Music and Nonmusical Abilities
E. GLENN SCHELLENBERG
Department of Psychology, University of Toronto at Mississauga,
Mississauga, Ontario, Canada L5L 1C6
ABSTRACT: Reports that exposure to music causes benefits in nonmusical do-
mains have received widespread attention in the mainstream media. Such re-
ports have also influenced public policy. The so-called “Mozart effect” actually
refers to two relatively distinct phenomena. One concerns short-term increases
in spatial abilities that are said to occur from listening to music composed by
Mozart. The other refers to the possibility that formal training in music yields
nonmusical benefits. A review of the relevant findings indicates that the short-
term effect is small and unreliable. Moreover, when it is evident, it can be ex-
plained by between-condition differences in the listener’s mood or levels of cog-
nitive arousal. By contrast, the effect of music lessons on nonmusical aspects of
cognitive development is still an open question. Several studies have reported
positive associations between formal music lessons and abilities in nonmusical
(e.g., linguistic, mathematical, and spatial) domains. Nonetheless, compelling
evidence for a causal link remains elusive.
KEYWORDS: Musical ability; Music exposure; Mozart effect; Spatial ability;
Nonmusical benefits of music training
The present report evaluates claims that exposure to music produces benefits in
nonmusical domains. These claims began to influence public policy as soon as they
came to public notice.1 For example, Zell Miller, the former Governor of Georgia,
budgeted for the distribution of classical music recordings to each infant born in
state. Moreover, Florida mandates daily doses of classical music in state-run pre-
Researchers2 and journalists3–6 have generated confusion by failing to clarify the
distinction between the short-term consequences of music listening and the long-
term consequences of formal training in music. Indeed, results from both types of
studies have been merged to yield the dictum, “music makes you smarter.” However,
passive listening to music, a ubiquitous activity, bears little resemblance to formal
training, which involves lessons and systematic practice. Thus, separate evaluation
of the short-term benefits of musical exposure and the long-term side effects of mu-
sic lessons could help to clarify these issues.
Examination of the effects of previous experience on learning and behavior has a
rich tradition in the history of psychology. “The transfer of training from old to new
Address for correspondence: E. Glenn Schellenberg, Ph.D., Department of Psychology, Uni-
versity of Toronto at Mississauga, Mississauga, Ontario, Canada L5L 1C6. Voice: 905-828-5367;
356ANNALS NEW YORK ACADEMY OF SCIENCES
situations is part and parcel of most, if not all, learning. In this sense the study of
transfer is coextensive with the investigation of learning”7 (p. 1019; emphasis add-
ed). In addition, hundreds of psychological examinations of priming have investigat-
ed how prior exposure to a stimulus affects subsequent processing of the same
stimulus or a closely related stimulus.8 If exposure to music causes welcome side ef-
fects, we would expect such effects to arise from transfer or priming. Moreover,
some researchers who argue for the nonmusical benefits of exposure posit a specific
neuropsychological basis for such benefits.2,10–12 Presumably, this hypothesized
cortical process would be compatible with other cortical processes that are demon-
strably relevant to music.
TRANSFER AND PRIMING
Transfer and priming occur in positive and negative forms. Positive transfer oc-
curs when previous experience in problem solving makes it easier to solve a new
problem,7,13 typically by accelerating learning. As such, positive transfer describes
successful generalization of a process or strategy. One example involves reasoning
with analogies.14 Previous exposure to analogies can lead to greater success at find-
ing the missing piece in new analogies (e.g., Lawyer is to client as doctor is to ????;
the correct answer is patient). Similar research is available on metaphor and the
transfer of skills.13 A common theme across transfer effects is similarity;13,15 posi-
tive transfer is more likely to occur when there are more similarities between the old
and new problems.
Negative transfer is the opposite of its positive counterpart; previous experience
interferes with solving a new problem.7 Negative transfer, which is often called in-
terference, can occur proactively or retroactively. Proactive interference is evident
when previous learning makes subsequent learning relatively difficult. For example,
a new problem is approached with an old mental set that is inefficient or inappropri-
ate for the new context. By contrast, retroactive interference refers to difficulty ac-
cessing mental representations because of intervening experience between initial
encoding and retrieval.
Roughly speaking, priming can be considered the “short-term” or “low-level” rel-
ative of transfer. Anderson16 defines priming as “an enhancement of the processing
of a stimulus as a function of prior exposure” (p. 459). In a classic experiment,17 par-
ticipants were asked to identify words presented briefly in the visual modality. Per-
formance was superior for words that were seen prior to the word-identification test.
The low-level nature of priming is evident in greater priming effects following open-
ended instructions (e.g., study the word) compared to compulsory semantic process-
ing (e.g., generate an antonym), the latter condition involving “deeper” levels of pro-
cessing.18 It is clear that priming does not require conscious awareness, as reflected
in the priming effects observed in amnesics.19
Negative priming refers to situations in which the processing of a “target” stimu-
lus is inhibited by prior exposure.20 For example, when participants are presented
with two words (a “target” and a “distractor”) and required to name only one (the
target), performance on subsequent trials is relatively slow when the target word was
previously a distractor. Most priming studies examine repetition priming, or subse-
quent processing of an identical stimulus.8 Nonetheless, cross-modal and cross-lan-
357SCHELLENBERG: MUSICAL ABILITY
guage priming effects are also observable. For people who are bilingual in Spanish–
English, auditory presentation of a partial sentence in Spanish (the priming stimulus)
can facilitate visual recognition of a target English word, provided that the target was
implied by the sentential prime.21 There are higher-level, associative16 or semantic9
priming effects, in which word processing (e.g., butter) is facilitated by previous pre-
sentation of an associated word (e.g., bread).22
This brief review of transfer and priming provides a context for evaluating spe-
cific claims about exposure to music. These claims posit remarkable positive side ef-
fects of exposure to certain types of music, side effects that, in principle, are closely
related to transfer and priming. My intention is to situate claims that music makes
you smarter in the context of cognitive psychology, which will permit a review and
evaluation of such claims with reference to well-established cognitive phenomena.
