Do Musicians with Perfect Pitch Have More Autism Traits

than Musicians without Perfect Pitch? An Empirical Study

Anders Dohn1,2*, Eduardo A. Garza-Villarreal1,2,3, Pamela Heaton4, Peter Vuust1,2
1 Center of Functionally Integrative Neuroscience (CFIN), University of Aarhus, Aarhus, Denmark, 2 The Royal Academy of Music, Aarhus/Aalborg, Denmark, 3 Department
of Neurology, University Hospital, and Neuroscience Unit, Center for Research and Development in Health Sciences (CIDICS), Universidad Autonoma de Nuevo Leon,
Nuevo Leon, Mexico, 4 Department of Psychology, Goldsmiths University of London, London, United Kingdom

Abstract
Perfect pitch, also known as absolute pitch (AP), refers to the rare ability to identify or produce a musical tone correctly
without the benefit of an external reference. AP is often considered to reflect musical giftedness, but it has also been
associated with certain disabilities due to increased prevalence of AP in individuals with sensory and developmental
disorders. Here, we determine whether individual autistic traits are present in people with AP. We quantified subclinical
levels of autism traits using the Autism-Spectrum Quotient (AQ) in three matched groups of subjects: 16 musicians with AP
(APs), 18 musicians without AP (non-APs), and 16 non-musicians. In addition, we measured AP ability by a pitch
identification test with sine wave tones and piano tones. We found a significantly higher degree of autism traits in APs than
in non-APs and non-musicians, and autism scores were significantly correlated with pitch identification scores (r = .46,
p = .003). However, our results showed that APs did not differ from non-APs on diagnostically crucial social and
communicative domain scores and their total AQ scores were well below clinical thresholds for autism. Group differences
emerged on the imagination and attention switching subscales of the AQ. Thus, whilst these findings do link AP with
autism, they also show that AP ability is most strongly associated with personality traits that vary widely within the normal
population.

Citation: Dohn A, Garza-Villarreal EA, Heaton P, Vuust P (2012) Do Musicians with Perfect Pitch Have More Autism Traits than Musicians without Perfect Pitch? An
Empirical Study. PLoS ONE 7(5): e37961. doi:10.1371/journal.pone.0037961
Editor: Frank Krueger, George Mason University/Krasnow Institute for Advanced Study, United States of America
Received February 23, 2012; Accepted April 30, 2012; Published May 30, 2012
Copyright: ß 2012 Dohn et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits
unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Funding: This study was financially supported by The Royal Academy of Music, Aarhus, the Danish Ministry of Culture, and the Danish National Research
Foundation’s Center of Functionally Integrative Neuroscience (CFIN). The authors declare that the funders had no influence in study design, data collection and
analysis, decision to publish, or preparation of the manuscript.
Competing Interests: The authors have declared that no competing interests exist.
* E-mail: dohn@pet.auh.dk

Introduction autism who had not experienced early musical training and
observed statistically superior pitch naming in comparison with
Absolute pitch (AP) is the ability to identify the pitch of a musically-trained typically-developing controls with AP. These
musical tone or to produce a musical tone at a given pitch without results suggest that the genesis of AP may be different in ASD, and
the use of an external reference pitch [1,2]. The estimated that pitch information is encoded with increased specificity in
prevalence of AP is frequently reported to be around 1 per 10,000 these individuals.
[1,3,4], although a higher prevalence has been reported among In the current version of the Diagnostic Statistical Manual
East Asian populations [5,6]. AP possessors are able to retain (DSM-IV-TR) [23], autism is one of a cluster of disorders that are
accurate information about an isolated pitch along the one- characterized by impaired social and communicative skills and co-
dimensional continuum of auditory frequency, and they are able occurring repetitive behaviors. Asperger syndrome is a subcate-
to label that pitch within the context of the western chromatic gory that is characterized by the same core deficits as autism, but
scale [7]. It has been argued that the development of AP depends unlike autism, is not associated with delays in attaining early
on musical exposure in a critical period in early childhood [5,8– language and cognitive milestones. However, research suggests
10] as well as on possible genetic contributions [3,4,6,11,12]. that at later developmental stages, individuals diagnosed with
However, many musicians who begin training early in life and Asperger syndrome may show comparable levels of symptom
come from musical families do not develop AP, and relatively little severity as those who had experienced language and cognitive
is known about the traits and personality features associated with delays but were intellectually able (high-functioning autism) [24]
AP ability. In recognition of this and other failures to reliably differentiate
It has been suggested that AP ability may be associated to some between the different sub-categories detailed in DSM-IV-TR, it is
extent with certain deficits, since there seems to be an increase in proposed that the upcoming revision of the manual (DSM-V) will
prevalence of AP among people with sensory and developmental include a single spectrum disorder (ASD).
disabilities. For example, AP is frequently reported in individuals A notable feature of ASD is that unusual skills, as noted in the
with congenital blindness [13,14], Williams syndrome [15,16], and study by Heaton et al. [22] appear to be fairly common.
autism spectrum disorder (ASD) [17–21]. In one study, Heaton Theoretical accounts of autism, for example, the Enhanced
et al. [22] investigated AP in an intellectually-able adult with Perceptual Functioning theory [25,26], propose that individuals

