On the other hand the almost absent MMNm innonmusicians in the present study is different from theprevious observations of Trainor et al (2002) of clearMMN to both contour and interval changes There arelikely two factors contributing to this difference First thebehavioral performance of nonmusicians in our studywas much lower than that of nonmusicians in Trainoret alrsquos study as we modified their contour and intervalstimuli Specifically we matched the size of pitch devi-ation across conditions with the result that the con-tour melodies in our study had a smaller pitch rangeand therefore a smaller size of deviation for contour

changes The present interval melody is also morecomplicated than that of Trainor et al in two importantmusical features First their standard melody consistingof the first five notes of an ascending diatonic scale washighly familiar and had a simple pitch contour whereas

Figure 3 Comparison between interval and contour condition inmusicianrsquos MMNm response The difference waveform was obtained

by subtraction of response for contour condition from interval

condition combined across both hemispheres

Figure 2 The source space

waveforms of MMNm For both

contour and interval condition

the time scale on the x-axisrefers to the onset of the fifth

note The thick line represents

the MMNm response and thin

lines above and below the zeroline show the upper and lower

limit of 95 confidence

interval The mean ofconfidence interval in the

whole time series was

subtracted from the response

to adjust the baseline ofMMNm waveform to the zero

line


the melody of the present study was unfamiliar andcontained a more complex melodic contour As well theinterval changes in Trainor et alrsquos study contained out-of-key notes whereas ours did not Naıve subjects moreeasily detect changes in a familiar melody than in anunfamiliar one especially for nondiatonic changes (Bes-son amp Faıta 1995 Besson et al 1994) Thus the taskdifficulty was certainly a contributing factor to the smallamplitude MMN seen in the present study A secondmajor difference between studies was that we used MEGand analyzed data as represented in a negative orienteddipole source whereas Trainor et al used EEG MEG isless sensitive than EEG in detecting radial orientedsource current signals which in the present case couldbe generated from multiple sources of MMN includingfrontal activation (Opitz Rinne Mecklinger Von Cra-mon amp Schroger 2002 Rinne Alho Ilmoniemi Virta-nen amp Naatanen 2000 Alho 1995 Giard et al 1990)

Interestingly MMNm responses were larger in theinterval condition compared to the contour conditiondespite the fact that we controlled for interval size andrange across the conditions Thus we have no evidencethat contour is a more fundamental process or that it isneurally privileged even the musically untrained Thereare two possible explanations for the larger MMN tointerval changes The first concerns the relation betweencontour and interval information Note that the melodicinterval between successive notes essentially includesthe up-and-down contour information as well In otherwords contour processing could be a part of the intervalprocessing even though it is hard to tell whether one ofthe processes comes first or both run in parallel In theinterval task the same melody is presented transposedto different pitch levels on each trial However in thecontour task different intervals are present on each trialIf the auditory cortex automatically extracts intervalinformation and this information is irrelevant in the

contour task the presence of the constantly varyingintervals could obscure the automatic response to thechange in contour In this case the regularities of thestandard melodies in the interval task could be extractedmore easily than those of the contour melodies forwhich the various interval sizes need to be ignored Ifcorrect this suggests that contour information is ex-tracted after or concurrently with interval informationbut not before it A second explanation concerns the factthat the changes in the present interval stimuli are ac-companied by a change in tonality That is the terminalnote of interval standard melody is the fifth note of thescale and functions strongly as the dominant in the keyThus it is possible that the MMNm was larger for intervalthan contour changes because not only was the intervalprocessed but also a difference in tonality was detected

An explanation of the larger MMNm for interval thanfor contour changes must also account for the mismatchbetween the MMNm elicited and behavioral perfor-mance The behavioral performance of musicians wasalmost perfect in both conditions On the other handbehavioral performance in nonmusicians was muchbetter for contour than for interval changes althoughMMNm was only significantly present for intervalchanges and then only in the right hemisphere Gener-ally MMN amplitude corresponds to behavioral accuracy(Winkler et al 1999 Kraus McGee Carrell amp Sharma

Table 1 Mean of Dipole Moment (plusmn SEM) as the MMNm Amplitude of Each Hemisphere in Control Contour and IntervalConditions

Dipole Moment (nAm)

Subject Group n Control Contour Interval

Left hemisphere

Musicians 12 9737 plusmn 1543 7572 plusmn 1202 9644 plusmn 1197

Nonmusicians 12 9742 plusmn 2418 0395 plusmn 1458 3061 plusmn 1086

Right hemisphere



p lt 05 (musicians control gt interval interval gt contour)

p lt 01 (nonmusicians control gt interval)

p lt 0001 (nonmusicians control gt contour)

Table 2 Mean Correct Performance (plusmn SEM) in theBehavioral Discrimination Task of Contour and IntervalConditions