A secondary goal is to situate the neuronal mechanism advanced as the basis of as-
sociations between music and nonmusical abilities in the domain of cognitive
THE MOZART EFFECT
The current debate about musical exposure and its side effects was inspired, in
part, by Rauscher et al.23 who reported that brief exposure (10 minutes) to a Mozart
sonata generates short-term increases in spatial-reasoning abilities (the Mozart ef-
fect). Each participant in their study was tested in three conditions. Participants in
one condition listened to a Mozart sonata before completing one of three tests of spa-
tial abilities. Participants in the other two conditions listened to a relaxation tape or
sat in silence before completing a test. Performance was superior in the “Mozart”
condition. This finding attracted considerable attention, because it appeared in a
highly prestigious journal, Nature, and because the investigators translated their
finding into an IQ-score improvement of approximately 8 points (i.e., half a standard
deviation). Indeed, the popular conclusion that “music makes you smarter” followed
directly from this IQ translation.
Closer examination of the method of that study raises questions about the validity
of the findings. The choice of comparison conditions is particularly problematic. Sit-
ting in silence or listening to a relaxation tape for 10 minutes is less arousing or in-
teresting than is listening to Mozart. Moreover, mood states are known to influence
performance on problem-solving tasks, with superior performance associated with
positive affect.24–26 Thus, the effect could have arisen from differences in mood or
arousal rather than from exposure to Mozart.
Because the Mozart effect is at odds with the literature on priming and transfer,
alternative explanations of the source of the effect (i.e., the mood/arousal hypothe-
sis) seem all the more credible. Improved spatial skills following exposure to a
Mozart sonata do not represent an instance of repetition priming (i.e., the priming
stimulus was not repeated) nor are they an instance of associative priming. How is
passive listening to a musical stimulus “associated” with performance on a visually
presented test of spatial skills? Evidence for associative priming typically involves
pairs of words with an obvious semantic association (nurse–doctor, bread–butter).
How, then, could an auditory (musical) stimulus prime performance on a task with
no obvious link to music? Transfer as an explanatory framework also raises more
358ANNALS NEW YORK ACADEMY OF SCIENCES
questions than it answers. Transfer typically involves applying a learned skill or a
strategy to a new context. But what is learned by listening passively to a piece of mu-
sic? Something about the music, no doubt, is learned, but it is difficult to rationalize
how the transfer of such knowledge could yield improved performance on a spatial
In short, the Mozart effect is a radical claim about cognitive processes that is dif-
ficult to reconcile with known principles and findings in cognitive psychology. It
comes as no surprise, then, that attempted replications have produced mixed re-
sults.10,27 As of May 2000, there were about 20 published tests of the Mozart effect.
Less than half replicated the effect. Because the Mozart-effect studies have been re-
viewed elsewhere, the present report focuses on the issues raised by selected studies.
Consider the replication reported by Rauscher et al.,12 who “pretested” their par-
ticipants with a paper-folding-and-cutting (PF&C) test (one of the spatial tests used
in the original study). Participants were then divided into three groups of equivalent
abilities. One group heard Mozart during three subsequent test sessions. A second
group sat in silence during the three sessions. A third group heard a minimalist piece
by Philip Glass during the first session, an audiotaped story in the second session,
and a repetitive piece of dance music in the third session. After each session, the
PF&C test was administered again. Although the Mozart group showed a signifi-
cantly larger improvement in performance than did the other two groups after the
first session, there was no difference between the Mozart and comparison groups af-
ter the next two sessions. The advantage of Mozart over silence and Glass conditions
in the first test session did not extend or clarify the original finding. Participants may
find repetitive, minimalist music as boring or unarousing as silence. The null find-
ings in the second and third sessions also raise doubts about the reliability of the
Rauscher and Shaw10 suggest that the numerous replication failures can be ex-
plained primarily by differences in the spatial tasks that have been used as outcome
measures. They claim that the effect can be obtained with “spatial-temporal” tasks
(e.g., the PF&C task and other tasks involving mental transformation of visual im-
ages), but not with “spatial-recognition” tasks. This distinction is based on the idea
that perceiving and remembering music involve identifying changes and systematic
transformations in musical patterns (e.g., motives) that occur over time. Thus, trans-
fer from music listening to the spatial domain should be limited to tasks involving
mental manipulation of visual images, which also takes time. Indeed, the time re-
quired is linearly related to the amount of manipulation.28 This distinction is curious
in light of the original findings,23 which indicated that the effect was identical across
spatial tasks, temporal or otherwise. In a subsequent reanalysis of the original data,10
however, the advantage of the Mozart effect proved to be significant on only one of
the three spatial tests that were administered, the “temporal” PF&C task, and not on
the two nontemporal tests. Nonetheless, mean scores were highest in the Mozart
condition across tests, and the design precluded tests of the two-way interaction be-
tween the listening conditions and the spatial tests. In other words, despite their con-
clusion and interpretation, the data did not support Rauscher and Shaw’s hypothesis,
that is, that the influence of Mozart’s music on spatial abilities depends on the tem-
poral nature of the tasks. Moreover, the temporal/nontemporal distinction cannot ex-
plain why several attempts to replicate the original findings failed to do so, even
359 SCHELLENBERG: MUSICAL ABILITY
though the outcome measure was a task that met Rauscher and Shaw’s criteria for
spatial-temporal status.29,30 Finally, the distinction does not address the problem
that the effect, when evident, may be a consequence of differences in mood or in
In all cases in which the Mozart effect has been evident, comparison conditions
involved repetitive music, sitting in silence, or listening to relaxation tapes. As not-
ed, these comparison conditions might seem boring to participants (compared to lis-
tening to music), promoting relatively negative mood states or low levels of
cognitive arousal. As a first attempt to address this possibility, Nantais and
Schellenberg31 replicated and extended the original findings. In their first experi-
ment, each participant was tested on the PF&C task twice, once after listening to 10
minutes of music and once after sitting in silence for 10 minutes. For some partici-
pants, the music was the same Mozart piece used by Rauscher and her colleagues.