PLoS ONE | www.plosone.org 1 May 2012 | Volume 7 | Issue 5 | e37961

 Perfect Pitch and Autism Traits

with ASD show superior perceptual discrimination and, in Recently, a preliminary interview-based study carried out by
addition, display an analytical cognitive style with increased Brown et al. [35] examined individual differences associated with
pattern discrimination abilities. Interestingly, Chin’s [27] two- the presence of AP in groups of classical trained musicians. The
factor model of AP describes a genetic predisposition toward an subjects were classified as being definitely socially eccentric,
‘‘analytical cognitive style’’, and this may account for increased somewhat eccentric, or not eccentric on the basis of interviewer’s
levels of AP in autism. notes regarding subjects’ communication style and nonverbal
Although ASD is considered to be a clearly defined neurode- behavior. The results showed that individuals classified as ‘‘socially
velopmental disorder there is an increasing recognition that some eccentric’’ were more likely to be AP possessors. Whilst these
of its defining characteristics can be observed at sub-clinical levels findings are intriguing, it must be noted that the data were largely
in the general population. Baron-Cohen et al. [28] have developed qualitative in terms of being interview-based and the groups were
the Autism-spectrum Quotient (AQ) questionnaire to measure not matched for age, age of onset of musical training, or musical
social skills, communication, imagination, attention to detail and preference.
attention switching in typical populations. The AQ questionnaire In the present study, we aimed to determine whether or not
includes 50 items (i.e. personal statements) and participants are musicians with AP show increased levels of autism traits compared
required to indicate whether or not the statements apply to them. to matched groups of musicians without AP and non-musicians.
For example, the statement ‘‘I enjoy social chit-chat’’ is included in Accordingly, we measured the autism traits of the participants
the communication subscale question set, and ‘‘I prefer to do things the quantitatively by administering the AQ and the level of AP ability
same way over and over again’’ is included in the attention switching using a pitch identification test, on the hypothesis of a correlation
subscale. The score ranges from 0 to 50, with higher scores between the two. Finding higher scores on the AQ subscales
indicating a higher prevalence of autistic traits. Baron-Cohen measuring social and communication deficits in our AP group
suggests that individuals who score above 32 points should be would be consistent with the results of Brown et al. [35], whereas
considered to have clinically significant levels of autistic traits [28]. differences on the subscales measuring attention-switching and
Three of the five different AQ subscales measure what has come to attention to detail would support the model by Chin [27] who
be known as the ‘‘triad’’ of impairments characterizing autism. claim that AP is associated with detailed processing style. Finally,
Hence, these questions probe social and communication skills and we examined whether musical abilities vary with degrees of autism
a tendency to repetitive behavior. The remaining subscales probe traits and whether the level of AP ability is reliably related to
for characteristics of autism that have been identified in musical aptitude.
experimental studies. These include difficulties in imagination,
difficulties in attention-switching, and exceptional attention to Methods
detail. Difficulties in attention switching are associated with poor
cognitive flexibility and are consistent with work showing impaired Ethics Statement
executive functions in ASD [29]. Exceptional attention has been This study was approved by the local ethics committee (De
associated with special skills and may be associated with an Videnskabsetiske Komiteer For Region Midtjylland, Denmark)
analytical cognitive style detailed in the models of autism outlined and was performed in accordance with the Code of Ethics of the
by Mottron [26] and Baron-Cohen [30]. Interestingly, given World Medical Association (Declaration of Helsinki). Written
questions about the prevalence of AP in autism, an analytical informed consent was obtained from each participant after
cognitive style is an important component in the model of AP detailed explanation of the experimental procedure and the object
proposed by Chin [27]. of the study.
Studies using the AQ with typical adults have shown that
natural science students have higher AQ than students from the Participants
social sciences and humanities, and that mathematicians have Fifty participants with a mean age of 27 years (SD = 6.5,
higher AQ than non-mathematician scientists [28]. These findings range = 19–43) were recruited for the study. They consisted of the
are consistent with previous studies showing an association following: 16 musicians with self-reported AP (APs), 18 musicians
between science/maths skills and autistic conditions [31] using without AP (non-APs), and 16 non-musicians. The three groups
other methods. Other studies using the AQ have shown that high were matched with regard to gender (APs: 13 male, 3 female; non-
AQ scorers are faster to complete the Embedded Figures (EF) test APs: 13 male, 5 female; non-musicians: 13 male, 3 female,
compared with low AQ scorers, independent of global IQ scores x2 = .55, p..05) and age (APs: M = 29.0, SD = 7.3; non-APs:
[32]. The EF task requires individuals to locate a simple form that M = 29.2, SD = 7.0; non-musicians: M = 23.6, SD = 2.8; H(3) = 5.7,
is embedded in a larger visual display. It provides a measure of the p..05). The two groups of musicians were also matched with
individual’s field independence, defined as the ability to isolate regard to age of onset of musical training (APs: M = 5.5, SD = 2.1;
details from their surrounding context. An early study by Shah non-APs: M = 5.6, SD = 2.0, t = 20.08, p..05) as well as to their
and Frith [33] revealed superior EF task performance in children preferred style of music (APs: 5 classical musicians, 11 jazz/rock/
with autism, and these more recent findings, showing similarly pop musicians; non-APs: 6 classical musicians, 12 jazz/rock/pop
superior performance in typical individuals with high AQ suggest musicians, x2 = .017, p..05). All participants received compensa-
similarities in cognitive style across these two groups. However, tion for being in the study.
there is some evidence that abnormalities in perceptual processing
are characteristics of individuals with high AQ scores. In a study Pitch Identification Test
by Gomot et al. [34] it has been observed that individuals with To confirm self-reported AP and to distinguish APs from non-
high AQ scores demonstrate superior auditory novelty detection, APs, the two musician-groups were asked to complete an online
revealed by shorter reaction time, in a task requiring response to pitch identification test (PIT) available from Athos et al. [36] and
target stimuli in an oddball paradigm. They also displayed developed by Baharloo et al. [37]. The non-musicians were not
activation of an unusually widespread network of brain regions asked to take this test since it requires familiarity to musical note
that are also observed in individuals with a formal diagnosis of names. The PIT consisted of 80 trials: 40 randomly-selected pure
autism [34]. tones (i.e. computer-generated sine waves without overtones) and