Subject Group n Contour Interval

Musicians 12 9650 plusmn 255 9583 plusmn 190

Nonmusicians 12 8617 plusmn 504 6300 plusmn 365

p lt 0001

Fujioka et al 1015

1995 Tiitinen May Reinikainen amp Naatanen 1994Naatanen Jiang Lavikainen Reinikainen amp Paavilainen1993) However some recent studies demonstrated thatMMN responses emerge even before subjects conscious-ly achieve the discrimination tasks in sound categoriza-tion (Allen Kraus amp Bradlow 2000 Dalebout amp Stack1999 Tremblay Kraus amp McGee 1998) The nonmusi-cians may have difficulty in performing the interval dis-crimination task behaviorally if it depends on highercognitive and attentive processes such as categorizationmemorization and decision making which utilize theoutput of the automatic MMN processes (Naatanen ampAlho 1997 Naatanen 1992)

No statistically significant laterality effect in the MMNmresponse was found in the present study in either groupalthough the MMNm was only significant for nonmusi-cians in the right hemisphere Previous behavioral stud-ies have shown evidence of discrete lateralization forcontour and interval processing as measured by psycho-physical performance in unilateral lesion patients andnormal listeners to monaural sound (Liegeois-ChauvelPeretz Babaı Laguitton amp Chauvel 1998 Peretz ampMorais 1987 Peretz Morais amp Bertelson 1987 Zatorre1985) It is possible that either the effect is too subtle todetect by our technique or it occurs at a processingstage after the automatic processes reflected in theMMN It should also be noted that our stimulation wasbinaural as was that of Trainor et al (2002) who also didnot show clear laterality effects Laterality effects mightbe seen more clearly with monaural stimulation andfuture studies should address this question

The clear MMNm differences between musicians andnonmusicians observed in our study contribute to thegrowing literature suggesting that musical training af-fects a whole network of brain areas from those in-volved in stimulus encoding and deviance detection tothose involved in conscious evaluation of the musicFor example during passive listening in addition to theMMN another preattentive negative response earlyright anterior negativity (ERAN peaking at about 200ndash250 msec to the violation of musical-harmony syntax)has been demonstrated (Koelsch et al 2001 Maess Ko-elsch Gunter amp Friederici 2001) which is also more pro-nounced in musicians than in nonmusicians (KoelschSchmidt amp Kansok 2002) During active discriminationtasks tonality violation elicits larger late event-relatedresponses in musicians than in nonmusicians such asthe P3 (Trainor et al 1999 Janata 1995 Cohen GranotPratt amp Barneah 1993) and a long-latency positive com-ponent (LPC) around 600 msec (Regnault Bigand ampBesson 2001 Patel Gibson Ratner Besson amp Holcomb1998 Besson amp Faıta 1995 Besson et al 1994 Levettamp Martin 1992 Besson amp Macar 1987) All these indi-cate that there must exist multiple parallel processingmodules related to various aspects of musical structuresOur results indicate that during the early automatic stagesof processing musical training particularly affects the

detection of changes at the abstract level of pitch contourand interval patterns but has less effect on the detectionof simple pitch changes

METHODS

Subjects

Twelve musicians (8 women) between 19 and 33 years ofage and 12 nonmusically trained adults (9 women)between 19 to 40 years of age participated in this studyThe musicians had studied more than one instrumentand practiced regularly for more than 10 years (10 to 23mean 143 years) with formal education including musi-cal schools or private lessons The nonmusicians had al-most no formal musical training (3 out of 12 had 2 yearsof lessons and quit playing more than 10 years ago therest had none) except in their regular school lessonsNone of the subjects in either group had absolute pitchperception All participants were right-handed as as-sessed by the Edinburgh handedness test and hadnormal hearing within the range of 250 to 8000 Hz astested by clinical audiometry The subjects consented toparticipate after they were completely informed aboutthe nature of the study The Ethics Commission of theBaycrest Centre for Geriatric Care approved all experi-mental procedures which are in accordance with theDeclaration of Helsinki

Stimuli

The stimuli for both the contour and the interval me-lodic conditions were composed of sequences of five-note standard and deviant melodies with standardmelodies occurring 80 of the time A sound file wascreated from digitally recorded piano timbres for eachnote in audio CD quality The duration of each notewas 300 msec for a total melody length of 1500 msecThe stimulus for the control condition was a se-quence of standard and deviant pure tones with differ-ent frequencies

Each melody in the contour condition was uniquelycomposed of different intervals from the C major dia-tonic scale and each started on one of five differentnotes from C5 to G5 (American notation) (Figure 4)Thus the eight melodies were not transpositions of eachother In the corresponding deviant melodies the firstfour notes were identical to those of each standardmelody However the last note was changed to a de-scending note Thus the contour and interval informa-tion was identical in the standard and deviant melodiesexcept for the last note which differed only in contour

In the interval condition the standard stimuli con-sisted of one five-note melody that was transposed toeight keys with starting notes in the same range as thoseof the contour stimuli The deviant melodies werederived from the set of standard melodies by raising


the final note by a whole tone (16th of an octave) achange that remained within the key of the melody(Figure 4) and did not change the contour The firstnotes of all contour and interval melodies were betweenC5 and G5 and the last notes were between G5 and F6The interval size deviations in the contour melodiesranged from a minor second to a major third (112 to4 of an octave) and its mean value was a major secondaround the median position of termination (B5) whichwas the same value of deviation exploited in the intervalconditions