For others, a piece by Schubert (from the same compact disk performed by the same
pianists) was used instead. This experiment was also the first to use a computer-
controlled procedure administered to participants individually. Indeed, the potential
impact of group dynamics on the results of earlier studies is unknown.12,23,29,32,33
(Imagine a classroom of undergraduates being required to sit in silence for 10
As shown in FIGURE 1 (upper panel), performance on the PF&C test was better
after listening to Mozart than sitting in silence. In other words, the Mozart effect was
replicated. Nonetheless, an identical effect was evident when the Mozart composi-
tion was substituted with the piece by Schubert (FIG. 1, middle panel). One would
predict such a “Schubert effect” if the comparison condition (silence) was depress-
ing levels of performance. For both groups, performance also improved from the first
to the second testing session, revealing a simple practice effect. In a second experi-
ment, a Mozart condition was contrasted with a comparison condition that involved
listening to a narrated short story (potentially as engaging as listening to music) in-
stead of sitting in silence. The Mozart effect disappeared (FIG. 1, lower panel), as one
would predict if the experimental (Mozart) and comparison (story) conditions were
equally engaging and if the source of the Mozart effect stemmed from differences in
mood or arousal. Perhaps even more important was the finding that performance in-
teracted with listeners’ preferences. Those who preferred Mozart over the story per-
formed better on the PF&C test after listening to Mozart. Those who preferred the
story performed better after listening to the story (FIG. 2). These findings provide
support for the suggestion that short-term effects of music on tests of spatial abilities
stem from differences in mood or arousal rather than from listening to Mozart. Al-
though the FIGURE implies that participants who preferred Mozart performed better
regardless of condition, the main effect of preference was marginal (p = 0.09).
Further support for the “mood or arousal” hypothesis comes from a metanalysis
that provided an overview of 20 Mozart–silence comparisons conducted to date.27
The overall advantage of listening to Mozart on subsequent tests of spatial skills was
nonsignificant (i.e., equivalent to 1.4 IQ points). The effect size for studies that used
a spatial-temporal outcome measure was also nonsignificant (2.1 IQ points) and
smaller than the average test-retest variability in IQ for a single person. Moreover,
successful replications of the Mozart effect were attributed to cognitive arousal,
which is predominantly a right hemisphere function,34–36 as are tests of complex
360ANNALS NEW YORK ACADEMY OF SCIENCES
FIGURE 1. Scores on the paper-folding-and-cutting task for participants tested by
Nantais and Schellenberg (1999). Each participant was tested twice. The upper panel illus-
trates scores after listening to Mozart or sitting in silence. The middle panel illustrates
scores after listening to Schubert or sitting in silence. The lower panel illustrates scores after
listening to Mozart or a narrated story. The line on the diagonal represents equivalent per-
formance across conditions. The maximum score was 17.
361 SCHELLENBERG: MUSICAL ABILITY
spatial abilities.36,37 This view helps to explain why the Mozart effect tends to be
slightly larger when the control condition consists of relaxation instructions, which
are designed to reduce arousal, instead of sitting in silence.27
Another way to interpret the Mozart effect is provided by a new theory based on
a large body of findings on the association between mood and cognition.38 The the-
ory proposes that positive mood states increase circulating levels of the neurotrans-
mitter dopamine. During periods of positive affect, dopamine is released from the
ventral tegmental area, which has projections to the prefrontal cortex. A variety of
cognitive tasks that show improvement when positive affect is induced25 may be in-
fluenced by the effects of dopamine on prefrontal function. It is possible, then, that
the Mozart effect is another way in which positive affect influences performance in
a problem-solving task. In short, although these seemingly mysterious effects of
cross-modal priming (i.e., the Mozart effect) may indeed have a neuropsychological
explanation, listening to music is just one of many ways to induce arousal or positive
The metanalysis presented by Chabris27 and the results of Nantais and
Schellenberg30 are consistent with the idea that differences in mood or arousal are
the actual source of the Mozart effect, but neither report tested this hypothesis direct-
ly. Thompson et al.39 attempted such a test using the PF&C task as their outcome
measure. Each of their participants was tested once in a music condition and once in
a silence condition (as in Nantais and Schellenberg,30 Experiment 1). Participants’
mood after listening to the music was measured using the Profile of Mood States.40
For some participants, the music condition consisted of the same Mozart piece used
in the original Mozart-effect study; for others, a piece by Albinoni was used instead.
The Albinoni Adagio was selected, because it is considered to be a stereotypical ex-
ample of slow, sad-sounding music.41 By contrast, the Mozart sonata is pleasant and
happy sounding. Hence, the prediction was that increases in performance on the
PF&C task would be evident for music compared to silence in the Mozart group but
not in the Albinoni group. This prediction was upheld by the data. Moreover, the per-
FIGURE 2. Scores from Nantais and Schellenberg’s (1999) participants as a function
of testing condition (Mozart or story) and preference (Mozart or story).
362ANNALS NEW YORK ACADEMY OF SCIENCES
formance advantage of the music over the silence condition in the Mozart group dis-
appeared when differences in participants’ mood scores were partialled out of the
analysis. In short, the results were completely consistent with the notion that the
Mozart effect is an epiphenomenon of mood or arousal.
The theoretical framework that Rauscher and Shaw use to explain the Mozart ef-
fect is called the “Trion model.”42,43 The model states that specific cortical firing
patterns are present over large areas of the cortex when one performs, composes, or
listens to music. Because these patterns are considered to be spatial-temporal in na-
ture, they are said to be highly similar to patterns evident during spatial-temporal
reasoning. Both processes involve perceiving and thinking about rule-governed
transformations that occur over time. The model describes more than a simple asso-
ciative or connectionist network, in which one group of neurons is connected to an-
other group. Rather, it posits actual similarities in cortical firing patterns for (1)
passive listening to music and (2) actively participating in a task requiring spatial-
If we examine the neuropsychological research on music processing, however,
the basic tenets of the Trion model seem implausible. The research of Peretz and her
colleagues is particularly relevant. Peretz has shown that much of music perception
and cognition is relatively modular, and, moreover, that individual aspects of music
cognition are relatively modularized and independent of other aspects.44 For exam-
ple, melody and rhythm are processed independently and in different parts of the
brain,45–47 lyrics are processed independently of tunes,48 and perceiving musical
emotion is independent of memory for music.49,50 Most importantly, Peretz has
studied brain-damaged patients with amusia, and none has exhibited accompanying
deficits in spatial abilities. For example, one of her amusic patients could not dis-
criminate tones that differed by gross differences in pitch, yet she continued to drive
safely around Montreal. In short, there is substantial evidence for modularity of mu-
sic processing and for independence of various aspects of music. Such evidence is
inconsistent with the notion that cortical activity is similar across a variety of musi-
cal activities (performing, composing, and listening) and that such patterns of acti-
vation are identical during spatial-temporal reasoning.