PLoS ONE | www.plosone.org 2 May 2012 | Volume 7 | Issue 5 | e37961

 Perfect Pitch and Autism Traits

40 randomly-selected digitized piano tones. The participants were Procedure

asked to listen to the presented tones and to identify them by The APs were primarily identified through word of mouth and
responding via an onscreen piano keyboard. A short practice test through advertisements at the Danish Royal Academy of Music
was given first to acquaint participants with the keyboard response and the Music Department at the local university. The non-APs
on the computer screen and to adjust the volume of the computer. and the non-musicians were found subsequently through adver-
The tones had duration of 1 second with a 2-second interlude tisements and were selected randomly using matching criteria.
between tone onsets. They were delivered in series of 10 tones, The participants were tested individually. First, the participant
giving the participant an opportunity to pause between sets. answered a short questionnaire regarding age, gender, musical
Four of the pure tones and four of the piano tones were background, and experience with AP. The non-APs and the non-
excluded from the scoring due to their position at the outermost musicians were told that they did not have to answer the questions
range of the keyboard, resulting in 72 counting trials. regarding AP experience. Then, the APs and non-APs completed
Participants were given 1 point for each correct answer the PIT on a computer with in-ear-stereo headphones. Next, each
(maximum score for the pure tone test and the piano tone participant completed the MET with paper and pen, with the
test together is 72) and 3/4 point for each error of a semitone. audio part played back from the computer. Next, they completed
We averaged the participant’s scores in mean pure tones and the AQ on the same computer.
mean piano tones, and those who scored above a threshold of The APs were tested for 90 minutes using EEG and MRI as
36 were designated APs, whereas the rest of the participants part of other ongoing experiments before the behavioral tests. No
were designated non-APs. The probability of testing AP by participants reported problems with either the auditory stimuli or
chance alone with this threshold of 36 in total (pure tones + the answering procedure in the PIT, the MET, and AQ.
piano tones) is 1.21610210 [36]. Note that the mean expected
score by chance is 14.25 with 95% of expected values lying Statistical Analysis
between scores of 8.5 and 20.75. Data from each group were assessed for normality with the
Kolmogorov-Smirnov test and the Shapiro-Wilk test, which
The Musical Ear Test revealed violations of normality assumptions. Accordingly, we
To examine the musical aptitude of the participants and to used non-parametric tests for subsequent statistical analyses.
clearly distinguish musicians from non-musicians, all participants We compared the AQ score between the three groups using the
completed the musical ear test (MET), a newly developed test Kruskal-Wallis test. To investigate which autistic factors contribute
designed for measuring musical abilities objectively and quantita- most to the difference between groups, we studied the five factors
tively in both musicians and non-musicians [38]. The test consists of the AQ using Kruskal-Wallis test. This analysis was followed up
of 104 trials in which participants listen to two musical phrases and by Mann-Whitney rank sum tests to investigate the differences
subsequently judge whether or not they were identical by between groups. Effect sizes were calculated as: r~ pZffiffiffi N
from the
responding on an answer sheet. The first half of the test is a Mann-Whitney rank sum tests. Furthermore, we performed
melodic subtest consisting of 52 pairs of melodic phrases, played Spearman correlation to investigate the relation between the