Each melody was separated by a 900-msec silent in-terval The experimental session consisted of 900 trialsof the contour and 900 trials of the interval conditionwhich were divided into successive blocks of 300 trialstaking 12 min each The successive deviant trials wereseparated with at least two standards In the contourcondition the order of melody variations was pseudo-random To avoid the appearance of the same notewithin two successive trials of the interval conditionwhich might be recognized as lsquolsquooddrsquorsquo or lsquolsquoprimedrsquorsquodespite the interval deviation the transposition fromtrial to trial was set to be an upward major third and adownward minor third until the starting note becameG5 Thus the repeating tonality change resulted in theseries lsquolsquoC-E-C-F-D-F-D-Grsquorsquo

The control condition consisted of two succes-sive blocks of 500 pure tones of 300 msec durationpresented with an ISI of 450 msec The frequency of the80 standard tones was 9907 Hz (B5) and of the 20deviant tones 11110 Hz (C6) The size of the changewas a whole tone which equals the mean size of that

used in the contour and interval conditions The tem-poral envelopes of the tones were derived from those ofthe B5 and C6 piano tones respectively

Hearing thresholds for each subject were determinedfor the left and right ears for two sounds the B5 pianosound and the pure tone of 9907 Hz In all conditionsthe stimuli were presented at 60 dB above those thresh-olds (the piano thresholds were used for the contourand interval tasks and the pure tone threshold for thecontrol condition)

MEG Recordings

The magnetic field responses were recorded with a 151-channel whole-cortex magnetometer system (OMEGACTF Systems Inc Port Coquitlam Canada) The averageintersensor spacing of this device is approximately31 cm The MEG pickup coils of 2 cm in diameter areconfigured as first-order axial gradiometers with a 5-cmbaseline The MEG signals were band-pass filtered be-tween 01 and 100Hz and sampled at a rate of 3125 sec1In the melodic conditions the duration of a recordingepoch was 22 sec including a 04-sec prestimulus periodIn the control condition one epoch was 05 sec induration including a 01-sec prestimulus interval Theonset of the first note of each stimulus synchronizedthe stimulus presentation and the data acquisition aswell For both melodic conditions three successive re-cordings consisting of 300 trials each were performedresulting in 72 min total recording time Two successiverecordings of 500 trials each were performed within15 min in the control condition

Figure 4 Musical stimuli for

the contour (left column) and

interval (right column) tasks

In each case the first fournotes of the melodies form a

common sequence which is

followed by the standard and

deviant terminal note In thecontour case the interval size

changes among melodies but

the standard terminal notesalways rise whereas the

deviant terminal notes always

fall In the interval case the

deviant terminal note ishigher by a whole tone than

standard terminal notes but

the contour of the melody

does not change

Fujioka et al 1017

The recordings were performed in a sitting positionwithin the magnetically shielded room The subjectswere instructed not to pay attention to the sound stimuliand to watch soundless movies of their own choicewhich were projected onto a screen placed in front ofthe chair The subjectrsquos compliance was verified by videomonitoring The recording session began with the con-trol condition for all subjects For half of the participantsthe contour condition was presented first while for theother half the interval condition was presented first Noexplanation about the stimuli was provided

Data Analysis

The recorded magnetic field data were averaged selec-tively for the standard and deviant stimuli and allstimulus types In order to detect the eye-artifact con-taminated epochs the magnetic field amplitude in achannel located just over the eyes was examined If itexceeded 10 pT in the latency interval between 02 to15 sec the data were excluded from the averaging

The analysis technique of signal space projection(SSP) (Tesche et al 1995) was applied to the MEG datawhich combined the multichannel magnetic field datainto a single time series of magnetic dipole momentThe weighting factor of each MEG sensor contributing tothe result was the sensitivity of each sensor to a sourceat the specified location in the brain This forms a virtualsensor which is maximally sensitive to a source at thespecified origin and orientation and less sensitive toother sources This results in considerable discrimina-tion against the sensor noise and uncorrelated brainactivity from distant brain regions The SSP is a usefulmethod under the assumption of a single time-varyingsource at a fixed location A necessary prerequisite forthe SSP is the determination of the source origin andorientation Therefore a source analysis using a singleequivalent current dipole (ECD) model for the N1mcomponent (latency around 100 msec after the stimulusonset) of the AEF to the first note of the melodyregardless of standard or deviant condition was donein both hemispheres for each recorded dataset For eachsubject the average of these dipole locations and ori-entations across all stimulus conditions served as anestimate for the source in the auditory cortex Basedon these source coordinates the dipole moment wave-forms over the whole stimulus-related epochs werecalculated for all stimulus conditions This methodallows the averaging of dipole moment waveforms fromrepeated measurements in the same subject or betweensubjects Grand average dipole moment waveformsacross both groups of subjects were obtained selectivelyfor the standard and deviant stimuli Individual differ-ence waveforms were calculated by subtracting theresponse to the standard from that to the deviantstimuli MMNm responses were examined after theonset of the fifth note in the melodic conditions and

after the onset of the pure tone stimulus in the controlcondition The baselines of all responses were adjustedto the mean in a 100-msec interval previous to the onsetof the deviation The 95 confidence intervals for thegrand-averaged response waveforms and the differencewaveforms were estimated from nonparametric boot-strap resampling analysis (Davison amp Hinkley 1997)This method empirically establishes the distribution ofthe mean from repeated samples of the data itself andallows estimating confidence limits without the assump-tion of the underlying distribution This analysis wasapplied to all data points of the difference waveformsand allowed identifying those time intervals with ampli-tudes significantly different from zero Although thismethod allowed us to measure peak amplitudes andlatencies signal-to-noise ratio was not sufficient to com-pare source locations across the different conditions