LONG-TERM SIDE EFFECTS OF MUSIC LESSONS
Although the short-term Mozart effect appears to be independent of Mozart in
particular and of music in general, it is still possible that positive, relatively long-
term cognitive side effects result from taking music lessons. Indeed, the two issues
may be orthogonal. To anticipate the conclusion, the relevant findings reviewed be-
low are consistent with the idea of an association between musical training and non-
musical benefits, but they fall far short of being conclusive. The studies are grouped
according to design (correlational, quasiexperimental, or experimental).
Several studies have examined whether musical ability (rather than musical train-
ing) is correlated with other kinds of abilities. Positive associations imply that im-
proving one’s musical ability through formal lessons would be accompanied by
363 SCHELLENBERG: MUSICAL ABILITY
nonmusical benefits. In correlational designs, however, it is always impossible to
make firm conclusions about the direction of causation when associations are dis-
covered. It is also impossible to rule out the possibility that the association stems
from a third, unidentified variable.
Gromko and Poorman51 examined children between the ages of 4 and 13 who
were enrolled in a private school. Their goal was to determine whether musical ap-
titude is related to children’s ability to use symbols. In an initial testing session, chil-
dren completed the tonal subtest of Gordon’s52,53 musical aptitude measures. During
a second session, children were tested on two tasks, one that required them to match
short melodies with graphic representations and another that required them to draw
graphic representations of the contour of short melodies. Performance on all three
measures improved with age, and each measure was significantly correlated with the
other two. These findings confirm that children’s musical aptitude is predictive of
their ability to interpret and produce symbolic representations of music. Because
each of the outcomes was associated with age, however, it is impossible to determine
whether the associations would still be in evidence if differences in age were held
constant (i.e., the authors did not report partial correlations).
In an examination of performance on musical and spatial tasks that required an-
alogical reasoning, children from 6 to 12 years of age were tested on their ability to
transfer a given relation between one pair of stimuli to a novel pair.54 As the age of
the children increased, performance on both tasks improved. Moreover, age-related
improvements were virtually identical across tasks. As with the study by Gromko
and Poorman,51 however, the association between the music and spatial tasks could
be a consequence of the fact that older children performed better on both tasks.
Lamb and Gregory55 studied the association between reading and musical abili-
ties in a sample of 5-year-old children. Reading abilities and phonemic awareness
were positively associated with pitch-discrimination abilities but not with the ability
to discriminate timbres. These associations remained in evidence when differences
in age and nonverbal intelligence were held constant. Virtually identical associations
between reading abilities and musical abilities (with differences in age and IQ held
constant) were reported for a sample of 9-year-old children.56 Although these find-
ings do not address the issue of causation, they provide evidence of an association
between reading and musical abilities that is independent of age or general
Douglas and Willatts57 tested a sample of 8 year olds to examine whether literacy
and musical ability are associated. Pairs of tones were presented in a pitch-discrim-
ination task that required children to identify whether the second tone was higher,
lower, or the same as the first. A rhythm-discrimination test required children to re-
spond “same” or “different” to pairs of sequences played on a wood block. Literacy
was measured with tests of reading and spelling. All measures showed significant
pairwise correlations. When differences in receptive vocabulary were held constant,
however, reading and spelling measures were associated with rhythm-discrimination
abilities but not with pitch-discrimination abilities. Whereas these findings suggest
that rhythm-discrimination abilities are better than pitch-discrimination abilities at
predicting literacy, the results of Lamb and Gregory55 imply that pitch-discrimina-
tion abilities are a better predictor than timbre-discrimination abilities.
Finally, Lynn et al.58 examined the association between musical aptitude and
general intelligence (Spearman’s g) in groups of children 10 years of age. Children
364ANNALS NEW YORK ACADEMY OF SCIENCES
were administered rhythm- and pitch-discrimination tasks as well as tests of general
intelligence. Each of the music measures was positively associated with each of the
measures of intelligence. These results suggest that musical aptitude is a function of
general intelligence. Alternatively, musical aptitude may be a valid estimate of g. Al-
though the association between musical aptitude and intelligence is provocative, it
remains to be seen whether music lessons actually promote improvements in cogni-
Other researchers have tested for the possibility of differences between naturally
occurring groups (e.g., those with and without musical training) in nonmusical abil-
ities. Again, because we can never be sure that the groups are identical on other po-
tentially relevant dimensions (e.g., socioeconomic status and overall IQ),
unequivocal determinations of causation are impossible.
A classic example of a relevant quasiexperiment is a study by Chan et al.59 of fe-
male college students in Hong Kong (mean age of 20 years). The authors compared
the verbal and visual memory abilities of women with no musical training to those
of women who had taken six years of music lessons before the age of 12. Although
the groups did not differ on the visual-memory task, the musically trained group out-
performed the untrained group on the verbal-memory task. Unfortunately, despite
the authors’ claim that the groups were matched according to years of education
(with alpha = 0.01), closer inspection of the findings revealed that the musically
trained group had significantly more education (with alpha set to a standard 0.05 val-
ue). In other words, it is impossible to determine whether the verbal advantage
stemmed from music lessons rather than from additional years of education. Indeed,
we would predict that better verbal skills would accompany higher levels of
Hassler et al.60 examined verbal fluency and visual-spatial abilities in children 9
to 14 years of age. The children were classified into one of three groups: (1) musi-
cally talented and capable of composing or improvising, (2) musically talented but
not capable of composing or improvising, or (3) nonmusicians. The groups did not
differ on a test of spatial relations, but significant differences were found on tests of
verbal fluency and visualization abilities, with the musically talented children out-
performing the nonmusicians. At a follow-up test two years later, significant differ-
ences were found for each of the three outcome variables.61 Nonetheless, students in
the composing/improvising group had more music lessons than did the other musi-
cally talented group, yet no differences between these groups on the outcome mea-
sures were evident. As such, this study provides equivocal support for the idea that
music lessons are accompanied by advantages in nonmusical domains.
