4

Ann. N.Y. Acad. Sci. 999: 4-14 (2003). 2003 New York Academy of Sciences.
doi: 10.1196/annals.1284.001
Music and the Brain
ROBERT J. ZATORRE
Montreal Neurological Institute. McGill University, Montreal, Canada
ABSTRACT: The aim of this paper is to illustrate how studying music from a
neuroscience perspective may be a valuable way to probe a variety of complex
cognitive functions and their neural substrate. Three different sets of issues are
described. First, studies dealing with the brain correlates of musical imagery
are discussed. This topic is of interest in that it illustrates how subjective sen-
sations may be studied via objective techniques, and gives insight into neural
systems associated with internal phenomena. Second, some findings pertaining
to absolute pitch are presented. Absolute pitch is a useful example of a highly
specific cognitive skill that is unevenly distributed in the population. Examina-
tion of its neural basis helps to understand aspects of memory function and
points to ways to explore individual differences in brain organization that un-
derlie differential skills. The final topic, music and emotion, has not been the
subject of much systematic research, but it is of great interest because it inter-
sects with a large literature on the neuroscience of affective processing. Find-
ings from some studies indicate that music may engage systems concerned with
biological reward, raising interesting but so far unanswered questions about
the broader role of music in human experience.
KEYWORDS: music; brain
INTRODUCTION
Why should music interest a neuroscientist as a legitimate object oI scientiIic
study? Such a question might be posed in two ways by persons coming Irom diIIer-
ent intellectual traditions. From a scientiIic perspective, we might wonder whether
studying something as complex and 'nonessential as music is likely to yield mean-
ingIul insights, the implicit objection being that more Iundamental cognitive or
perceptual Iunctions are more valid to Iocus on. From a humanistic perspective, one
might harbor suspicions oI a perceived reductionist agenda in science, to oversim-
pliIy artistic endeavors as mere products oI mechanistic organisms. Such points oI
view may be, to some extent, caricatures oI the reactions to the neuroscience oI
music, but they do represent important critiques that should not be ignored. Happily
Ior those within this Iield, as exempliIied by this volume oI proceedings, these ques-
tions are increasingly being addressed to the satisIaction oI both scientiIic and
artistic communities. Our hope is that the contents oI these papers will demonstrate
the value oI music as a window onto complex brain Iunctions, while at the same time
illustrating how a scientiIic understanding oI music can yield deep insights into the
nature oI human thought and expression.
Address Ior correspondence: Dr. Robert J. Zatorre, Montreal Neurological Institute, 3801 Uni-
versity St., Montreal, Quebec H3A 2B4, Canada. Voice: 1-514-398-8903; Iax: 1-514-398-1338.
5 ZATORRE: MUSIC AND THE BRAIN
In this essay, my aim is to highlight the diIIerent approaches that can be taken to
the study oI music and brain Iunction and how they throw light on three distinct
aspects oI the neural substrates oI cognition, rather than to provide a detailed review
oI any one research question. Because music encompasses so many aspects oI
human cognition, it provides a rich source oI materials Ior psychologists and neuro-
scientists to investigate. This very complexity also renders the study oI music inher-
ently complicated, however. It is thereIore oI great importance to deIine the speciIic
aspect oI musical Iunction to be studied and, wherever possible, to identiIy the cog-
nitive components associated with those Iunctions. The research to be described in
this paper comes Irom three separate threads oI work that have been carried out in
my laboratory over the last Iew years. The Iirst set oI studies deal with musical
imagery; the second, with absolute pitch; and the third, with music and emotion.
Each oI these topics touches upon separate questions, yet they have in common that
musical processes are central to them. This work also needs to be put in perspective
as part oI a much broader research agenda, which is sketched out in part by the other
papers in this volume. It is thereIore important to remember, Iirst, that any one set oI
studies cannot be interpreted without the context oI other inIormation Irom many
other domains; second, that the study oI music and its neural correlates is clearly still
in its early years and consequently that many more questions than answers are
apparent at this point. The latter point is precisely what makes this Iield oI inquiry
dynamic; it is thereIore a good choice Ior an investigator who values stimulating
debate over clear-cut conclusions.
MUSICAL IMAGERY
The Iirst set oI studies to be discussed deal with a cognitive Iunction that is inher-
ently diIIicult to study because it is entirely internal and does not necessarily have
observable concomitants. This work, done over the last Iew years in collaboration
with Andrea Halpern oI Bucknell University, has beneIited Irom prior research in
cognitive psychology dealing with visual imagery.