In order to identify the amplitude of MMNm in theindividual data the peak latencies of grand-averageddifference waveforms were identified in all 12 cases withdifferent parameters (two groups musicians and non-musicians three conditions control contour and inter-val two hemispheres) The latency interval was definedas mean of those peak latencies The single subjectrsquosMMNm amplitude was defined as the mean value of thewaveforms within the 40-msec time interval centered atthe mean peak latency This procedure was necessarybecause the identification of peak latency and amplitudein the individual data was not always feasible Theamplitudes of MMNm were statistically examined by arepeated measures ANOVA with one between-subjectsfactor (group) and two within-subjects factors (condi-tion and hemisphere) The post hoc comparison wascalculated with Fisherrsquos PLSD tests using the level ofsignificance as 5

Behavioral Test

After the MEG recordings all subjects participated in abehavioral test consisting of two contour and intervaldiscrimination tasks The tests were designed as twoalternative forced-choice tasks (2AFC) with two melo-dies presented sequentially on each trial The subjectswere instructed to judge whether both melodies werethe same or different in terms of one of contour orinterval structure One melody of each trial was chosenfrom the set of standard stimuli whereas the othermelody was either another standard melody (same) ora deviant melody (different) The melodies were pre-sented in the same order as in the MEG recordings(randomized for the contour task the same order in theinterval task) The same and different pairs occurredwith equal probability The presentation of stimuli andthe recording of the subjectrsquos responses were controlledby specially developed software on a desktop computerThe stimuli were presented at an intensity of about60 dBSL through headphones and the subjects re-


sponded by a mouse click on buttons shown on thecomputer monitor The silent interval between the firstand the second melody was 900 msec The next trialstarted after the subjectrsquos response The subjects wereinstructed in detail about the tasks and briefly trainedby a few trials with feedback until they understood thetask correctly In the actual testing condition no feedbackwas provided

All behavioral data were examined statistically byrepeated measures ANOVA with one between-subjectsfactor (group musicians vs nonmusicians) and onewithin-subjects factor (task contour vs interval) Thepost hoc comparison was done with paired t tests forboth tasks

Acknowledgments

We thank Dr Claude Alain for helpful comments on a previousversion of the manuscript Ms Judy Vendramini and MsHaydeh Shaghaghi for technical assistance and the studentsfrom the Faculty of Music University of Toronto for theirparticipation This research has been supported by theInternational Foundation for Music Research CA USA

Reprint requests should be sent to Dr Christo Pantev Institutefor Biomagnetism and Biosignalanalysis Munster Univer-sity Hospital University of Munster Kardinal-von-Galen-Ring 10 48129 Munster Germany or via e-mail pantevuni-muensterde

REFERENCES

Alain C Achim A amp Woods D L (1999) Separatememory-related processing for auditory frequency andpatterns Psychophysiology 36 737ndash744

Alain C Cortese F amp Picton T W (1999) Event-relatedbrain activity associated with auditory pattern processingNeuroReport 10 2429ndash2434

Alain C Woods D L amp Knight R T (1998) A distributedcortical network for auditory sensory memory in humansBrain Research 812 23ndash37

Alain C Woods D L amp Ogawa K H (1994) Brain indices ofautomatic pattern processing NeuroReport 6 140ndash144

Alho K (1995) Cerebral generators of mismatch negativity(MMN) and its magnetic counterpart (MMNm) elicited bysound changes Ear and Hearing 16 38ndash51

Alho K Woods D L Algazi A Knight R T amp NaatanenR (1994) Lesions of frontal cortex diminish the auditorymismatch negativity Electroencephalography and ClinicalNeurophysiology 91 353ndash362

Allen J Kraus N amp Bradlow A (2000) Neural representationof consciously imperceptible speech sound differencesPerception and Psychophysics 62 1383ndash1393

Bartlett J C amp Dowling W J (1980) Recognition oftransposed melodies A key-distance effect in developmentalperspective Journal of Experimental Psychology HumanPerception and Performance 6 501ndash515

Besson M amp Faıta F (1995) An event-related potential (ERP)study of musical expectancy Comparisons of musicianswith non-musicians Journal of Experimental PsychologyHuman Perception and Performance 21 1278ndash1296

Besson M Faıta F amp Requin J (1994) Brain wavesassociated with musical incongruities differ for musiciansand non-musicians Neuroscience Letters 168 101ndash105

Besson M amp Macar F (1987) An event-related potentialanalysis of incongruity in music and other non-linguisticcontexts Psychophysiology 24 14ndash25

Bever T G amp Chiarello R J (1974) Cerebral dominancein musicians and nonmusicians Science 185 537ndash539

Brattico E Naatanen R amp Tervaniemi M (2000) Contexteffects on pitch perception in musicians and nonmusiciansEvidence from event-related-potential recordings MusicPerception 19 199ndash222