with sampled piano sounds, and the other half is a rhythm subtest PIT score and the AQ score.
consisting of 52 trials with rhythmical phrases, played with wood We used the Kruskal-Wallis test for comparison of age and for
block sound. Before each subtest, participants are given two the total MET scores from all three groups, and Mann-Whitney
example trials with feedback. Half of the trials in each session (26) rank sum test was used for comparison of PIT scores from the APs
and non-APs. All analyses were corrected for multiple comparisons
are ‘‘same’’ trials and half are ‘‘different’’ trials, with the order
using the Bonferroni correction for p,.05.
randomized in both sessions. The participant’s score is the
percentage of correct answers out of the 104 trials.
Results
Autism-spectrum Quotient Autism-spectrum Quotient
All participants also completed the Autism-spectrum quotient The mean AQ scores are shown in Table 1. We found a
(AQ), a self-administered questionnaire developed by Baron- statistically significant difference between the three groups
Cohen et al. [28] that quantitatively measures the degree to which (H(3) = 11.32, p = .003). To follow up this finding, Mann-Whitney
an adult with normal intelligence has autistic traits. Although the tests revealed a significant difference between APs and non-AP’s
scoring scale is not a diagnostic measure, those with autism (U = 51.00, p = .007, r = 2.55) and between APs and non-
spectrum disorder have been shown to score at the high end of the musicians (U = 62.00, p = .036, r = 2.43). However, no significant
scale [28,39], and it has been validated against clinical diagnoses difference was found between non-APs and non-musicians
[40]. The test consists of 50 items, made up of 10 questions (U = 128.50, p..05) (Fig. 1). These results show that absolute
assessing five subscales: ‘‘social skills’’, ‘‘attention switch’’, ‘‘atten- pitch possessors score significantly higher than individuals without
tion to detail’’, ‘‘communication’’ and ‘‘imagination’’. Half the absolute pitch in the autism-spectrum quotient.
questions are formulated to elicit an ‘agree’ response and the other Correlation. We found a positive correlation between the
half a ‘disagree’ response. PIT score and the AQ score (r = .46, p = .003) (Fig. 2). This shows
The participants answered the questions on a computer by that high scores among musicians on the pitch identification test
clicking their response on a multiple choice table with four are associated with high scores on the autism-spectrum quotient.
response boxes entitled ‘‘definitely agree’’, ‘‘slightly agree’’, slightly There was no correlation between the MET and the AQ scores,
disagree’’, or ‘‘definitely disagree’’. We used the scoring procedure suggesting that musical abilities may not vary with the level of
described by Baron-Cohen et al. [28] ranging from 0 to 50, with autism traits.
higher scores indicating greater inclination towards autistic traits.
Hence, higher scores in the subscales indicate poorer social skill, Autism-spectrum Quotient Factors
poorer attention switching, stronger attention to detail, poorer The descriptive data and statistical results are presented in
communication skill, and poorer imagination. Table 2 and Figure 3. We found a statistically significant difference

PLoS ONE | www.plosone.org 3 May 2012 | Volume 7 | Issue 5 | e37961

 Perfect Pitch and Autism Traits

Figure 1. AQ score. Box plot showing the autism-spectrum quotient (AQ) score of absolute pitch possessors (APs), musicians without absolute
pitch (non-APs), and non-musicians. The red dashed line shows the proposed AQ cut-off for distinguishing individuals who have clinically significant
levels of autistic traits, according to Baron-Cohen (2001).
doi:10.1371/journal.pone.0037961.g001