Two studies compared the nonmusical abilities of children enrolled in a Kodály
music program with those of a comparison group who were not taking music les-
sons.62,63 The Kodály program is known for intensive training and for placing great
emphasis on singing and on the development of sequential skills. The program also
incorporates clapping, the use of hand signs, and simple musical notation. Hurwitz
and his colleagues examined the sequencing and spatial skills of a group of seven
year olds. Children in the Kodály group had taken music lessons for approximately
seven months, with 40-minute lessons five days per week. The sequencing task in-
365SCHELLENBERG: MUSICAL ABILITY
volved tapping mechanical keys in a regular manner or in time with a metronome af-
ter the metronome was turned off or its rate had been changed. Children were also
given tests of spatial abilities, plus a Stroop-like test of interference. The Kodály
group outperformed the comparison children on the Stroop test and on some of the
spatial tests. In a separate examination of children who had completed 1 year of
Kodály instruction, the Kodály group performed better than a comparison group on
a reading test even though the two groups had performed identically a year earlier.
A subsequent study of four- and five-year-olds’ understanding of prenumber con-
cepts showed a benefit of Kodály training only for five-year-old girls.63 These results
suggest that training in music may lead to nonmusical improvements, yet it is impos-
sible to ascertain whether nonmusical aspects of Kodály training or preexisting dif-
ferences between groups may have influenced the results.
The next group of studies had more-or-less random assignment of participants to
experimental conditions. Thus, provided that comparison conditions were selected
appropriately, we should be able to determine whether music lessons actually
“cause” nonmusical cognitive advantages. As with most of the short-term (Mozart
effect) studies, however, none of the studies in this group used comparison condi-
tions that preclude the possibility of alternative explanations for the findings.
For example, six-year-old children who were taught music for seven months by
means of the Kodály method showed improvements in mathematical and reading
abilities that surpassed those of children without such training.64 The researchers’
goal was to examine possible by-products of a “test arts” (Kodály) program that was
implemented in some first-grade classes but not in others. They examined two first-
grade classes in each of two schools that were designated as “test arts” classrooms
and another two from both schools that were “standard arts” classrooms. If we as-
sume that the classrooms were assigned to the two arts programs at random, we can
consider the design to approximate a “true” experiment. The reported advantage for
the test-arts classes is remarkable when we consider that in the previous year chil-
dren in the test-arts classes were actually behind the standard-arts children in terms
of the proportion who had reached the national average grade level. Although these
results are promising, children in the standard-arts classrooms did not participate in
activities focusing on “sequenced skill development” as did children in the test-arts
(Kodály) classrooms. Again, this confounding makes it impossible to attribute the
remarkable recovery and achievements of the test-arts classrooms to training in mu-
sic per se, rather than to other nonmusical aspects of the Kodály program.
In another study, four-year-old children who received individual 10-minute piano
lessons once or twice a week for six to eight months performed better on a test of
spatial skills than children assigned to comparison conditions.11 Nonetheless, other
aspects of the design question the reliability of the effect. For example, some of the
children had 33% more lessons than other children, yet this additional training in
music had no effect on performance. Moreover, the primary comparison condition
involved playing with commercial software programs on a computer. Although a
computer instructor provided one-on-one instruction about how to use the computer
and open the programs, the software (not the instructor) was designed to teach the
children basic skills in reading and arithmetic. As such, superior levels of perfor-
366ANNALS NEW YORK ACADEMY OF SCIENCES
mance in the piano group could be the consequence of additional instruction from an
Standley and Hughes65 found that children in prekindergarten classes (four to
five years of age) who took 15 music lessons over a period of two months showed
enhanced pre-reading and writing skills compared to other children. Children in the
comparison condition were exposed to the regular prekindergarten curriculum but
had no additional lessons of any kind. Again, it is impossible to determine whether
the observed numerical and verbal benefits arose specifically from music instruction
or from pedagogical differences that were independent of musical training. The in-
vestigators noted that “it was also apparent from the children’s reaction that the mu-
sic activities provided pleasure and excitement about academic participation,
possibly generating long range motivation for reading and writing” (p. 83). Nonmu-
sical activities that generate similar levels of pleasure and excitement could generate
similar increases in motivation.
Gromko and Poorman’s66 study of three- and four-year-old children enrolled in a
private Montessori school is similar to Standley and Hughes’65 study just described.
Children in the music group were provided with weekly group music lessons in ad-
dition to the regular curriculum, but the comparison group received no additional
lessons of any sort. As such, the modest gains in nonverbal IQ witnessed for the mu-
sic group relative to the comparison group can be attributed simply to additional ed-
ucational instruction from an adult.
Three recent experimental studies suffer from similar methodological problems.
Each compared young children enrolled in music-education programs with children
in “control” groups who had no comparable extraschool activities.67–69 One study
provided three years of piano lessons free of charge to children in the fourth to sixth
grades.68 These “piano” children performed better than children in a control group
on a comprehensive test of cognitive abilities after the first and second years, but the
difference disappeared after the third year. Between-group differences during the
first two years stemmed solely from differences in spatial abilities. In another study,
kindergartners were provided with group keyboard lessons for eight months.69 The
keyboard children showed greater improvement than a control group on tests of spa-
tial abilities, but there was no difference between groups on a test of recognition. A
third study examined the influence of a 30-week structured music curriculum on
cognitive development.67 Treatment and control groups of six year olds were admin-
istered six subtests from the Stanford-Binet Intelligence Test before and after the
curriculum. The treatment group showed relatively larger gains on a single subtest
that measured capacity of short-term memory (Bead Memory).
Another recent study examined possible side effects of group keyboard lessons
that were provided free of charge to children six to eight years of age.70 A control
group had computer lessons with a commercial software program designed to im-
prove English language skills. Both groups were also given lessons intended to en-
hance spatial abilities by playing with a software program designed by the
researchers. Unfortunately, the main outcome variable consisted of scores on a “test-
ing” version of the same spatial software, which has unknown reliability and valid-
ity. Moreover, aggregate scores on the outcome tests did not differ between groups.
The investigators reported a significant advantage for the keyboard group on a sub-
test of mathematical fractions and proportions, and they concluded that improved
musical and spatial skills lead to improved mathematical abilities. These results
367 SCHELLENBERG: MUSICAL ABILITY
would be more convincing if they had been obtained with standardized tests and if
the piano group had performed better overall or at least on subtests for which clear
predictions were made a priori.