1,2
Until the 1960s, mental imag-
ery was generally considered suspect as an object oI serious research precisely
because oI its subjective, internal nature. The behaviorist model eschewed unobserv-
able processes as not subject to scientiIic scrutiny, and rightly so, Ior earlier attempts
at characterizing mental processes too oIten devolved into nothing more than 'arm-
chair psychology. Starting with the work oI Shepard,
3
however, it was shown that
by clever experimental manipulations, it was indeed possible to measure objective
behavioral indices oI what otherwise would be inscrutable internal events. Thus, Ior
instance, the demonstration that the time taken to make a decision on a rotated letter
was strictly proportional to the degree oI rotation indicated that a visual imagery pro-
cess could be isolated and measured. This and other Iindings quickly became clas-
sics in the Iield oI cognition, because they powerIully showed how we could probe
internal processes by studying their external correlates. In this sense, cognitive psy-
chology was shown to be no diIIerent Irom other Iields oI science, where unobserv-
able Iorces or events are examined indirectly via the traces they leave (think particle
physics).
Most work in mental imagery has involved vision. Yet, both introspection and
experimentation reveal that most people experience auditory imagery,
4
in particular
6 ANNALS NEW YORK ACADEMY OF SCIENCES
musical imagery, which might be described intuitively as simply 'hearing music in
one`s mind. The phenomenology oI this eIIect suggests that musical imagery
retains many oI the properties oI real perception (as with visual images), particularly
its temporal characteristics, but that it is not conIusable with real hearing. Based on
this, Halpern designed a series oI studies showing that judgments oI imagined music
generally Iollowed similar patterns to judgments oI heard music. For example, in an
important set oI experiments,
5
it was shown that iI persons are asked to judge the
pitch change corresponding to two syllables in a well-known song, they can do so
Iairly accurately, demonstrating that they have access to an internal representation oI
the pitch structure. That, in itselI, is useIul because it serves to veriIy a person`s
response, because there is an objectively correct answer. But more importantly,
Halpern showed that the time taken to make the response was a near-linear Iunction
oI the time between probed items in the real music. This demonstration oI chron-
ometry indicates that the imagery event unIolds in something like real time, much as
visual imagery processes are also time dependent. This is important, because it sug-
gests an analogous mental process in imagery and perception. (It is important to note
that the alternative, a nontime-dependent process, would be easy to posit: it could
certainly be the case, Ior instance, that the pitches oI a tune are stored in memory in
something like a look-up table, where it would take no longer to search Ior two items
close together than two items Iar apart. That this is not how imagery proceeds tells
us that the imaginal event unIolds over time just as real perception does.)
The behavioral similarity between imagery and perception leads naturally to the
hypothesis that these two Iunctions share an underlying neural representation. This
hypothesis was put to the test in a series oI studies that can brieIly be summarized.
The Iirst attempt to address this question made use oI a classic paradigm in neuro-
psychology: the lesion-deIicit approach. The basic idea is simply that damage to cer-
tain areas oI the brain leads to speciIic disturbances, which yield clues to the
Iunction oI the damaged areas. In the context oI the imagery hypothesis, the predic-
tion would be that iI both processes depend on shared neural territory, then a lesion
that disrupts perception should also disrupt imagery. It had previously been estab-
lished that damage to the superior temporal gyrus, a brain area in the temporal lobe
that contains the auditory cortices, leads to perceptual deIicits on musical tasks, es-
pecially when that damage is in the right cerebral hemisphere. (For a review, see ReI.
6.) When we tested patients with cortical excision within these areas, we Iound that
perception and imagery suIIered a common Iate aIter damage to the right temporal
lobe: perIormance was poorer than normal on a pitch-judgment task whether per-
Iormed on a heard tune or one that was merely imagined. By contrast, patients with
similar lesions on the leIt side perIormed as well as controls on both tasks.
7
This association between the two Iunctions was tested Iurther in two studies using
Iunctional neuroimaging. These techniques have become the mainstay oI a great deal
oI work in cognitive neuroscience and have added immeasurably to our ability to
make inIerences about human brain Iunction by allowing scientists to probe how
regional brain activity patterns change as a Iunction oI stimulus characteristics or
task demands. In the Iirst oI several experiments on musical imagery, listeners were
scanned while perIorming a version oI the pitch-judgment task just described.
8
The
principal result conIirmed earlier Iindings that activity was observed within the
superior temporal gyri on both sides, not only when listening and judging heard
tunes, but also when simply imagining the tunes without overt input. This demon-
7 ZATORRE: MUSIC AND THE BRAIN
stration oI auditory cortex activity in the absence oI an external stimulus is important
in supporting the contention that neural events within sensory areas are responsible
Ior the subjective experience associated with imagery, a conclusion that is also sup-
ported in other modalities. To explore Iurther the issue oI lateralization oI response,
a second neuroimaging experiment was carried out in which the listener was asked
to imagine the continuation oI a Iamiliar musical excerpt aIter having heard its Iirst
Iew notes.