in the ‘‘imagination’’ factor (H(3) = 10.2, p = .049); the APs showed Musical Ear Test
significantly higher AQ score in imagination (i.e. less imaginative). The MET scores were distributed as follows: APs mean = 87.3%
This indicates that the imagination factor contributed strongly to (SD = 6.1), non-APs mean = 84.2% (SD = 6.8), and non-musicians
the AQ score. However, we cannot rule out contributions from mean = 69.4% (SD = 8.6) (Fig. 5). A statistically significant
other factors, such as attention switching, because we noted also difference was found between the three groups (H(3) = 26.23,
differences in that factor, although they failed to reach conven- p,.0001). Planned post-hoc Mann-Whitney tests revealed a
tional levels of statistical significance. significant difference between APs and non-musicians
(U = 10.00, p,.0001) and between non-APs and non-musicians
Pitch Identification Test (U = 20.00, p,.0001). However, no statistically significant differ-
The APs had a mean PIT score of 60.1 (SD = 11.1, ranging from ence was found between APs and non-APs (U = 108.50, p..05),
39 to 71.75), whereas the non-APs had a mean PIT score of 15.5 and no significant correlation between PIT and MET was found
(SD = 5.1, ranging from 9.5 to 30.75) (Fig. 4). This shows that the (r = .24, p..05).
APs are unambiguously distinguished from the non-APs (U These results suggest a clear distinction between musicians (with
,.0001, p,.0001) in that, they do not overlap. or without AP) and non-musicians, and that the level of AP ability
among musicians is unrelated to the level of musical abilities as
measured by the MET.

Discussion
Table 1. Descriptive statistics of AQ scores.
Here, we show that musicians with absolute pitch (AP) score
higher than individuals without AP (musicians and non-musicians)
AQ scores Mean SD on the autism-spectrum quotient (AQ), and that AP accuracy
correlates with AQ. However, our results showed that the
APs 19.6 6.0
association between AQ traits and AP resulted from group
Non-APs 12.8 6.0 differences in scores on the imagination and attention shifting
Non-musicians 13.8 5.9 subscales rather than on scores on the social and communication
AS/HFA 35.8 6.5 subscales.
This finding is surprising given that the results from Brown et al.
Means and standard deviations (SD) of autism-spectrum quotient (AQ) scores of [35] concluded that AP possessors (APs) are more likely to have
musicians with absolute pitch (APs), musicians without absolute pitch (non-
APs), non-musicians, and a group of adults with Asperger syndrome (AS) or
impairments in social behavior than non-possessors (non-APs).
high-functioning autism (HFA) as reported in Baron-Cohen [28]. Furthermore, they speculated that AP is an example of piecemeal
doi:10.1371/journal.pone.0037961.t001 information processing, an enhanced attention to isolated details of

PLoS ONE | www.plosone.org 4 May 2012 | Volume 7 | Issue 5 | e37961

 Perfect Pitch and Autism Traits

Figure 2. AQ/PIT scores. Scatterplot showing the autism-spectrum quotient (AQ) score as a function of the pitch identification score (PIT). The red
line indicates the absolute pitch (AP) cut-off. To the right side of the red line are the musicians with AP (APs) and to the left side are the musicians
without AP (non-APs).
doi:10.1371/journal.pone.0037961.g002

Figure 3. AQ factors. Bar plot showing the mean autism-spectrum quotient (AQ) factor score of all groups. The error bars indicate the standard
error of the mean.
doi:10.1371/journal.pone.0037961.g003

PLoS ONE | www.plosone.org 5 May 2012 | Volume 7 | Issue 5 | e37961

 Perfect Pitch and Autism Traits

Table 2. Statistical results of AQ factors.

APs Non-APs Non-musicians Kruskal-Wallis

AQ factors Mean (SD) Mean (SD) Mean (SD) H p

Social skills 2.3 (1.9) 1.3 (1.5) 1.2 (1.4) 4.1 Ns

Attention switching 5.7 (2.3) 3.6 (2.3) 3.4 (2.1) 9.1 (.08)
Attention to detail 5.2 (2.7) 5.2 (2.4) 5.0 (1.6) 0.1 Ns
Communication 2.8 (1.9) 1.1 (1.4) 2.1 (1.8) 8.1 Ns
Imagination 3.6 (1.8) 1.6 (1.5) 2.1 (1.8) 10.2 .049