The studies just reviewed provide consistent suggestive evidence that music les-
sons have positive nonmusical side effects. Nonetheless, specifics of the reported as-
sociations vary widely from study to study. If we suspend our disbelief, however, and
assume that music education affects abilities in other areas, how could we account
for this influence?
A number of neurological studies describe ways in which music lessons affect
cortical development. Compared to nonmusicians, accomplished players of string
instruments show increased representation in the cerebral cortex for the fingers of
their left hand,71 which implies that musical training can alter patterns of cortical or-
ganization. Indeed, cortical representations are especially large for those who begin
music lessons at an early age when the brain is relatively plastic. Although the size
of the corpus callosum is larger in musicians than in nonmusicians, this effect is par-
ticularly notable in musicians who began taking lessons before the age of 7.72 Rela-
tively large brain asymmetries are also evident among musicians who have absolute
(perfect) pitch,73 and this relatively rare ability to name and produce pitches in iso-
lation is evident predominantly among musicians who begin lessons in early child-
hood.74 Moreover, the representation of piano tones in the auditory cortex differs in
musicians than in nonmusicians,75 although genetic factors or simple exposure to
music could also play a role.76 Finally, specific cortical areas in the right hemisphere
are activated when reading a musical score but not when reading one’s primary or
Consequences of an enriched environment on other species (e.g., rats and mice)
include denser patterns of dendritic branching and a greater number of hippocampal
neurons.78,79 If music education represents an enrichment of a child’s environment,
such enrichment could promote neurological development, which could, in turn, in-
fluence abilities in other domains. Music, however, is simply one of many ways to
enrich a child’s environment. Moreover, music education is a complex process that
involves many different dimensions. As such, it may be more fruitful to examine the
effects of music education at a behavioral level instead of attempting to map such
effects directly onto cortical architecture.
We know that schooling improves a wide variety of cognitive skills and that this
association is not simply a by-product of maturation.80–83 For young children in par-
ticular, schooling is more effective in smaller classes.84,85 Reviews of intervention
programs for children who are at risk of academic failure suggest that extended one-
on-one contact with a supportive adult is a common feature of successful interven-
tions.86–87 Thus, music lessons, which are typically taught individually or in small
groups, may confer nonmusical benefits for children by providing close and extend-
ed contact with an adult other than a parent or teacher. If this is the case, then similar
side effects should be evident with other types of lessons that provide similar levels
of contact (e.g., chess, drawing).
Music lessons may be unique, however, because of their focus on a particular
combination of factors, such as hours of individual practice, learning to read music,
attention and concentration, timing, ear training, sight reading, constructive feed-
back from the instructor, and exposure to music.88 Thus, positive transfer effects to
nonmusical domains, such as language, mathematics, or spatial reasoning, could be
368 ANNALS NEW YORK ACADEMY OF SCIENCES
similarly unique for individuals who take music lessons. On the other hand, music
lessons are likely to improve many general skills, such as attending to rapidly chang-
ing temporal information, honing skills of auditory stream segregation, developing
the ability to detect temporal groups, becoming attentive to signals of closure and
other gestalt cues of form, developing emotional sensitivity and expressiveness, and
developing fine motor skills. These general skills should be particularly likely to
transfer to a variety of nonmusical domains.
As someone who took music lessons from the age of five and practiced regularly
for the next 11 years, I feel changed—probably for the better—in ways that seem
specific to my involvement with music. It remains to be seen, however, whether this
personal observation will withstand the test of rigorous experimental investigation.
Preparation of this article was supported by a grant from the International Foun-
dation for Music Research. Bill Thompson and Sandra Trehub provided helpful
comments on an earlier draft.
1. KOLATA, G. 1999. Muddling fact and fiction and policy. New York Times, WK5.
2. SHAW, G.L. 2000. Keeping Mozart in Mind. Academic Press. San Diego, CA.
3. CALAMAI, P. 1999. Mozart theory creates discord in the scientific community. The
Toronto Star, A28, Aug. 26.
4. INGRAM, J. 1997. More proof music lifts young IQs. The Toronto Star, F8, May 11.
5. MARTIN, S. 1994. Music lessons enhance spatial reasoning skills. APA Monitor,
6. MOTCHUK, A. 1997. Can Mozart make maths add up? New Scientist, 17, March 15.
7. POSTMAN, L. 1971. Transfer, interference and forgetting. In Woodworth & Schlos-
berg’s Experimental Psychology, 3rd Edit. J.W. Kling & L.A. Riggs, Eds.: 1019–
1132. Holt, Rinehart & Winston. New York.
8. TULVING, E. & D.L. SCHACHTER. 1990. Priming and human memory systems. Science
9. NEELY, J.H. 1991. Semantic priming effects in visual word recognition: a selective
review of current findings and theories. In Basic Processes in Reading: Visual Word
Recognition. D. Besner & G.W. Humphreys, Eds.: 264–336. Erlbaum. Hillsdale, NJ.
10. RAUSCHER, F.H. & G.L. SHAW. 1998. Key components of the Mozart effect. Percept.
Motor Skills 86: 835–841.
11. RAUSCHER, F.H., G.L. SHAW, L.J. LEVINE, et al. 1997. Music training causes long-term
enhancement of preschool children’s spatial-temporal reasoning. Neurol. Res. 19: 2–8.
12. RAUSCHER, F.H., G.L. SHAW & K.N. KY. 1995. Listening to Mozart enhances spatial
temporal reasoning: towards a neurophysiological basis. Neurosci. Lett. 185: 44–47.
13. ROBINS, A. 1996. Transfer in cognition. Connec. Sci. 8: 185–203.
14. GICK, M.L. & K.J. HOLYOAK. 1983. Schema induction and analogical transfer. Cognit.
15. GENTNER, D., M.J. RATTERMAN & K.D. FORBUS. 1993. The roles of similarity in trans-
fer: separating retrievability from inferential soundness. Cognit. Psychol. 25: 524–
16. ANDERSON, J.R. 1995. Cognitive Psychology and Its Implications, 4th Edit. Freeman.
17. JACOBY, L.L. 1983. Perceptual enhancement: persistent effects of experience. J. Exp.
Psychol. Learn. Mem. Cognit. 9: 21–38.
369 SCHELLENBERG: MUSICAL ABILITY
18. CRAIK, F.I.M. & R.S. LOCKHART. 1972. Levels of processing: a framework for memory
research. J. Verb. Learn. Verb. Behav. 11: 671–684.