Cheour M Ceponiene R Lehtokoski A Luuk A Allik JAlho K amp Naatanen R (1998) Development oflanguage-specific phoneme representations in the infantbrain Nature Neuroscience 1 351ndash353

Cohen D Granot R Pratt H amp Barneah A (1993)Cognitive meanings of musical elements as disclosed byevent-related potential (ERP) and verbal experimentsMusic Perception 11 153ndash184

Cuddy L L amp Cohen A J (1976) Recognition of transposedmelodic sequences Quarterly Journal of ExperimentalPsychology 28 255ndash270

Dalebout S D amp Stack J W (1991) Mismatch negativity toacoustic differences not differentiated behaviorally Journalof the American Academy of Audiology 10 388ndash399

Davison A C amp Hinkley D V (1997) Bootstrap methods andtheir application Cambridge Cambridge University Press

Deutsch D (1999) The psychology of music (2nd ed) SanDiego Academic Press

Dowling W J (1978) Scale and contour Two componentsof a theory of memory for melodies Psychological Review85 341ndash354

Dowling W J (1982) Contour in context Comments onEdworthy Psychomusicology 2 47

Elbert T Pantev C Wienbruch C Rockstroh B amp TaubE (1995) Increased cortical representation of the fingersof the left hand in string players Science 270 305ndash307

Giard M H Perrin F Pernier J amp Bouchet P (1990)Brain generators implicated in the processing of auditorystimulus deviance A topographic event-related potentialstudy Psychophysiology 27 627ndash640

Halpern A R amp Zatorre R J (1999) When that tune runsthrough your head A PET investigation of auditory imageryfor familiar melodies Cerebral Cortex 9 697ndash704

Janata P (1995) ERP measures assay the degree of expectancyviolation of harmonic contexts in music Journal ofCognitive Neuroscience 7 153ndash164

Janata P Birk J L Van Horn J D Leman M Tillmann Bamp Bharucha J J (2002) The cortical topography of tonalstructures underlying Western music Science 2982167ndash2170

Jancke L Schlaug G amp Steinmetz H (1997) Hand skillasymmetry in professional musicians Brain and Cognition34 424ndash432

Jancke L Shah N J amp Peters M (2000) Corticalactivations in primary and secondary motor areas forcomplex bimanual movements in professional pianistsBrain Research Cognitive Brain Research 10 177ndash183

Koelsch S Gunter T C Schroger E Tervaniemi MSammler D amp Friederici A D (2001) Differentiating ERANand MMN An ERP study NeuroReport 12 1385ndash1389

Koelsch S Gunter T C Von Cramon D Zysset SLohmann G amp Friederici A D (2002) Bach speaks Acortical lsquolsquolanguage-networkrsquorsquo serves the processing ofmusic Neuroimage 17 956ndash966

Koelsch S Schmidt B H amp Kansok J (2002) Effects ofmusical expertise on the early right anterior negativity An

Fujioka et al 1019

event-related brain potential study Psychophysiology 39657ndash663

Koelsch S Schroger E amp Tervaniemi M (1999) Superiorpre-attentive auditory processing in musicians NeuroReport10 1309ndash1313

Kraus N McGee T Carrell T D amp Sharma A (1995)Neurophysiologic bases of speech discrimination Earand Hearing 16 19ndash37

Levanen S Ahonen A Hari R McEvoy L amp Sams M(1996) Deviant auditory stimuli activate human left andright auditory cortex differently Cerebral Cortex 6288ndash296

Levett C amp Martin F (1992) The relationship betweencomplex music stimuli and the late components of theevent-related potential Psychomusicology 11 125ndash140

Liegeois-Chauvel C Peretz I Babaı M Laguitton V ampChauvel P (1998) Contribution of different cortical areasin the temporal lobes to music processing Brain 1211853ndash1867

Maess B Koelsch S Gunter T C amp Friederici A D (2001)Musical syntax is processed in Brocarsquos area An MEG studyNature Neuroscience 4 540ndash545

Menning H Imaizumi S Zwitserlood P amp Pantev C (2002)Plasticity of the human auditory cortex induced bydiscrimination learning of non-native mora-timedcontrasts of the Japanese language Learning andMemory 9 253ndash267

Menning H Roberts L E amp Pantev C (2000) Plasticchanges in the auditory cortex induced by intensivefrequency discrimination training NeuroReport 11817ndash822

Naatanen R (1992) Attention and brain function HillsdaleNJ Erlbaum

Naatanen R amp Alho K (1997) Mismatch negativitymdashThemeasure for central sound representation accuracyAudiology Neurootology 2 341ndash353

Naatanen R Jiang D Lavikainen J Reinikainen Kamp Paavilainen P (1993) Event-related potentials reveala memory trace for temporal features NeuroReport 5310ndash312

Naatanen R Lehtokoski A Lennes M Cheour MHuotilainen M Iivonen A Vainio M Alku P IlmoniemiR J Luuk A Allik J Sinkkonen J amp Alho K (1997)Language-specific phoneme representations revealed byelectric and magnetic brain responses Nature 385432ndash434