9
The data again showed activity within the superior temporal gyrus when
imagery was involved, as compared to a condition in which tones were perceived
without imagery. This time, the pattern showed strong lateralization to the right side,
most likely because the tunes were selected to avoid all verbal associations, adding
additional support Ior the predominance oI right-hemisphere mechanisms in melodic
processing.
In more recent research (Halpern et al., submitted), we approached the problem
Irom a somewhat diIIerent perspective. The aim was to test the generality oI the
musical imagery phenomenon using materials other than songs. Songs have both a
sequential component as well as semantic associations in long-term memory, which
may give rise to complex patterns oI brain activity; we wanted to avoid those Iactors.
We thereIore selected timbre as the dimension to Iocus on. During IMRI scanning,
listeners with moderate musical training were asked to imagine the timbres oI two
instruments cued by a visual word or they listened to the sounds oI musical instru-
ments in the perceptual control condition. To obtain a behavioral index oI successIul
imagery, subjects were asked to make a dissimilarity rating oI the two sounds in
question. Comparison oI these ratings using a multidimensional scaling approach
showed signiIicant similarities in both perceived and imagined ratings, indicating
that subjects can indeed generate internal representations oI timbre. The IMRI data
were consistent with prior Iindings in that brain activity elicited during imagery was
observed in portions oI the auditory cortex and overlapped with the activity elicited
by real perception. In addition, there was again a tendency Ior a greater response in
the right than in the leIt auditory cortex.
Thus, this additional study conIirms and extends the general conclusion that
imagery does depend on a partially shared neural substrate that Iorms part oI the
perceptual analysis mechanism. When we experience an imaginal event, this reIlects
an engagement oI neural systems that are involved in perception. More generally,
these studies are oI interest because they illustrate one way to probe the complex
inner workings oI conscious thought processes. It is now possible to state with a
certain degree oI conIidence that we have the tools to begin to unravel what earlier
generations oI scientists had thought closed to scientiIic inquiry. However, much
work lies ahead, and it should be emphasized that no one technique will be suIIicient
to give us more than a small piece oI the puzzle. Among the many questions that
remain to be explained are how the neural traces in sensory cortex that seem to
correspond to imagery are activatedwhat are the pathways and sources oI this pro-
cess? Another question that begs to be studied is how individuals diIIer in their de-
gree oI imagery ability and what in their brains might explain the variability. We
need think no Iurther than Beethoven`s celebrated ability to compose while utterly
deaI to realize the extent to which some people can generate internal representations
not only oI previously heard sounds, but also oI new ones. How does this happen?
An inquiry into this process would undoubtedly tell us something interesting about
creativity, that is, how the mind can recombine known elements into novel patterns.
8 ANNALS NEW YORK ACADEMY OF SCIENCES
ABSOLUTE PITCH
A consideration oI how individuals diIIer Irom one another in any domain inevi-
tably leads to the question oI nature versus nurture. This is such a general issue in
all oI psychology and biology that we can only touch upon it here. But one circum-
stance in which these issues come into particularly clear Iocus concerns the phenom-
enon oI absolute pitch. A review oI the many threads related to this phenomenon is
beyond our scope (Ior reviews see ReIs. 10 and 11), but what is oI relevance Ior the
present discussion is the link between absolute pitch and brain Iunction. Indeed, this
phenomenon has begun to generate considerable interest in some scientiIic circles,
because it may yield some clues to brain organization associated with special skills
and also because oI growing hints that there may be a genetic component to its
development (Ior a review, see ReI. 12). Moreover, it is well-established that the
development oI absolute pitch is highly dependent on the age at which musical train-
ing is begun.
13,14
Thus, absolute pitch may serve as a useIul paradigm Ior under-
standing the interactions between genetic mechanisms, their role in brain
development, and how environmental input modiIies and inIluences behavior by
shaping or tuning neural processes that depend on these genetic Iactors.
The contribution made by Iunctional imaging to this domain has so Iar been mod-
est, but it seems destined to grow. One oI the Iirst questions addressed in our own
laboratory concerning absolute pitch was simply to understand the pattern oI brain
activity that distinguished a possessor oI absolute pitch Irom musicians with similar
training but relative pitch.
15
To this end we measured regional cerebral blood Ilow
among possessors and non-possessors under two conditions: while listening to pairs
oI tones with no explicit instructions and while listening to the same tones and
determining the musical interval Iormed by them. We reasoned that musicians with
absolute pitch would engage their ability automatically even when listening
'passively and that this condition would reveal the greatest diIIerence across
groups. Making relative judgments oI musical intervals, however, should yield more
similar patterns oI brain activity, although musicians with absolute pitch might also
use pitch labels in this task.
