295tests of musical aptitude (for a review, see Ref. 4) and by hit rates and reaction times.5–8In
addition, the degree of perceptual similarity between different musical instrument timbres
highly correlates with the MMN amplitude.^9 This correlation between the MMN para-
meters and behavioural responses imply that preattentive neural functions determine the
accuracy of the subsequent attentive processes.2,3,10
Since the MMN is elicited without attention, it received particular interest during past
years among clinicians who need to evaluate the integrity of neurocognitive functions in
patients unable or unwilling to participate in a neuropsychological (behavioural) testing.
For instance, it was found that the MMN elicitation predicts the recovery of consciousness
in comatose patients11,12) (for reviews, see Refs 13, 14). In addition, the deficits in MMN
elicitation might help target perceptual training or rehabilitation in milder perceptual dis-
orders. For instance, recently it was revealed that automatic discriminative functions of
dyslexic adults are worse than those of control subjects.15–18In addition, the MMN is also
correlated with the improvement of the reading performance in dyslexic children.^19
Since differences in subjects’ attention or motivation might confound studies in music
psychology especially in between-group comparisons, the MMN recordings offer a promis-
ing means in determining the accuracy of neural sound representations prior to involve-
ment of the perceptual and cognitive functions also in music studies. In the following, an
overview on studies using musical sounds and sound successions to investigate the auto-
matic neural sound processing will be given. In addition, recent studies comparing musi-
cians and nonmusicians will be reviewed.
MMN generators for speech vs musical sounds
MMN is generated mainly in the primary auditory cortex or in its immediate vicinity. This
has been indicated by several brain research methods such as magnetoencephalography
(MEG),20,21intracranial electrode recordings,22,23positron emission tomography (PET),^24
and functional magnetic resonance imaging (fMRI).25–27In addition, this is reflected in
electric recordings by polarity reversal above the Sylvian fissure from the fronto-central
negative maximum to positivity in mastoids leads (when the nose reference is used) (for a
review, see Ref. 28).
In late 1990s, a whole-head MEG experiment was conducted to determine whether the
auditory cortex has spatially distinct areas for encoding musical vs phonetic sounds.^29 The
subjects, while watching a silent movie with subtitles, were presented with frequent and
infrequent phonemes (/e/ vs /o/) or chords (A major vs A minor). These phonetic and
musical stimuli were matched in complexity as well as in the magnitude of the frequency
change embedded in them. It was found that the source of the MMNm elicited by the
deviant phoneme or chord was located posteriorly to the source of an earlier P1m compo-
nent (see footnote a). In addition, the MMNm source for a phoneme change was located
superiorly to that of the chord change. On the contrary, P1m sources did not differ between
aEarly exogenous ERP components and their MEG equivalents are mainly determined by the physical sound
characteristics, for instance, the intensity and rise time, while endogenous ERP/MEG components with longer
onset latencies are modulated by cognitive and attentional factors.^2