Naatanen R amp Picton T (1987) The N1 wave of the humanelectric and magnetic response to sound A review and ananalysis of the component structure Psychophysiology 24375ndash425

Ohnishi T Matsuda H Asada T Aruga M Hirakata MNishikawa M Katoh A amp Imabayashi E (2001) Functionalanatomy of musical perception in musicians CerebralCortex 11 754ndash760

Opitz B Rinne T Mecklinger A Von Cramon D ampSchroger E (2002) Differential contribution of frontaland temporal cortices to auditory change detection fMRIand ERP results Neuroimage 15 167ndash174

Paavilainen P Jaramillo M amp Naatanen R (1998) Binauralinformation can converge in abstract memory tracesPsychophysiology 35 483ndash487

Pantev C Oostenveld R Engelien A Ross B RobertsL E amp Hoke M (1998) Increased auditory corticalrepresentation in musicians Nature 392 811ndash814

Pantev C Roberts L E Schulz M Engelien A amp Ross B(2001) Timbre-specific enhancement of auditory corticalrepresentations in musicians NeuroReport 12 169ndash174

Patel A D Gibson E Ratner J Besson M amp Holcomb P J

(1998) Processing syntactic relations in language and musicAn event-related potential study Journal of CognitiveNeuroscience 10 717ndash733

Patel A D Peretz I Tramo M amp Labreque R (1998)Processing prosodic and musical patterns Aneuropsychological investigation Brain and Language61 123ndash144

Peretz I amp Babaı M (1992) The role of contour and intervalsin the recognition of melody parts Evidence from cerebralasymmetries in musicians Neuropsychologia 30 277ndash292

Peretz I amp Morais J (1987) Analytic processing in theclassification of melodies as same or differentNeuropsychologia 25 645ndash652

Peretz I Morais J amp Bertelson P (1987) Shifting eardifferences in melody recognition through strategyinducement Brain and Cognition 6 202ndash215

Phillips C Pellathy T Marantz A Yellin E Wexler KPoeppel D McGinnis M amp Roberts T (2000) Auditorycortex accesses phonological categories An MEG mismatchstudy Journal of Cognitive Neuroscience 12 1038ndash1055

Picton T W Alain C Otten L Ritter W amp Achim A (2000)Mismatch negativity Different water in the same riverAudiology Neurootology 5 111ndash139

Platel H Price C Baron J C Wise R Lambert JFrackowiak R S Lechevalier B amp Eustache F (1997)The structural components of music perception Afunctional anatomical study Brain 120 229ndash243

Regnault P Bigand E amp Besson M (2001) Different brainmechanisms mediate sensitivity to sensory consonance andharmonic context Evidence from auditory event-relatedbrain potentials Journal of Cognitive Neuroscience 13241ndash255

Ridding M C Brouwer B amp Nordstrom M A (2000)Reduced interhemispheric inhibition in musiciansExperimental Brain Research 133 249ndash253

Rinne T Alho K Ilmoniemi M K Virtanen J amp NaatanenR (2000) Separate time behaviors of the temporal andfrontal mismatch negativity sources Neuroimage 1214ndash19

Saarinen J Paavilainen P Schroger E Tervaniemi M ampNaatanen R (1992) Representation of abstract attributesof auditory stimuli in the human brain NeuroReport 31149ndash1151

Satoh M Takeda K Nagata K Hatazawa J amp KuzuharaS (2001) Activated brain regions in musicians during anensemble A PET study Brain Research Cognitive BrainResearch 12 101ndash108

Scherg M Vajsar J amp Picton T W (1989) A source analysisof the late human auditory evoked potentials Journal ofCognitive Neuroscience 1 336ndash355

Schlaug G Jancke L Huang Y Staiger J F amp SteinmetzH (1995) Increased corpus callosum size in musiciansNeuropsychologia 33 1047ndash1055

Schlaug G Jancke L Huang Y amp Steinmetz H (1995) Invivo evidence of structural brain asymmetry in musiciansScience 267 699ndash701

Schneider P Scherg M Dosch H G Specht H J GutschalkA amp Rupp A (2002) Morphology of Heschlrsquos gyrus reflectsenhanced activation in the auditory cortex of musiciansNature Neuroscience 5 688ndash694

Shahin A Bosnyak D Trainor L J amp Roberts L E (2003)Enhancement of neuroplastic P2 and N1c auditory evokedpotentials in musicians Journal of Neuroscience 235545ndash5552

Tervaniemi M Ilvonen T Karma K Alho K amp NaatanenR (1997) The musical brain Brain waves reveal theneurophysiological basis of musicality in human subjectsNeuroscience Letters 226 1ndash4


Tervaniemi M Maury S amp Naatanen R (1994) Neuralrepresentations of abstract stimulus features in the humanbrain as reflected by the mismatch negativity NeuroReport5 844ndash846

Tervaniemi M Rytkonen M Schroger E Ilmoniemi R Jamp Naatanen R (2001) Superior formation of corticalmemory traces for melodic patterns in musicians Learningand Memory 8 295ndash300

Tesche C D Uusitalo M A Ilmoniemi R J Huotilainen MKajola M amp Salonen O (1995) Signal-space projections ofMEG data characterize both distributed and well-localizedneuronal sources Electroencephalography and ClinicalNeurophysiology 95 189ndash200