Means and standard deviation (SD) of autism-spectrum quotient (AQ) factor score for musicians with absolute pitch (APs), musicians without absolute pitch (non-APs),
and non-musicians with standard deviation (SD) in parentheses. The H value relates to a Kruskal-Wallis test between the three groups. Ns = not statistically significant.
doi:10.1371/journal.pone.0037961.t002

a configuration at the expense of attention to the whole. In the Brown et al [35], and may also challenge the importance of
earliest study of AP in autism, Heaton, Hermelin & Pring [17] analytical style as outlined in the model of AP by Chin [27]. In one
observed a highly significant correlation between pitch memory study of a musical savant with AP, Mottron et al. [41] noted
scores and scores from a cognitive task (block design) that was features commonly observed in individuals with executive function
taken as a marker for a local bias or an analytical cognitive style deficits and suggested that absolute pitch may result from
characteristic of autism. However, later studies have failed to executive function difficulties, most notably a lack of cognitive
observe this association [18,19], and it does not appear that a local flexibility, in a person with a marked interest for auditory stimuli.
processing style is a necessary precursor for AP in individuals with Whilst our comparison of AP and non-AP groups did not reveal a
autism. Consistent with this conclusion are the results from the significant difference on the attention-switching factor, scores were
current study showing that APs and non-APs did not differ in their markedly higher for the AP group, and that trend may add
attention to detail. Hence, our findings do not indicate piecemeal support to the suggestion that reduced cognitive flexibility is
information processing by musicians with AP as suggested by implicated in AP.
It is important to emphasize that even though our AP possessors
achieved reliably higher AQ scores than our non-possessors, they
did not, with one exception, have scores above 32 which is the cut-
off for the DSM-IV-TR criteria for high functioning autism as
suggested by Baron-Cohen et al. [28]. The only person in our
study who exceeded that score was an AP possessor who obtained
a score of 33 but did not evidence any social or communication
disability and had never been given a diagnosis of autism or
related disorder. Thus, our findings, whilst showing that AP
possessors exhibit more traits associated with the broad autism
phenotype than non-possessors and non-musicians, do not support
the notion of increased social and communication disability in
musicians with AP.
Interestingly, the musicians without AP had a lower mean AQ
than the non-musicians (albeit not statistically significant). The
results showing that the mean scores for the communication factor
were twice as high for the non-musicians as for the musicians
without AP are particularly striking. Whilst the sample size may
not be large enough to allow for a statistically significant difference
across groups, the result does suggest that musicians without AP
show minimal AQ traits. However, musicians constantly commu-
nicate with sound to create their musical artwork, and commu-
nication within musical ensembles and with audiences is an
essential element of playing music [42–44]. This may explain why
musicians without AP display even fewer communication impair-
ments than non-musicians as measured by the AQ.
An interesting finding, possibly related to communicative
abilities in musicians, was the observed difference between groups
Figure 4. PIT score. Bar plot showing the mean pitch identification on the imagination subscale of the AQ. This factor clearly
test (PIT) score of the absolute pitch possessors (APs) and the musicians contributed notably to the difference in total AQ scores across the
without absolute pitch (non-APs). The error bars indicate the standard three groups. However, this finding should not be interpreted as
error of the mean. The red line indicates the threshold for possessing evidence that the APs have high deficits in imagination. When
absolute pitch (.36), and the dashed lines indicate the range of scores
expected by chance distribution, with a mean expected score by
comparing the mean imagination score of our groups of APs, non-
chance of 14.25, and with 95% of expected values lying between scores APs, and non-musicians (see Table 2) with the controls and student
of 8.5 and 20.75. controls in Baron-Cohen’s AQ study [28] (Controls: M = 2.3,
doi:10.1371/journal.pone.0037961.g004 SD = 1.7; Students: M = 2.5, SD = 1.9), it becomes clear that the

PLoS ONE | www.plosone.org 6 May 2012 | Volume 7 | Issue 5 | e37961

 Perfect Pitch and Autism Traits

Figure 5. MET score. Bar plot showing the mean total musical ear test (MET) score of absolute pitch possessors (APs), musicians without absolute
pitch (non-APs), and non-musicians. The error bars indicate the standard error of the mean.
doi:10.1371/journal.pone.0037961.g005