The principal Iinding oI this study was that subjects with absolute pitch demon-
strated signiIicant brain activity in a region oI the Irontal cortex while listening to
tones without explicit instructions, whereas nonpossessors did not. We might be
tempted to conclude Irom this that the area in question is the 'absolute pitch region.
But the pattern oI results Irom the interval-judgment task, in Iact, showed a similar
pattern oI activity in this Irontal area Ior both absolute and relative pitch musicians.
Thus, the most parsimonious explanation is that the activity in this region seems to
be an index oI a neural process related to labeling. Under conditions in which label-
ing is only possible by absolute pitch subjects, only these subjects show the eIIect,
whereas when labeling oI an interval is possible, both groups show similar activity
in this area. The particular Irontal cortical region in question is known Irom other
research to be concerned with establishing and maintaining conditional associations
in memory,
16
that is, the ability to make a speciIic response in relation to a given
stimulus. Thus, it constitutes a logical candidate area Ior the link between a pitch and
its label in absolute pitch subjects. Additional support Ior this explanation is that the
same region oI Irontal cortex can be recruited, even in nonmusicians, once they are
taught to identiIy chords with arbitrary labels.
17
Thus, it seems that Ior whatever as
9 ZATORRE: MUSIC AND THE BRAIN
yet unknown reason, the associative Iunction oI this part oI the Irontal lobe is some-
how Iacilitated in these individuals so that they Iorm tone-label associations more
readily. It should be pointed out that the labeling, while usually oI a verbal nature,
need not be limited to a note name, because absolute pitch may also be demonstrated
by any number oI other, nonverbal codes;
18
in particular, auditory imagery and
sensorimotor responses such as those associated with playing a tone on a given
instrument are oIten reported among musicians who have absolute pitch.
So what brain mechanism(s) do distinguish people with absolute pitch? At
present, the answer to this question remains unknown, but an interesting anatomical
correlate has been Iound that bears mention. The picture emerging Irom several stud-
ies is that there are signiIicant diIIerences in the degree oI lateral asymmetry oI
cortical structures concerned with auditory processing (in the posterior portion oI the
superior temporal gyrus). Persons with absolute pitch tend to show a greater degree
oI anatomical hemispheric asymmetry oI this area oI the brain as compared to either
musicians without absolute pitch or nonmusicians.
15,19
However, how to interpret
this eIIect remains uncertain; whereas initial indications were that the enhanced
asymmetry might reIlect greater growth oI leIt-hemisphere auditory cortical struc-
tures, the latest evidence suggests instead that it may be a reduced volume oI right-
hemisphere auditory cortical areas that accounts Ior the diIIerences.
20
Thus, perhaps
it is not simply the case that 'bigger is better, but instead that there may be more
complex interactions between gross morphological Ieatures and the underlying
cortical Iunction. Regardless oI the details oI this line oI investigation, anatomical
diIIerences are important to the extent that they may serve as an indication oI possi-
ble genetic or epigenetic Iactors. Because asymmetries in this part oI the brain are
reported to exist even prior to birth,
21
we need to know iI they are also somehow
markers oI a propensity Ior developing absolute pitch under appropriate environ-
mental conditions. Longitudinal studies to unravel this issue are thereIore oI great
interest. More generally, thereIore, the study oI this otherwise somewhat arcane
phenomenon may reveal important knowledge about how speciIic cognitive and
perceptual skills depend on neural development and its interaction with the environ-
ment, a problem that is at the heart oI essentially all oI cognitive neuroscience.
MUSIC AND EMOTION
The third topic selected to illustrate the value and interest oI studying music Irom
a neuroscience perspective is emotion. Studying music and emotion is inherently
tricky Ior a number oI reasons, and only relatively recently are eIIorts being made to
achieve a more systematic understanding oI this complex phenomenon. (For a recent
overview oI this Iield, see ReI. 22.) Ignoring the aIIective aspects oI musical experi-
ence (which music cognition has oIten been guilty oI) may be a dangerous approach,
however, because we may be missing some oI the most salient and important aspects
oI human response to music. Once again, the contribution oI brain imaging to the
knowledge base is so Iar rather limited, and what little has been done tends to oIIer
more questions than answers. Yet, as mentioned at the outset oI this essay, at this
early stage in our understanding oI these phenomena, it may be more valuable to
generate appropriate questions than to provide premature answers. Work on these
questions done in my laboratory by Anne Blood represents our Iirst attempt to
10 ANNALS NEW YORK ACADEMY OF SCIENCES
explore the neural substrates oI music and emotion, and we hope that it will serve to
spur others to address the topic as well.
Among the Iactors that contribute to the diIIiculty oI studying emotions and
music in a scientiIic context is that emotional responses to music tend to be idio-
syncratic and heterogeneous, and depend on a variety oI complex and diIIicult-to-
control individual sociocultural, historical, educational, and contextual variables.
This presents a particular problem Ior examining brain responses to music-elicited
emotion, because iI reactions vary widely Ior unknown reasons, it will be diIIicult to
set up a consistent experimental protocol. However, there are approaches that we can
use to mitigate these problems. In the Iirst study Irom our group on this topic,
23
we
reasoned that one way to elicit consistent responses would be to make use oI disso-
nance. Whereas people oIten disagree on what music they like, there is a certain
amount oI agreement that highly dissonant music tends to be unpleasant. We need to
determine what is meant by dissonance, because in the appropriate tonal context
dissonance serves a critical role in creating esthetically pleasant music; an entire
literature exists in music cognition and psychoacoustics dealing with these issues.
(For a general review, see ReI. 24.) But intuitively, it is clear that iI we listen to music
with many 'wrong notes, it simply does not sound very good!
Based on this straightIorward, albeit potentially simplistic consideration, we
designed a stimulus set in which a tonal melody was accompanied by chords that
varied in their degree oI dissonance. The advantage oI this approach is that the
degree oI dissonance could be varied parametrically to measure changes in brain
activity that correlate with the variable oI interest; this method avoids having to
choose an appropriate baseline condition (oIten a problem in Iunctional neuro-
imaging studies),
25
because any changes observed must be related to the parameter
varied. Subjects were scanned while listening to these varying levels oI dissonant
music and making judgments about their pleasantness, which generally conIirmed
that the more dissonant the accompaniment, the less pleasant the experience.
The pattern oI brain activity associated with increasing and decreasing disso-
nance revealed distinct sets oI correlations in parahippocampal and orbitoIrontal
regions, respectively. These sites are typically considered paralimbic cortical areas:
cortex that stands in an intermediate anatomical relationship between association
areas and the limbic system. One way to think oI these areas, thereIore, is that they
mediate between perceptual and cognitive representations, on the one hand, and
emotion, on the other. Many prior studies oI emotion in the context oI both neuro-
imaging and pathology have identiIied orbitoIrontal and adjacent Irontal areas as
well as parahippocampal areas as being related to various aspects oI mood and pleas-
ant versus unpleasant emotional responses.
2629
It thereIore seems that the emotion-
al eIIects oI music are mediated via the same areas that mediate emotion elicited by
other stimuli or situations. In addition, reciprocal regional interactions were Iound
between increasing versus decreasing dissonance conditions. That is, increasing
activity in regions active during negative emotions was observed to be associated
with a corresponding decrease in activity in regions that are active during positive
emotions. This type oI eIIect suggests a Iunctional interaction between regions
mediating opposite emotions, which in turn may indicate that music exerts its eIIects
not only by evoking one reaction, but also by inhibiting incompatible ones.
Having explored aspects oI negative aIIect elicited by music, we turned our atten-
tion to positive responses. Here we were once again Iaced not only with the issue oI
11 ZATORRE: MUSIC AND THE BRAIN
heterogeneity oI responses, but also with the problem oI subjectivity. How do we
know whether one person`s reaction to music is Iundamentally similar to another`s,
even iI they both say that they enjoy what they are listening to? This problem is a
variant oI the issue brought up in the context oI imagery, and there is no one solution
to exploring subjective states with objective methods (this is, aIter all, psychology`s
principal task). However, one phenomenon associated with music that helps greatly
in dealing with this problem is the so-called 'musical chills or 'shivers that many
people report experiencing on hearing certain musical passages. This eIIect has been
documented by behavioral measures.
30
It occurs in a reasonably large proportion oI
the population, including persons with no Iormal musical training; it also has the
advantage oI being elicited relatively reliably. Moreover, emotions elicited by music
can be associated with objective physiological markers, which thereIore serve as an
external index oI the internal phenomenon.
31,32
The chills eIIect is experienced as a
very positive emotion and is sometimes described as ecstatic or euphoric by many
individuals. Thus, it seems to capture one oI the most intense aspects oI the aIIective
response to music.
But what is the neural basis Ior this response? To address this question we
measured the brain activity associated with musically elicited chills using positron
emission tomography.
33
First, we had the test subjects select passages oI instrumen-
tal music that consistently gave them this Ieeling at a certain speciIic point in time.
We careIully selected both the subjects and their music, such that we Iound stimuli
that elicited a strong chills response in one person, while leaving another person
unaIIected. The latter served as our control condition, in that any given stimulus
served as the experimental condition Ior one person and the control Ior another; thus,
across the entire subject sample the chills-inducing and neutral musical samples
were completely balanced. We obtained subjective numerical ratings on the intensity
oI the chills experience Irom each subject aIter scanning and also measured a
number oI psychophysiological responses during scanning, including respiration,
heart rate, and muscle tension.
The participants reported Ieeling the chills sensation on an average oI 77 oI the
trials, which is suIIicient Ior purposes oI analysis and a reasonable value considering
that the listening conditions inside a brain scanner tend not to be as conducive to pos-
itive emotional responses as one`s Iavorite easy chair Ior music listening! Important-
ly, the psychophysiological measures were consistent with the subjective ratings oI
chills, so that when subjects reported Ieeling chills, there were signiIicant increases
in all three variables.
Several complementary approaches were taken to analyze the data. One oI the
Iindings that emerged most consistently was that several brain areas were signiIi-
cantly engaged as a Iunction oI increasing subjective intensity oI the chills response.
These eIIects could not be attributed to diIIerences in Iamiliarity among the selI-
selected and control stimuli (a potential conIound), because they were largely
preserved even when data Irom the selI-selected music conditions alone were
analyzed. OI greatest interest in the complex neural response observed were brain
areas previously implicated by other types oI studies in response to highly rewarding
or motivationally important stimuli. The circuitry implicated, which includes areas
in the dorsal midbrain, ventral striatum (which contains the nucleus accumbens),
insula, and orbitoIrontal cortex, among others, was correlated with the subjective
experience oI chills. In addition, and reminiscent oI the previous study on disso-
12 ANNALS NEW YORK ACADEMY OF SCIENCES
nance, reciprocal interactions were observed such that portions oI the amygdala, a
brain region implicated in Iear and negative emotion,
34,35
showed activity decreases
as the chills response became more intense. The amygdala receives inhibitory pre-
synaptic input Irom cholinergic neurons intrinsic to the nucleus accumbens,
36
sug-
gesting a possible mechanism Ior decreased activity in these regions as a
consequence oI activity increases in ventral striatum.
The pattern oI activity observed in correlation with music-induced chills is
similar to that observed in other brain imaging studies oI euphoria and/or pleasant
emotion. For example, activity in the nucleus accumbens, dorsal midbrain, and
insula has been reported to increase, and that in the leIt amygdala and ventromedial
preIrontal cortex to decrease in response to cocaine administration in cocaine-depen-
dent subjects.
37
In addition, animal studies support a critical role Ior ventral stria-
tum, several midbrain areas, amygdala, and medial preIrontal cortex in circuitry
underlying reward processes, including hedonic impact, reward learning, and moti-
vation.
36,38
Animal studies oI endogenous reward in response to natural stimuli,
such as Iood and sexual stimulation, also show involvement oI brain activity in
similar regions.
39,40
Activity in these regions in relation to reward processes is-
known to involve dopamine and opioid systems as well as other neurotransmitters.
The pleasant experience oI chocolate consumption in humans has also been corre-
lated with activity in midbrain, insula, and ventromedial and orbitoIrontal cortex.
41
In essence, the Iindings indicate that under some particular conditions, the plea-
surable Ieeling elicited by listening to music seems to depend on a neural substrate
that is primarily associated with biologically signiIicant environmental events and is
present in many other species. It is easy to understand, in an evolutionary Iramework,
why such reward mechanisms exist Ior certain stimuli. Intake oI Iood is required Ior
survival oI the individual, and sexual reproduction is required Ior survival oI the
species; hence, both types oI stimuli induce pleasure, and most animals are thereIore
strongly motivated to seek them out. Certain drugs, such as cocaine, appear to mimic
the rewarding Ieatures oI natural stimuli, no doubt leading to addiction and other
negative consequences. But music is not a substance with pharmacological proper-
ties. Neither can music be said to be truly essential Ior survival; iI deprived oI music,
people will neither die oI starvation nor Iail to reproduce. Yet, we see evidence that
it recruits some oI the same brain systems associated with these more basic biolog-
ically driven stimuli.
Whether music is unique in this respect remains to be seen; it may be one oI a
class oI human constructs that elicit pleasure by coopting ancient neural systems via
inputs Irom the neocortex. We know Irom much other research that most cognitive
aspects oI music (pattern processing, working memory, and so Iorth) depend on neo-
cortical systems in auditory areas oI the brain and in Irontal and parietal cortex. (For
an overview, see ReI. 42.) In Iact, damage to cortical areas that results in loss oI per-
ceptual capacity also abolishes aIIective reactions.
43
But music thus presents an
interesting puzzle. Appreciation and understanding oI music clearly depends on
learning, cultural, and even social Iactors (part oI the reason why people diIIer,
sometimes radically, in their preIerences). Thus, research into music and emotion
opens up some interesting questions not only Ior the neuroscientist, but also perhaps
Ior the musicologist and even the philosopher. It may prove to be a paradigm Ior the
understanding oI interactions between higher-level cognition and subcortically
mediated aIIective responses.
13 ZATORRE: MUSIC AND THE BRAIN
CONCLUSION
This highly selective overview highlights how music can illuminate a variety oI
distinct aspects oI complex human cognitive Iunctions and their neural correlates.
The rest oI this volume gives many more examples oI the important issues that have
been tackled by others in this domain. The Iield is now beginning to show some
evidence oI maturity, in that there is a larger body oI literature on which to draw and
also in that investigators are increasingly able to integrate the knowledge gained with
more established Iindings in psychology, neuroscience, or other Iields. It is hoped
that this volume will contribute towards this evolution by summarizing the state oI
knowledge and, more importantly perhaps, by indicating the still large gaps in our
understanding that Iuture researchers will attempt to cover. Finally, as the theme oI
the meeting underscored, the continued interaction oI musicians and scientists will
be important, as the study oI music and neuroscience is mutually revealing.
REFERENCES
1. FARAH, M.J. 1988. Is visual imagery really visual? Overlooked evidence Irom neuro-
psychology. Psychol. Rev. 95: 307317.
2. KOSSLYN, S.M., G. GANIS & W.L. THOMPSON. 2001. Neural Ioundations oI imagery.
Nature Rev. Neurosci. 2: 635642.
3. SHEPARD, R.N. & J. METZLER. 1971. Mental rotation oI three-dimensional objects. Sci-
ence 171: 791793.
4. REISBERG, D., Ed. 1992. Auditory Imagery. Lawrence Erlbaum. Hillsdale, NJ.
5. HALPERN, A.R. 1988. Mental scanning in auditory imagery Ior tunes. J. Exp. Psychol.
LMC 14: 434443.
6. ZATORRE, R.J., P. BELIN & V.B. PENHUNE. 2002. Structure and Iunction oI auditory cor-
tex: music and speech. Trends Cognit. Sci. 6: 3746.
7. ZATORRE, R.J. & A.R. HALPERN. 1993. EIIect oI unilateral temporal-lobe excision on
perception and imagery oI songs. Neuropsychologia 31: 221232.
8. ZATORRE, R.J., A.R. HALPERN, D.W. PERRY, et al. 1996. Hearing in the mind`s ear: a
PET investigation oI musical imagery and perception. J. Cognit. Neurosci. 8: 2946.
9. HALPERN, A.R. & R.J. ZATORRE. 1999. When that tune runs through your head: a PET
investigation oI auditory imagery Ior Iamiliar melodies. Cereb. Cortex 9: 697704.
10. WARD, W.D. 1999. Absolute pitch. In The Psychology oI Music, 2nd Ed. D. Deutsch,
Ed. :265298. Academic Press. San Diego.
11. TAKEUCHI, A. & S. HULSE. 1993. Absolute pitch. Psychol. Bull. 113: 345361.
12. ZATORRE, R.J. 2003. Absolute pitch: a paradigm Ior understanding the inIluence oI
genes and development on neural and cognitive Iunction. Nature Neurosci. 6: 692
695.
13. BAHARLOO, S., S. SERVICE, N. RISCH, et al. 2000. Familial aggregation oI absolute
pitch. Am. J. Hum. Genet. 67: 755758.
14. MIYAZAKI, K. 1988. Musical pitch identiIication by absolute pitch possessors. Percept.
Psychophysiol. 44: 501512.
15. ZATORRE, R.J., D.W. PERRY, C.A. BECKETT, et al. 1998. Functional anatomy oI musical
processing in listeners with absolute pitch and relative pitch. Proc. Natl. Acad. Sci.
USA 95: 31723177.
16. PETRIDES, M. 1995. Functional organization oI the human Irontal cortex Ior mnemonic
processing: evidence Irom neuroimaging studies. Ann. N.Y. Acad. Sci. 769: 8596.
17. BERMUDEZ, P. & R.J. ZATORRE. 2002. Conditional associative memory Ior musical stim-
uli in non-musicians: relationship to absolute pitch. Soc. Neurosci. Abstr.: Abstract
Viewer/Itinerary Planner. Society Ior Neuroscience, Online. Washington, DC.
18. ZATORRE, R.J. & C. BECKETT. 1989. Multiple coding strategies in the retention oI
musical tones by possessors oI absolute pitch. Mem. Cognit. 17: 582589.
14 ANNALS NEW YORK ACADEMY OF SCIENCES
19. SCHLAUG, G., L. JNCKE, Y. HUANG, et al. 1995. In-vivo evidence oI structural brain
asymmetry in musicians. Science 267: 699701.
20. KEENAN, J., V. THANGARAJ, A. HALPERN, et al. 2001. Absolute pitch and planum tem-
porale. NeuroImage 14: 14021408.
21. WITELSON, S. & W. PALLIE. 1973. LeIt hemisphere specialization Ior language in the
newborn: neuroanatomical evidence oI asymmetry. Brain 96: 641646.
22. JUSLIN, P.N. & J.A. SLOBODA. 2001. Music and Emotion: Theory and Research. OxIord
University Press. OxIord, UK.
23. BLOOD, A.J., R.J. ZATORRE, P. BERMUDEZ, et al. 1999. Emotional responses to pleasant
and unpleasant music correlate with activity in paralimbic brain regions. Nature Neu-
rosci. 2: 382387.
24. HANDEL, S. 1989. Listening. MIT Press. Cambridge, MA.
25. FRISTON, K., C. PRICE, P. FLETCHER, et al. 1996. The trouble with cognitive subtraction.
NeuroImage 4: 97104.
26. LANE, R.D., E.M. REIMAN, M.M. BRADLEY, et al. 1997. Neuroanatomical correlates oI
pleasant and unpleasant emotion. Neuropsychologia 35: 14371444.
27. DIAS, R., T.W. ROBBINS & A.C. ROBERTS. 1996. Dissociation in preIrontal cortex oI
aIIective and attentional shiIts. Nature 380: 6972.
28. HORNAK, J. & E.T. ROLLS. 1996. Face and voice expression identiIication in patients
with emotional and behavioural changes Iollowing ventral Irontal lobe damage.
Neuropsychologia 34: 247261.
29. DAMASIO, A.R. 1996. The somatic marker hypothesis and the possible Iunctions oI the
preIrontal cortex. Phil. Trans. Roy. Soc. Lond. 351: 14131420.
30. PANKSEPP, J. 1995. The emotional sources oI 'chills induced by music. Mus. Percept.
13: 171207.
31. KRUMHANSL, C.L. 1997. An exploratory study oI musical emotions and psychophysiol-
ogy. Canad. J. Exp. Psychol. 51: 336353.
32. GOLDSTEIN, A. 1980. Thrills in response to music and other stimuli. Physiol. Psychol.
8: 126129.
33. BLOOD, A.J. & R.J. ZATORRE. 2001. Intensely pleasurable responses to music correlate
with activity in brain regions implicated in reward and emotion. Proc. Natl. Acad.
Sci. USA 98: 1181811823.
34. ZALD, D.H. & J.V. PARDO. 1997. Emotion, olIaction, and the human amygdala:
amygdala activation during aversive olIactory stimulation. Proc. Natl. Acad. Sci.
USA 94: 41194124.
35. MORRIS, J.S., C.D. FRITH, D.I. PERRETT, et al. 1996. A diIIerential neural response in
the human amygdala to IearIul and happy Iacial expressions. Nature 383: 812.
36. BARDO, M.T. 1998. Neuropharmacological mechanisms oI drug reward: beyond
dopamine in the nucleus accumbens. Crit. Rev. Neurobiol. 12: 3767.
37. KESTER, D.B., A.J. SAYKIN, M.R. SPERLING, et al. 1991. Acute eIIect oI anterior tempo-
ral lobectomy on musical processing. Neuropsychologia 29: 703708.
38. BERRIDGE, K.C. & T.E. ROBINSON. 1998. What is the role oI dopamine in reward:
hedonic impact, reward learning, or incentive salience? Brain Res. Rev. 28: 309369.
39. SCHILSTROM, B., H.M. SVENSSON, T.H. SVENSSON, et al. 1998. Nicotine and Iood
induced dopamine release in the nucleus accumbens oI the rat: putative role oI
alpha7 nicotinic receptors in the ventral tegmental area. Neuroscience 85: 1005
1009.
40. PFAUS, J.G., G. DAMSMA, D. WENKSTERN, et al. 1995. Sexual activity increases dopa-
mine transmission in the nucleus accumbens and striatum oI Iemale rats. Brain Res.
693: 2130.
41. SMALL, D., R.J. ZATORRE, A. DAGHER, et al. 2001. Changes in brain activity related to
eating chocolate: Irom pleasure to aversion. Brain 124: 17201733.
42. ZATORRE, R.J. & I. PERETZ. 2001. The Biological Foundations oI Music. Ann. N.Y.
Acad. Sci. 930: 1462.
43. PERETZ, I., A.J. BLOOD, V.B. PENHUNE, et al. 2001. Cortical deaIness to dissonance.
Brain 124: 928940.