19. GRAF, P., L.R. SQUIRE & G. MANDLER. 1984. The information that amnesic patients do
not forget. J. Exp. Psychol. Learn. Mem. Cognit. 10: 164–178.
20. MAY, C.P., M.J. KANE & L. HASHER. 1995. Determinants of negative priming. Psychol.
Bull. 118: 35–54.
21. HERNANDEZ, A.E., E.A. BATES & L.X. AVILA. 1996. Processing across the language
boundary: a cross-modal priming study of Spanish-English bilinguals. J. Exp. Psy-
chol. Learn. Mem. Cognit. 22: 846–864.
22. MEYER, D.E. & R.W. SCHVANEVELDT. 1971. Facilitation in recognizing pairs of words:
evidence of a dependence between retrieval operations. J. Exp. Psychol. 90: 227–234.
23. RAUSCHER, F.H., G.L. SHAW & K.N. KY. 1993. Music and spatial task performance.
Nature 365: 611.
24. BENEDICT, R.H.B., M. DOBRASKI & M.Z. GOLDSTEIN. 1999. A preliminary study of the
association between changes in mood and cognition in a mixed geriatric psychiatry
sample. J. Gerontol. B Psychol. Sci. Soc. Sci. 54B: 94–99.
25. ISEN, A.M. 1999. Positive affect. In The Handbook of Cognition and Emotion. T. Dal-
gleish & M. Power, Eds.: 521–539. Wiley. New York.
26. O'HANLON, J.F. 1981. Boredom: practical consequences and a theory. Acta Psychol.
27. CHABRIS, C.F. 1999. Prelude or requiem for the ‘Mozart effect’? Nature 400: 826–827.
28. SHEPARD, R. & J. METZLER. 1971. Mental rotation of three-dimensional objects.
Science 171: 701–703.
29. STEELE, K.M., K.E. BASS & M.D. CROOK. 1999. The mystery of the Mozart effect: fail-
ure to replicate. Psycholog. Sci. 10: 366–369.
30. STEELE, K.M., K.M. BROWN & J.A. STOECKER. 1999. Failure to confirm the Rauscher
and Shaw description of recovery of the Mozart effect. Percept. Motor Skills 88:
31. NANTAIS, K.M. & E.G. SCHELLENBERG. 1999. The Mozart effect: an artifact of prefer-
ence. Psycholog. Sci. 10: 370–373.
32. STEELE, K.M., T.N. BALL & R. RUNK. 1997. Listening to Mozart does not enhance
backwards digit span performance. Percept. Motor Skills 84: 1179–1184.
33. STEELE, K.M., S. DALLA BELLA, I. PERETZ, et al. 1999. Prelude or requiem for the
‘Mozart effect’? Nature 400: 827.
34. HELLER, W. & J.B. NITSCHKE. 1997. Regional brain activity in emotion: a framework
for understanding cognition in depression. Cognit. Emotion 11: 637–661.
35. ROBERTSON, I.H., J.B. MATTINGLEY, C. RORDEN & J. DRIVER. 1998. Phasic alerting of
neglect patients overcomes their spatial deficit in visual awareness. Nature 395: 169–
36. TUCKER, D.M., A. HARTRY-SPEISER, L. MCDOUGAL, et al. 1999. Mood and spatial
memory: emotion and right hemisphere contribution to spatial cognition. Biolog.
Psychol. 50: 103–125.
37. DITUNNO, P.L. & V.A. MANN. 1990. Right hemisphere specialization for mental rota-
tion in normals and brain damaged subjects. Cortex 26: 177–188.
38. ASHBY, F.G., A.M. ISEN & A.U. TURKEN. 1999. A neuropsychological theory of posi-
tive affect and its influence on cognition. Psychol. Rev. 106: 529–550.
39. THOMPSON, W.F., E.G. SCHELLENBERG & G. HUSAIN. 2001. Arousal, mood, and the
Mozart effect. Psycholog. Sci. 12: 248–251.
40. MCNAIR, D.M., M. LORR & L.F. DROPPLEMAN. 1992. The Profile of Mood States, 3rd
Edit. Educational and Industrial Testing Service. San Diego, CA.
41. KRUMHANSL, C.L. 1997. An exploratory study of musical emotions and psychophysiol-
ogy. Can. J. Exp. Psychol. 51: 336–352.
42. LENG, X. & G.L. SHAW. 1991. Toward a neural theory of higher brain function using
music as a window. Concepts Neurosci. 2: 229–258.
43. LENG, X., G.L. SHAW & E.L. WRIGHT. 1990. Coding of musical structure and the trion
model of the cortex. Music Percept. 8: 49–62.
44. PERETZ, I. 2001. Music perception and recognition. In The Handbook of Cognitive
Neuropsychology. B. Rapp, Ed.: 521–540. Psychology Press. Hove, UK.
370 ANNALS NEW YORK ACADEMY OF SCIENCES
45. PERETZ, I. 1996. Can we lose memories for music? A case of music agnosia in a non-
musician. J. Cognit. Neurosci. 8: 481–496.
46. PERETZ, I. & R. KOLINSKY. 1993. Boundaries of separability between melody and
rhythm in music discrimination: a neuropsychological perspective. Q. J. Exp. Psy-
chol. 46A: 301–325.
47. PERETZ, I. & J. MORAIS. 1989. Music and modularity. Contemp. Music Rev. 4: 277–
48. BESSON, M., F. FAÏTA, I. PERETZ, et al. 1998. Singing in the brain: independence of lyr-
ics and tunes. Psycholog. Sci. 9: 494–498.
49. PERETZ, I. & L. GAGNON. 1999. Dissociation between recognition and emotional judge-
ments for melodies. Neurocase 5: 21–30.
50. PERETZ, I., L. GAGNON & B. BOUCHARD. 1998. Music and emotion: perceptual determi-
nants, immediacy, and isolation after brain damage. Cognition 68: 111–141.
51. GROMKO, J.E. & A.S. POORMAN. 1998. Developmental trends and relationships in chil-
dren’s aural perception and symbol use. J. Res. Music Ed. 46: 16–23.