score of the non-AP musicians in our study was extraordinarily the group difference in attention switching failed to reach
low. This could be explained by the fact that musicians in general statistical significance in our study, scores from this subscale
use their imagination in order to express a certain musical style contributed substantially to the observed difference in total AQ
with creativity and empathy, especially in music that involves scores and it would be interesting to explore cognitive flexibility
improvisation. However, when we compare the imagination score and its relationship with imagination in a larger group of
of the APs in our study with Baron-Cohen’s group of adults with individuals with AP. Results obtained by Mottron et al., suggested
Asperger syndrome or high-functioning autism (M = 6.4, SD = 2.1), that reduced cognitive flexibility may be implicated in AP in
we see that the two groups differ markedly. Hence, we propose autism, and extending this work to the non-autistic population
that while musicians with AP may be less imaginative than may provide important new insights into absolute pitch ability.
musicians without AP and non-musicians, they do not show
clinically significant imagination deficits. Acknowledgments
One limitation in the study of musicians with and without AP by
Brown et al. [35] was that they did not include a group of non- We wish to thank all the musicians with and without absolute pitch and the
musicians in their study. Our inclusion of a non-musician group non-musicians for participating in this study. We also wish to thank Line
enabled us to draw comparisons between musician and non- Gebauer for her help with the data acquisition from the non-musicians and
Donald F. Smith for final language revision.
musician groups. Further, our use of the AQ enabled us to provide
a more detailed account of the traits associated with AP.
In conclusion, our findings show that AP ability is not associated Author Contributions
with deficits in social and communication abilities in typical Conceived and designed the experiments: AD PV. Performed the
populations and this challenges previous work making such links. experiments: AD. Analyzed the data: AD EAGV PH. Contributed
Future research might focus on further exploring our key finding reagents/materials/analysis tools: AD EAGV. Wrote the paper: AD
showing differences in imagination in APs and non-APs. Whilst EAGV PH PV.

References
1. Takeuchi AH, Hulse SH (1993) Absolute pitch. Psychol Bull 113: 345–361. 9. Miyazaki K, Ogawa Y (2006) Learning absolute pitch by children: A cross-
2. Ward WD (1999) Absolute Pitch. In: Deutsch D, ed. The Psychology of Music. sectional study. Music Percept 24: 63–78.
San Diego: Academic. pp 265–298. 10. Miyazaki K (2004) How well do we understand absolute pitch? Acoust Sci
3. Profita J, Bidder TG (1988) Perfect pitch. Am J Med Genet 29: 763–771. Technol 25: 426–432.
4. Bachem A (1955) Absolute Pitch. J Acoust Soc Am 27: 1180–1185. 11. Baharloo S, Service SK, Risch N, Gitschier J, Freimer NB (2000) Familial
5. Deutsch D, Henthorn T, Marvin E, Xu H (2006) Absolute pitch among aggregation of absolute pitch. Am J Hum Genet 67: 755–758.
American and Chinese conservatory students: prevalence differences, and 12. Brown WA, Sachs H, Cammuso K, Folstein SE (2002) Early music training and
evidence for a speech-related critical period. J Acoust Soc Am 119: 719–722. absolute pitch. Music Percept 19: 595–597.
6. Gregersen PK, Kowalsky E, Kohn N, Marvin EW (1999) Absolute pitch: 13. Hamilton R, Pascual-Leone A, Rodriguez D, Schlaug G (2000) Increased
prevalence, ethnic variation, and estimation of the genetic component. prevalence of absolute pitch in blind musicians. Society for Neuroscience
Am J Hum Genet 65: 911–913. Abstracts 26.
7. Levitin DJ (2004) Absolute pitch: Self-reference and memory. Annee 14. Hamilton RH, Pascual-Leone A, Schlaug G (2004) Absolute pitch in blind
Psychologique 104: 103–120. musicians. Neuroreport 15: 803–806.
8. Sergeant D (1969) Experimental Investigation of Absolute Pitch. Journal of
Research in Music Education 17: 135–143.

PLoS ONE | www.plosone.org 7 May 2012 | Volume 7 | Issue 5 | e37961

 Perfect Pitch and Autism Traits

15. Martens MA, Reutens DC, Wilson SJ (2010) Auditory cortical volumes and high-functioning autism, and normal sex differences. Philosophical Transactions
musical ability in Williams syndrome. Neuropsychologia 48: 2602–2609. DOI of the Royal Society of London Series B-Biological Sciences 358: 361–374.
10.1016/j.neuropsychologia.2010.05.007. 31. Baron-Cohen S (1998) Does Autism Occur More Often in Families of Physicists,
16. Lenhoff HM, Perales O, Hickok G (2001) Absolute pitch in Williams syndrome. Engineers, and Mathematicians? Autism 2: 296–301.
Music Percept 18: 491–503. 32. Grinter EJ, Van Beek PL, Maybery MT, Badcock DR (2009) Brief Report:
17. Heaton P, Hermelin B, Pring L (1998) Autism and pitch processing: A precursor Visuospatial Analysis and Self-Rated Autistic-Like Traits. Journal of Autism and
for savant musical ability? Music Percept 15: 291–305. Developmental Disorders 39: 670–677.
18. Heaton P, Pring L, Hermelin B (1999) A pseudo-savant: A case of exceptional 33. Shah A, Frith U (1983) An Islet of Ability in Autistic-Children - A Research
musical splinter skills. Neurocase 5: 503–509. Note. Journal of Child Psychology and Psychiatry and Allied Disciplines 24:
19. Heaton P (2003) Pitch memory, labelling and disembedding in autism. J Child 613–620.
Psychol Psychiatry 44: 543–551. 34. Gomot M, Belmonte MK, Bullmore ET, Bernard FA, Baron-Cohen S (2008)
20. Bonnel A, Mottron L, Peretz I, Trudel M, Gallun E, et al. (2003) Enhanced Brain hyper-reactivity to auditory novel targets in children with high-functioning
pitch sensitivity in individuals with autism: a signal detection analysis. J Cogn autism. Brain 131: 2479–2488.
Neurosci 15: 226–235.
35. Brown WA, Cammuso K, Sachs H, Winklosky B, Mullane J, et al. (2003)
21. Rimland B, Fein D (1988) Special Talents of Autistic Savants. In: Obler LK,
Autism-related language, personality, and cognition in people with absolute
Fein D, eds. The exceptional brain. New York: Guilford Press. pp 474–492.
pitch: results of a preliminary study. J Autism Dev Disord 33: 163–167.
22. Heaton P, Davis RE, Happe FGE (2008) Research note: Exceptional absolute
pitch perception for spoken words in an able adult with autism. Neuropsycho- 36. Athos EA, Levinson B, Kistler A, Zemansky J, Bostrom A, et al. (2007)
logia 46: 2095–2098. 10.1016/j.neuropsychologia.2008.02.006. Dichotomy and perceptual distortions in absolute pitch ability. Proc Natl Acad
23. APA (2000) Diagnostic and Statistical Manual of Mental Disorders. Washington Sci U S A 104: 14795–14800.
DC: American Psychiatric Association. 37. Baharloo S, Johnston PA, Service SK, Gitschier J, Freimer NB (1998) Absolute
24. Howlin P (2003) Outcome in high-functioning adults with autism with and pitch: an approach for identification of genetic and nongenetic components.
without early language delays: Implications for the differentiation between Am J Hum Genet 62: 224–231.
autism and Asperger syndrome. Journal of Autism and Developmental Disorders 38. Wallentin M, Nielsen AH, Friis-Olivarius M, Vuust C, Vuust P (2010) The
33: 3–13. Musical Ear Test, a new reliable test for measuring musical competence. Learn
25. Mottron L, Dawson M, Soulieres I, Hubert B, Burack J (2006) Enhanced Individ Differ 20: 188–196.
perceptual functioning in autism: An update, and eight principles of autistic 39. Wheelwright S, Baron-Cohen S, Goldenfeld N, Delaney J, Fine D, et al. (2006)
perception. Journal of Autism and Developmental Disorders 36: 27–43. Predicting autism spectrum quotient (AQ) from the systemizing quotient-revised
26. Mottron L, Dawson M, Soulieres I (2009) Enhanced perception in savant (SQ-R) and empathy quotient (EQ). Brain Research 1079: 47–56.
syndrome: patterns, structure and creativity. Philosophical Transactions of the 40. Woodbury-Smith MR, Robinson J, Wheelwright S, Baron-Cohen S (2005)
Royal Society B-Biological Sciences 364: 1385–1391. Screening adults for asperger syndrome using the AQ: A preliminary study of its
27. Chin CS (2003) The Development of Absolute Pitch: A Theory Concerning the diagnostic validity in clinical practice. Journal of Autism and Developmental
Roles of Music Training at an Early Developmental Age and Individual Disorders 35: 331–335.
Cognitive Style. Psychology of Music 31: 155–171. 41. Mottron L, Peretz I, Belleville S, Rouleau N (1999) Absolute pitch in autism: A
28. Baron-Cohen S, Wheelwright S, Skinner R, Martin J, Clubley E (2001) The case study. Neurocase 5: 485–501.
Autism-Spectrum Quotient (AQ): Evidence from Asperger syndrome/high- 42. Vuust P, Roepstorff A (2008) Listen up! Polyrhythms in Brain and Music.
functioning autism, males and females, scientists and mathematicians. Journal of Cognitive Semiotics 135–159.
Autism and Developmental Disorders 31: 5–17. 43. Monson I (1996) Saying something: Jazz Improvisation and Interaction.
29. Hill EL (2004) Executive dysfunction in autism. Trends in Cognitive Sciences 8: Chicago: University of Chicago Press.
26–32. 44. Berliner PF (1994) Thinking in Jazz: The Infinite Art of Improvisation. Chicago:
30. Baron-Cohen S, Richler J, Bisarya D, Gurunathan N, Wheelwright S (2003) The University of Chicago Press.
The systemizing quotient: an investigation of adults with Asperger syndrome or

PLoS ONE | www.plosone.org 8 May 2012 | Volume 7 | Issue 5 | e37961