Tiitinen H May P Reinikainen K amp Naatanen R (1994)Attentive novelty detection in humans is governed bypre-attentive sensory memory Nature 372 90ndash92

Tillmann B Janata P amp Bharucha J J (2003) Activationof the inferior frontal cortex in musical priming BrainResearch Cognitive Brain Research 16 145ndash161

Trainor L J Desjardins R N amp Rockel C (1999) Acomparison of contour and interval processing in musiciansand nonmusicians using event-related potentials AustralianJournal of Psychology 51 147ndash153

Trainor L J McDonald K L amp Alain C (2002) Automaticand controlled processing of melodic contour and intervalinformation measured by electrical brain activity Journalof Cognitive Neuroscience 14 1ndash13

Trehub S E Trainor L J amp Unyk A M (1993) Music andspeech perception in the first year of life In H W Reese ampL P Lipsitt (Eds) Advances in child development andbehavior (Vol 24 pp 1ndash35) New York Academic Press

Tremblay K Kraus N amp McGee T (1998) The time courseof auditory perceptual learning Neurophysiological changesduring speech-sound training NeuroReport 9 3557ndash3560

Wallin N L Merker B amp Brown S (2000) The origins ofmusic Cambridge MIT Press

Winkler I Kujala T Tiitinen H Sivonen P Alku PLehtokoski A Czigler I Csepe V Ilmoniemi R J ampNaatanen R (1999) Brain responses reveal the learning offoreign language phonemes Psychophysiology 36 638ndash642

Zatorre R J (1985) Discrimination and recognition oftonal melodies after unilateral cerebral excisionsNeuropsychologia 23 31ndash41

Zatorre R J Evans A C amp Meyer E (1994) Neuralmechanisms underlying melodic perception and memoryfor pitch Journal of Neuroscience 14 1908ndash1919

Fujioka et al 1021

2002 Koelsch et al 2002 Ohnishi et al 2001 SatohTakeda Nagata Hatazawa amp Kuzuhara 2001 Halpernamp Zatorre 1999 Platel et al 1997 Zatorre Evans ampMeyer 1994) For example a functional magnetic reso-nance imaging (fMRI) study has shown that listening tomusic produces an enhanced activation in the planumtemporale and the left posterior dorsolateral prefrontalcortex in musicians in comparison to pure tones (Oh-nishi et al 2001) Long-term musical training has bothanatomical and functional consequences For exampleanatomical asymmetries in the auditory cortices of mu-sicians were reported using high-resolution MRI Theleft planum temporale was larger than the right inskilled musicians especially in those with absolute pitch(the ability to recognize and name the pitch of a musicaltone without reference to a comparison tone) (SchlaugJancke Huang amp Steinmetz 1995) Furthermore thespecific anterior subregion of the corpus callosum wasfound to be larger in musicians who commenced musi-cal training before the age of seven (Schlaug JanckeHuang Staiger amp Steinmetz 1995) Functional differ-ences in musicians have also been demonstrated mainlyin the sensorimotor and in the auditory systems Violin-ists practice rapid independent movement of the fingersof their left hand on the fingerboard of their instrumentevery day for many years Somatosensory-evoked mag-netic fields reveal the results of this experience in largercortical representations of the left-hand fingers of musi-cians in comparison to either the right-hand fingers ofmusicians or the fingers of control subjects who neverplayed a violin (Elbert Pantev Wienbruch Rockstroh ampTaub 1995) In contrast behavioral and functional MRIstudies on pianists (Jancke Shah amp Peters 2000 JanckeSchlaug amp Steinmetz 1997) have found less asymmet-rical activity in the motor cortices of musicians than inthose of control subjects since keyboard performancerequires similar motor control skills for both handsFurthermore transcranial magnetic stimulation (TMS)data demonstrated that bimanual motor activities inkeyboard musicians were less inhibited than in normalsubjects (Ridding Brouwer amp Nordstrom 2000) as ispredicted from the observation of an enlarged corpuscallosum in musicians (Schlaug Jancke Huang Staigeret al 1995) Musicians and nonmusicians also differ intheir evoked responses from auditory cortex A magne-toencephalography (MEG) study reveals that musiciansshow larger amplitude auditory evoked N1m responses(peaking at about 100 msec after stimulus onset) forpiano sounds (Pantev et al 1998) and an electroen-cephalography (EEG) study reveals that they showlarger P2 and N1c responses (Shahin Bosnyak Trainoramp Roberts 2003) Furthermore the effect of N1m en-hancement is specific to the instrument of training(Pantev Roberts Schulz Engelien amp Ross 2001) Eventrelated potential (ERP) studies have also demonstrateddifferences between musicians and nonmusicians inlater processing that is likely outside primary or second-

ary sensory areas Besson Faıta amp Requin (1994) andTrainor et al (1999) report enhanced late positive wavesbetween 300 and 600 msec Relations between function-al and anatomical measures are also emerging Schneideret al (2002) combined structural information from MRIand functional information from MEG to reveal a corre-lation between Heschlrsquos gyrus enlargement and en-hanced evoked magnetic field response in the latencyof 19ndash30 msec in musicians In summary the comparisonof brain responses between skilled musicians and naivesubjects can give us new insights into brain plasticityassociated with musical training

The ability to encode an acoustical context is reflectedelectrophysiologically in the mismatch negativity (MMN)component of ERPs The MMN and its magnetic coun-terpart the MMNm are elicited within about 100 to200 msec after any discriminable change that occurs in-frequently in a repeatedly presented auditory stimulieven when the stimulus is not attended (Picton et al2000 Naatanen amp Picton 1987 Naatanen 1992) TheMMN reflects processing in neural memory traces bywhich the auditory cortex handles representations ofthe recent acoustic past and its repetitive aspects Thesources of the MMN have been located mainly in thesupratemporal plane (Alho 1995) by dipole modeling(Scherg Vajsar amp Picton 1989) and scalp current den-sity (SCD) maps obtained from EEG (Giard PerrinPernier amp Bouchet 1990) and MEG (Levanen AhonenHari McEvoy amp Sams 1996) The contribution of fron-tal generators of MMN has been also suggested not onlyby SCD (Giard et al 1990) but also by lesion studies(Alain Woods amp Knight 1998 Alho Woods AlgaziKnight amp Naatanen 1994) MMN is elicited not onlyfor changes in single acoustic features but also for morecomplex and abstract features Recent studies show thatMMN can be obtained even by changes in an auditorypattern (Alain Cortese amp Picton 1999 Alain Woods ampOgawa 1994) and a change from ascending to descend-ing tone pairs (Saarinen Paavilainen Schroger Terva-niemi amp Naatanen 1992) Furthermore MMN is affectedby experience such as phoneme categorization in aparticular language (Phillips et al 2000 Cheour et al1998 Naatanen et al 1997) Even after a short time ofintense listening training the increase in MMN ampli-tude parallels the increased discrimination performancewith tone frequency discrimination (Menning Robertsamp Pantev 2000) or with foreign phoneme categories(Menning Imaizumi Zwitserlood amp Pantev 2002)These studies indicate that the conscious process of dis-crimination and the unconscious process of extractingchanges in memory traces interact and that these twoprocesses are affected by training

The goal of the present study is to investigate therelationship between long-term musical training andautomatic melodic processing Changes in both pitchcontour (Trainor McDonald amp Alain 2002 TervaniemiRytkonen Schroger Ilmoniemi amp Naatanen 2001 Paa-

Fujioka et al 1011


RESULTS








DISCUSSION






















the vague responses

Fujioka et al 1013



















line







Dipole Moment (nAm)


Left hemisphere



Right hemisphere










p lt 0001

Fujioka et al 1015





METHODS

Subjects


Stimuli










MEG Recordings
















does not change

Fujioka et al 1017


Data Analysis





Behavioral Test





Acknowledgments



REFERENCES
































Fujioka et al 1019






















































Fujioka et al 1021


RESULTS








DISCUSSION






















the vague responses

Fujioka et al 1013



















line







Dipole Moment (nAm)


Left hemisphere



Right hemisphere










p lt 0001

Fujioka et al 1015





METHODS

Subjects


Stimuli










MEG Recordings
















does not change

Fujioka et al 1017


Data Analysis





Behavioral Test





Acknowledgments



REFERENCES
































Fujioka et al 1019






















































Fujioka et al 1021




















the vague responses

Fujioka et al 1013



















line







Dipole Moment (nAm)


Left hemisphere



Right hemisphere










p lt 0001

Fujioka et al 1015





METHODS

Subjects


Stimuli










MEG Recordings
















does not change

Fujioka et al 1017


Data Analysis





Behavioral Test





Acknowledgments



REFERENCES
































Fujioka et al 1019






















































Fujioka et al 1021



















line







Dipole Moment (nAm)


Left hemisphere



Right hemisphere










p lt 0001

Fujioka et al 1015





METHODS

Subjects


Stimuli










MEG Recordings
















does not change

Fujioka et al 1017


Data Analysis





Behavioral Test





Acknowledgments



REFERENCES
































Fujioka et al 1019






















































Fujioka et al 1021






Dipole Moment (nAm)


Left hemisphere



Right hemisphere










p lt 0001

Fujioka et al 1015





METHODS

Subjects


Stimuli










MEG Recordings
















does not change

Fujioka et al 1017


Data Analysis





Behavioral Test





Acknowledgments



REFERENCES
































Fujioka et al 1019






















































Fujioka et al 1021





METHODS

Subjects


Stimuli










MEG Recordings
















does not change

Fujioka et al 1017


Data Analysis





Behavioral Test





Acknowledgments



REFERENCES
































Fujioka et al 1019






















































Fujioka et al 1021






MEG Recordings
















does not change

Fujioka et al 1017


Data Analysis





Behavioral Test





Acknowledgments



REFERENCES
































Fujioka et al 1019






















































Fujioka et al 1021


Data Analysis





Behavioral Test





Acknowledgments



REFERENCES
































Fujioka et al 1019






















































Fujioka et al 1021



Acknowledgments



REFERENCES
































Fujioka et al 1019






















































Fujioka et al 1021






















































Fujioka et al 1021

musical training enhances automatic encoding of melodic ... · of melodic contour and interval...

Documents