52. GORDON, E. 1979. Primary Measures of Music Audiation. G.I.A. Publications.
53. GORDON, E. 1982. Intermediate Measures of Music Audiation. G.I.A. Publications.
54. NELSON, D.J. & A.L. BARRESI. 1989. Children’s age-related intellectual strategies for
dealing with musical and spatial analogical tasks. J. Res. Music Ed. 37: 93–103.
55. LAMB, S.J. & A.H. GREGORY. 1993. The relationship between music and reading in
beginning readers. Ed. Psychol. 13: 19–27.
56. BARWICK, J., E. VALENTINE, R. WEST & J. WILDING. 1989. Relations between reading
and musical abilities. Br. J. Ed. Psychol. 59: 253–257.
57. DOUGLAS, S. & P. WILLATTS. 1994. The relationship between musical ability and liter-
acy skills. J. Res. Reading 17: 99–107.
58. LYNN, R., R.G. WILSON & A. GAULT. 1989. Simple musical tests as measures of Spear-
man's g. Personality Individ. Diff. 10: 25–28.
59. CHAN, A.S., Y.-C. HO & M.-C. CHEUNG. 1998. Music training improves verbal mem-
ory. Nature 396: 128.
60. HASSLER, M., N. BIRBAUMER & A. FEIL. 1985. Musical talent and visual-spatial ability:
a longitudinal study. Psychol. Music 13: 99–113.
61. HASSLER, M., N. BIRBAUMER & A. FEIL. 1987. Musical talent and visual-spatial ability:
onset of puberty. Psychol. Music 15: 141–151.
62. HURWITZ, I., P.H. WOLFF, B.D. BORTNICK & K. KOKAS. 1975. Nonmusical effects of the
Kodaly music curriculum in primary grade children. J. Learn. Disabil. 8: 167–174.
63. NEUFELD, K.A. 1986. Understanding of selected pre-number concepts: relationships to
a formal music program. Alberta J. Ed. Res. 32: 134–139.
64. GARDINER, M.F., A. FOX, F. KNOWLES & D. JEFFREY. 1996. Learning improved by arts
training. Nature 381: 284.
65. STANDLEY, J.M. & J.E. HUGHES. 1997. Evaluation of an early intervention music cur-
riculum for enhancing prereading/writing skills. Music Therapy Perspect. 15: 79–85.
66. GROMKO, J.E. & A.S. POORMAN. 1998. The effect of music training on preschoolers'
spatial-temporal task performance. J. Res. Music Ed. 46: 173–181.
67. BILHARTZ, T.D., R.A. BRUHN & J.E. OLSON. 2000. The effect of early music training on
child cognitive development. J. Applied Dev. Psychol. 20: 615–636.
68. COSTA-GIOMI, E. 1999. The effects of three years of piano instruction on children’s
cognitive development. J. Res. Music Ed. 47: 198–212.
69. RAUSCHER, F.H. & M.A. ZUPAN. 2000. Classroom keyboard instruction improves kin-
dergarten children’s spatial-temporal performance: a field experiment. Early Child-
hood Res. Q. 15: 215–228.
70. GRAZIANO, A.B., M. PETERSON & G.L. SHAW. 1999. Enhanced learning of proportional
math through music training and spatial-temporal reasoning. Neurol. Res. 21: 139–152.
71. ELBERT, T., C. PANTEV, C. WIENBRUCH, et al. 1995. Increased cortical representation of
the fingers of the left hand in string players. Science 270: 305–307.
72. SCHLAUG, G., L. JÄNCKE, Y. HUANG, et al. 1995. Increased corpus callosum size in
musicians. Neuropsychologia 33: 1047–1055.
371 SCHELLENBERG: MUSICAL ABILITY Download full-text
73. SCHLAUG, G., L. JÄNCKE, Y. HUANG & H. STEINMETZ. 1995. In vivo evidence of struc-
tural brain asymmetry in musicians. Science 267: 699–700.
74. TAKEUCHI, A.H. & S.H. HULSE. 1993. Absolute pitch. Psychol. Bull. 113: 345–361.
75. PANTEV, C., R. OOSTENVELD, A. ENGELIEN, et al. 1998. Increased auditory cortical rep-
resentation in musicians. Nature 392: 811–813.
76. MONAGHAN, P., N.B. METCALFE & G.D. RUXTON. 1998. Does practice shape the brain?
Nature 394: 434.
77. NAKADA, T., Y. FUJII, K. SUZUKI & I. KWEE. 1998. ‘Musical brain’ as revealed by high-
field (3 tesla) functional MRI. Neuroreport 9: 3853–3856.
78. GREENOUGH, W.T. & F.R. VOLKMAR. 1973. Pattern of dendritic branching in occipital
cortex of rats reared in complex environments. Exp. Neurol. 40: 491–504.
79. KEMPERMAN, G., H.G. KUHN & F.H. GAGE. 1997. More hippocampal neurons in adult
mice living in an enriched environment. Nature 386: 493–495.
80. CECI, S.J. 1991. How much does schooling influence general intelligence and its cog-
nitive components? A reassessment of the evidence. Dev. Psychol. 27: 703–722.
81. CECI, S.J. & W.M. WILLIAMS. 1997. Schooling, intelligence, and income. Am. Psychol.
82. MORRISON, F.J., E.M. GRIFFITH & D.M. ALBERTS. 1997. Nature-nurture in the class-
room: entrance age, school readiness, and learning in children. Dev. Psychol. 33:
83. MORRISON, F.J., L. SMITH & M. DOW-EHRENSBERGER. 1995. Education and cognitive
development: a natural experiment. Dev. Psychol. 31: 789–799.
84. BLATCHFORD, P. & P. MORTIMORE. 1994. The issue of class size for young children in
schools: what can we learn from research? Oxford Rev. Ed. 20: 411–428.
85. MOSTELLER, F. 1995. The Tennessee study of class size in the early school grades.
Future Child. 5: 113–127.
86. BARNETT, W.S. 1995. Long-term effects of early childhood programs on cognitive and
school outcomes. Future Child. 5: 25–50.
87. LAZAR, I. & R. DARLINGTON. 1982. Lasting effects of early education: a report from the
Consortium for Longitudinal Studies. Monogr. Soc. Res. Child Dev. 47 (2–3, Serial
88. LAMONT, A. 1998. Responses to Katie Overy's paper: respondent 3. Psychol. Music 26: