Pitch discrimination and melodic memory in children with autism spectrum disorders

Abstract

Background:
Pitch perception is enhanced among persons with autism. We extended this finding to memory for pitch and melody among school-aged children.

Objective:
The purpose of this study was to investigate pitch memory in musically untrained children with autism spectrum disorders, aged 7-13 years, and to compare it to that of age- and IQ-matched typically developing children.

Methods:
The children were required to discriminate isolated tones in two differing contexts as well to remember melodies after a period of 1 week. The tasks were designed to employ both short- and long-term memory for music. For the pitch discrimination task, the children first had to indicate whether two isolated tones were the same or different when the second was the same or had been altered to be 25, 35, or 45 cents sharp or flat. Second, the children discriminated the tones within the context of melody. They were asked whether two melodies were the same or different when the leading tone of the second melody was the same or had been altered to be 25, 35, or 45 cents sharp or flat. Long-term memory for melody was also investigated, as the children attempted to recall four different two-bar melodies after 1 week.

Results:
The children with autism spectrum disorders demonstrated elevated pitch discrimination ability in the single-tone and melodic context as well as superior long-term memory for melody. Pitch memory correlated positively with scores on measures of nonverbal fluid reasoning ability.

Conclusion:
Superior short- and long-term pitch memory was found among children with autism spectrum disorders. The results indicate an aspect to cognitive functioning that may predict both enhanced nonverbal reasoning ability and atypical language development.

Full text

Autism
2014, Vol. 18(2) 137 –147
© The Author(s) 2012
Reprints and permissions:
sagepub.co.uk/journalsPermissions.nav
DOI: 10.1177/1362361312462905
aut.sagepub.com
Introduction
Persons with autism show enhanced perception of pitch in
comparison to typically developing (TD) persons (Bonnel
et al., 2003; Heaton, 2003; Heaton et al., 1998). Enhanced
pitch perception in autism appears to be attributable to
strengths in both short- and long-term memory. For exam-
ple, enhanced short-term memory for pitch was supported
by the finding that young adults and adolescents with high-
functioning autism discriminated between two tones 18
cents apart, or less than 1/4 of a semitone, more accurately
than TD persons matched at the group level for IQ (Bonnel
et al., 2003).
Similarly, enhanced long-term memory for pitch was
supported by evidence that school-aged boys with autism
had more accurate long-term memory for pitch sounds over
speech sounds when paired with pictures, as compared to
IQ-matched TD children who were better able to remember
the speech sounds (Heaton et al., 1998). In addition, an
exceptional form of long-term memory for pitch, referred
to as absolute pitch (AP) ability, is well documented among
musical savants who have autism (Hermelin, 2001; Miller,
1999; Rimland and Fein, 1988; Sloboda et al., 1985). AP is
a rare form of memory that enables the pitches of notes to
be classified out of context. In typical populations, this
exceptional memory has an estimated prevalence of 1 in
Pitch discrimination and melodic
memory in children with autism
spectrum disorders
Sandy Stanutz, Joel Wapnick and Jacob A Burack
Abstract
Background: Pitch perception is enhanced among persons with autism. We extended this finding to memory for pitch
and melody among school-aged children.
Objective: The purpose of this study was to investigate pitch memory in musically untrained children with autism
spectrum disorders, aged 7–13 years, and to compare it to that of age- and IQ-matched typically developing children.
Methods: The children were required to discriminate isolated tones in two differing contexts as well to remember
melodies after a period of 1 week. The tasks were designed to employ both short- and long-term memory for music.
For the pitch discrimination task, the children first had to indicate whether two isolated tones were the same or
different when the second was the same or had been altered to be 25, 35, or 45 cents sharp or flat. Second, the children
discriminated the tones within the context of melody. They were asked whether two melodies were the same or
different when the leading tone of the second melody was the same or had been altered to be 25, 35, or 45 cents sharp
or flat. Long-term memory for melody was also investigated, as the children attempted to recall four different two-bar
melodies after 1 week.
Results: The children with autism spectrum disorders demonstrated elevated pitch discrimination ability in the single-
tone and melodic context as well as superior long-term memory for melody. Pitch memory correlated positively with
scores on measures of nonverbal fluid reasoning ability.
Conclusion: Superior short- and long-term pitch memory was found among children with autism spectrum disorders.
The results indicate an aspect to cognitive functioning that may predict both enhanced nonverbal reasoning ability and
atypical language development.
Keywords
absolute pitch, autism, melodic memory, pitch discrimination, visual nonverbal reasoning ability
Article
McGill University, Canada
Corresponding author:
Sandy Stanutz, 1105 Edward Street, Manotick, ON K4M 1G8, Canada.
Email: sandy.stanutz@mail.mcgill.ca

138 Autism 18(2)
10,000 (Profita and Bidder, 1988), but the incidence of AP
among persons with autism is estimated to be as high as 1
in 20 (Brown et al., 2003). This estimate was calculated
from a survey of 5400 parents of children with autism,
where musical memory ability for melody and individual
pitch were by far the most reported of the savant abilities
(Rimland, 1988). Although tests of AP ability were not con-
ducted, parental reports suggested behaviors associated
with AP ability across the spectrum of autism. To what
degree these children would have tested positively for a test
of AP ability is unknown. However, the indication of
enhanced musical memory in children with autism raises
the question of whether the AP ability of musical savants is
reflected in nonsavant populations with autism as enhanced
pitch memory. Traditionally, AP ability has generally been
considered to be an absolute ability that one either pos-
sesses or does not possess, but researchers have recognized
varying levels to the skill (Baharloo et al., 1998; Bermudez
and Zatorre, 2009). This kind of classification may help to
explain enhanced memory for pitch and melody among
persons with autism who do not have musical savant
syndrome.
Children with autism demonstrate high levels of musical
ability in comparison to TD children (Armstrong and
Darrow, 1999) both in the replication of tone sequences
(Thaut, 1988) and in the sophistication of improvised
melodic sequences (Applebaum et al., 1979). This is evi-
dence that young children with autism appear to be both
attending to and retaining the musical harmonic conven-
tions of culture, with respect to maintaining the key of a
musical excerpt, in a developmentally unusual manner. For
example, TD preschool children have difficulty in retaining
a tonal center or singing in key (Flowers and Dunne-Sousa,
1990; Moog, 1976; Moorhead and Pond, 1978), and there is
evidence that many TD children up to 11 years of age do not
accurately remember semitones (Bentley, 1966). In contrast,
Heaton (2003) found that children with high-functioning
autism have enhanced memory for chord tones, a surprising
finding given the sophistication of this musical task. Thus,
developmental differences in pitch memory are evident
between TD children and children with autism spectrum
disorders (ASD), and there is evidence that this develop-
mental difference may reveal the cognitive profile of a new
subgrouping of children along the spectrum of autism. For
example, Heaton et al. (2008b) found an unusually superior
memory for melodic intervals among a subgroup of high-
and low-functioning children and adolescents with autism,
while Altgassen et al. (2005) found enhanced pitch mem-
ory for chord tones among a group of children with Asperger
syndrome. However, Bonnel et al. (2010) found no differ-
ences in short-term memory between a group of persons
with Asperger syndrome and a TD group aged 15–32 years
in their pitch discrimination of two tones less than a semi-
tone apart, while enhanced performance was reported for
an age-matched group with autism.
Enhanced memory for pitch among TD populations is
associated with differences in neural structure. For exam-
ple, evidence from event-related potential (ERP) studies
indicates that AP possessors do not update working mem-
ory (Zatorre, 2003). In addition, musicians with AP ability
have a leftward asymmetry of the brain, reflective of a
larger structural difference in the auditory association areas
located on the planum temporale (Schlaug et al., 1995).
Taken together, these findings suggest that AP possessors in
TD populations access pitch information through a differ-
ent neural pathway than non-AP possessors and point to
possible developmental differences in the maturation of
enhanced pitch memory.
Therefore, enhanced pitch memory in autism could
also be associated with developmental differences in neu-
ral structure. This idea is supported by ERP evidence of
left versus right hemisphere asymmetry during the pro-
cessing of pitch information among children with autism.
For example, 6-year-old children with autism showed
faster left hemisphere reactivity in comparison to TD
children matched for age when passively listening to
infrequent changes of pitch (Gomot et al., 2002).
However, 12-year-old children with ASD showed faster
right brain reactivity than age-matched comparison par-
ticipants when passively listening to three infrequent
tones (Gage et al., 2003). These studies indicate develop-
mental asymmetries of brain function that may predict
strong pitch memory in adulthood. Similarly, develop-
mental differences in neural structure emerge when com-
paring speech and pitch processing strategies between
children with autism and TD children (Lepistö et al.,
2005). Children with autism showed enhanced discrimi-
nation of both pure tones and vowel sounds as indicated
by enlarged ERPs. However, when children with autism
passively listen to speech-like sounds, there is less acti-
vation of the left speech-related areas, a reverse asym-
metry than that found among TD children (Boddaert
et al., 2004). Similarly, the left hemisphere language
areas are enlarged relative to the right among TD chil-
dren, but the reverse is often the case among children
with autism (Siegal and Blades, 2003). This asymmetry
may be attributed to developmental differences in the
maturation of pitch memory as areas in the right hemi-
sphere, such as Heschl’s gyrus and superior temporal
gyrus, play an important role in pitch computation
(Peretz, 2001; Zatorre, 1988). Such differences could
underlie a structural component to many of the auditory
difficulties in autism, such as increased sensitivity to
complex sounds (Khalfa et al., 2004; Samson et al.,
2006) and problems segregating speech sounds from
vowel sounds in certain situations (Alcantára et al.,
2004). Atypical development of pitch memory among
children with ASD may be a factor in processing complex
speech sounds, thereby contributing to both language
difficulties and enhanced pitch memory in autism.

Stanutz et al. 139
Enhanced memory for pitch information in autism is
thought to be analogous to abilities in the visual domain
with regard to attention to detail (Mottron and Belleville,
1993). Mottron and Belleville observed that the drawings
of one savant draftsman always grew from one unimpor-
tant detail. In contrast, a more common strategy among
TD persons is to make an outline and fill in the details.
The draftsman’s strategy, in which the details are taken
out of their context in order to recreate the big picture in a
reverse contextualizing process, is an example of weak
central coherence (WCC), and it is thought to be a cogni-
tive style that is specific to autism (Shah and Frith, 1993).
According to WCC theory, cognition in autism is driven
by the need to process detail at the expense of integrating
details within the large framework of context (Happé,
1999). AP ability has been viewed as an example of WCC
in that enhanced memory for individual pitch information
results from taking individual notes out of context apart
from the scales and melodies they form. WCC is also
thought to be a strategy used in tests of nonverbal reason-
ing ability, such as the block design task in which children
with autism typically show elevated performance. Similar
cognitive processes are likely accessed among persons
with autism when remembering pitch or visual informa-
tion, and this may also be the case for TD populations.
Scores on tests of nonverbal reasoning ability were
found to improve in TD groups after listening to music
(Rauscher et al., 1994), indicating possible symbiotic
cognitive processes between nonverbal reasoning and
pitch processing.
An alternative approach with regard to cognitive style in
autism is the enhanced perceptual functioning (EPF) model,
according to which, persons with autism show a preference
in processing detail over context, although the preference
does not necessarily create an imbalance between the two
levels of processing (Mottron 2005). However, studies that
indicate persons with autism use both detail and contextual
strategies when discriminating pitch have not been repli-
cated consistently. For example, Heaton (2005) found that
superior pitch discrimination of tone pairs in a group of
children with ASD did not predict superior detection of
small contextual changes in melody pairs. Furthermore,
Altgassen (2005) failed to replicate Heaton’s findings of
either superior long-term memory for individual pitch
(Heaton et al., 1998) or superior short-term memory for
pitch within the context of a chord (Heaton, 2003) in a
group of children with autism.
An underlying theme in the research literature concern-
ing pitch processing among persons with autism is the role
of enhanced memory for pitch information. Whether the
aim of the research was to discriminate between two pitches
as the same or different (Bonnel et al., 2003), to remember
individual pitches over the long term (Heaton et al., 1998b)
or within the context of a chord (Heaton 2003), to catego-
rize intervals (Heaton et al., 2008b), or to recognize an
altered note within a melody (Mottron, et al., 2000) enhanced
memory for pitch was demonstrated.
The purpose of the present study was to extend the pre-
vious research in a number of ways with regard to both
short- and long-term memory for pitch in autism and to
examine the relationship between pitch memory and other
areas of cognitive functioning. The research questions were
as follows:
1. Would the finding of superior short-term memory
for pitch in young adults and adolescence with
autism in their discrimination of tone pairs (Bonnel
et al., 2003) apply to school-aged children with
ASD?
2. Would children with ASD outperform TD children
in their discrimination of pitch within the context of
a melody when more demands on memory were
required? How would their performance fit with the
current theories of cognitive function in autism?
3. Would the previous finding in autism of superior
long-term memory of single tones (Heaton et al.,
1998) extend to long-term memory for melody?
Would children with ASD outperform TD children
in remembering a melody over a period of a week?
4. Would superior pitch memory in children with ASD
positively correlate with parental reports of their
child’s sensitivities to noise?
5. Is nonverbal reasoning ability among children with
autism connected to enhanced pitch memory? Are
the various measures of pitch memory correlated
with visual nonverbal reasoning ability among both
children with ASD and TD children?
Method
Participants
The participants included 25 children with ASD (6 girls)
ranging in age from 7 years 10 months to 13 years 2 months
(M = 10 years 8 months) and 25 TD children (14 girls)
ranging in age from 8 years 3 months to 12 years 6 months
(M = 10 years 5 months). The mean chronological ages of
the two groups were not significantly different from one
another (t = 0.628, df = 48, p > .05). Based on the Brief IQ
measure of the Leiter International Performance Scale–
Revised (Leiter-R) (Roid and Miller, 1997), the average
mean IQ of the children with ASD was 100.7 (standard
deviation (SD) = 18.3) and TD children was 106.4 (SD =
10.6). These mean scores did not differ from each other (t =
1.34, df = 48, p > .05).
All participants with ASD were diagnosed based on
Diagnostic and Statistical Manual of Mental Disorders
(4th ed.; DSM-IV, 1994) criteria for ASD by psychologists
in Ottawa, Canada. Two of the participants had been diag-
nosed using the Autism Diagnostic Observation Schedule.

140 Autism 18(2)
In all, 16 children were diagnosed at hospitals, 4 in schools,
and 5 in private practice. In all, 16 children were diagnosed
as having pervasive developmental disorder–not otherwise
specified (PDD-NOS), 7 were diagnosed with Asperger
syndrome, and 2 children had been diagnosed with autistic
disorder. None of these children had a dual diagnosis of
ASD and other developmental disorders. All of the children
with ASD had IQs above 70. Four of these children had
tested within the gifted range of intelligence as determined
by school psychologists. Of the four gifted children, three
were diagnosed with Asperger syndrome and one with
autistic disorder. The characteristics of the participants with
ASD are presented in Table 1.
In all, 19 of the children in the ASD group had under-
gone a standard audiometric procedure. All children who
underwent the procedure scored within the normal range of
hearing; 25 dB hearing level within the standard range of
frequencies (250–8000 Hz). Normal hearing in the remain-
der of the children with ASD and in the TD group was
ascertained by parental account. In all, 14 of the children
with ASD experienced aversions to loud sounds as reported
by the parent.
All but four of the participants in the study were musi-
cally untrained. One 10-year-old child with autistic disor-
der and one 10-year-old TD child had been taking piano
lessons for 1 year while one 13-year-old child with PDD-
NOS and one 12-year-old TD child had been involved in a
band program at school for a period of 1 year; thus, both the
TD and ASD groups were similar in their background of
limited musical training.
Procedure
Prior to commencing the research, the Research and Ethics
Board of Mcgill University approved the study and granted
a certificate of research involving humans. The parent of
each participant signed a letter of informed consent allow-
ing their child to participate in the study, and all the chil-
dren assented to participate. The study was conducted in
two sessions separated by 1 week. Prior to beginning the
first session, referred to as Music Game 1, the Brief IQ
measure of the Leiter-R (Roid and Miller, 1997) was
administered while the parent of each child filled out a
questionnaire concerning the specific diagnoses, birthdate,
and auditory sensitivities of their child. Music Game 1 con-
sisted of a paired single-tone pitch discrimination task, fol-
lowed by a melodic memory encoding task. A week later,
the child returned to complete the second session referred
to as Music Game 2, during which a melodic memory task
was presented first followed by a pitch discrimination task
within the context of melody. Each child achieved 70% on
a series of practice trials prior to commencing each music
game to ensure that each child understood the directives.
Apparatus
Microsoft PowerPoint 2004 for Macintosh was used to cre-
ate and present all tasks in the study. All narrations, single-
tone pairs, and both single melodies and melody pairs were
recorded using a Yamaha keyboard model P 90, in conjunc-
tion with Peak LE 5 recording software. All sounds were
recorded in stereo using a 44.1 kHz sampling rate and
32-bit resolution. They were saved as Audio Interchange
Format Files (AIFF) on an Apple MacBook computer. The
AIFF files then were imported into PowerPoint. The spo-
ken narrations of the music games were recorded using a
microphone attached to a set of Casonic model EP-790
headphones. The narration files were then modified by the
SoundSoap plug-in within Peak in order to eliminate ambi-
ent noise recorded during this process.
Measures
Brief IQ Leiter-R. The children completed the visualization
and reasoning battery Brief IQ measures of the Leiter-R
(Roid and Miller, 1997). The Leiter-R is a measure of non-
verbal fluid reasoning ability that is uninfluenced by educa-
tional, social, or family experience. The Leiter-R yields a
Table 1. Characteristics of participants with ASD.
Autistic disorder Asperger syndrome PDD-NOS
Participants 2 6 17
Gifted 1 2
Profoundly gifted 1
Music lessons 1 year 1 1
Normal hearing audiology 2 3 14
Normal hearing parental account 3 3
Hospital diagnoses 2 2 12
School psychologist diagnoses 3 1
Private practice diagnoses 1 4
ASD: autism spectrum disorders; PDD-NOS: pervasive developmental disorder–not otherwise specified; Gifted: performance is in the top 2% in
two out of three areas on measures of IQ; Profoundly gifted: performance is in the top 1% in two out of three areas on measures of IQ.

Stanutz et al. 141
nonverbal IQ measurement that correlates well with other
IQ instruments.
Music Game 1 design: pitch discrimination of tone pairs and
presentation of four melodies. For the pitch discrimination
task, children listened to two consecutive tones and then
indicated whether the tones were the same or different by
activating a blue button labeled “same” or gray button
labeled “different” on the computer screen. Once a choice
had been made, the next trial was activated. Four middle-
range piano tone pairs were presented randomly throughout
the 36-pitch discrimination trials: G3 (196.00 Hz), C4
(261.63 Hz), F4 (349.23 Hz), and A4 (440 Hz). Each tone
and its pair were presented a total of nine times. In three of
the presentations, the pairs of notes were identical in fre-
quency. The remaining pairs were altered so that the second
note of the pair was 25, 35, or 45 cents sharp or flat. Thus,
all pitch alterations were smaller than half of a semitone.
The tones were 1 s in duration, and they were separated by
1 s of silence. The participants chose to listen to sounds
either through a set of NexxTech headphones, model
3319154, or through the internal speakers of the computer.
Each participant was required to achieve a 70% success
rate in a series of practice trials before commencing the 36
trials of the study.
In the second half of Music Game 1, 4 two-measure
songs were presented (see Figure 1). Each song was paired
with a picture of a bird, fish, rabbit, or cat. The songs were
in 4/4 time. The bird, cat, and rabbit songs were in the major
keys of F#, Bb, and Ab, respectively, and the fish song was in
the key of c# minor. There were six blocks of four trials. The
four animal songs played randomly within each block. In
blocks 1, 3, and 5, the animal songs were presented paired
with their respective pictures. In the second block, the ani-
mal songs played without the presentation of the pictures. In
blocks 4 and 6, an animal song played while all four animal
pictures were presented. The child was instructed to click
the mouse on the picture of the animal “that liked the song
best.” After the child made a choice, the computer indicated
which animal song had played. In this way, the child
received feedback for each choice. Once the child accu-
rately identified three or more correct melodies in one block,
Music Game 1 terminated. If a child identified three or more
correct melodies by the fourth block, the researcher allowed
the child to finish blocks 5 and 6 and then the game was
terminated. If the participant had not achieved a 75% suc-
cess rate by the sixth block, the game was reset to blocks 5
and 6. The purpose of this section was for the participants to
memorize the melodies and to associate each melody with
its respective picture. No data was reported at this time.
Music Game 1 was approximately 40 min long.
Music Game 2 design: melodic memory and pitch discrimination
within a melodic context. A week after being administered
Music Game 1, each participant returned to complete the
second 40-min game, Music Game 2, which involved one
block of 16 melodic memory trials and one block of 36
melodic pitch discrimination trials. For the melodic memory
Figure 1. Animal songs: (a) bird song, (b) cat song, (c) rabbit song, and (d) fish song.

142 Autism 18(2)
trials, the participants listened to an animal song and then
clicked the mouse on the picture of the animal “that liked
the song best.” Feedback concerning the participants’ choice
was not given at this time nor was there exposure to the
melodies prior to the commencement of the 16 trials.
For the melodic pitch discrimination task, the partici-
pants were required to discriminate whether two melodies
were the same or different depending on whether the lead-
ing tone of the second melody was altered. The animal
melodies were altered and ordered as in the paired single-
tone pitch discrimination task in Music Game 1. Animal
melodies were paired with either the same or different ver-
sion of the same song. Each animal song pair was presented
a total of nine times. Three of the trials included identical
melody pairs, whereas in six of the pairs, the leading tone
in the second bar of the second melody was altered, so that
it was 25, 35, or 45 cents sharp or flat. The leading tone
preceded the tonic in each animal song, implying a strong
dominant tonic harmony. The order was randomized as in
the pitch discrimination task of Music Game 1. Each par-
ticipant responded by indicating whether the melodies were
the same or different. Before commencing these melodic
discrimination trials, each participant achieved a 70% suc-
cess rate on a series of practice trials. Paired melodies were
10 s in duration. The melody pairs began after 1 s of silence,
and 1.5 s of silence separated each melody of the pair.
Results
Analysis of variance: pitch discrimination
levels and contexts by diagnosis
A three-way mixed design analysis of variance was per-
formed to determine differences in pitch discrimination
between the children with ASD and the TD children. The
dependent variables were the means for the correct responses
out of 12 for each difficulty level (see Table 2). The between-
group factor was diagnosis (children with ASD and TD chil-
dren). Within-group factors were difficulty levels (25, 35, or
45 cents flat or sharp and same) and context (single-note
task vs melodic task). A main effect was found for diagnosis
(p < .03) and levels (p < .01). In addition, a significant inter-
action was found between levels and diagnosis (p < .04) and
between levels and context (p < .01).
The means of the children with ASD and the TD chil-
dren on each level (25, 35, or 45 cents sharp or flat and
same) and context (single-tone task and melodic task) were
subjected to a Fisher’s post hoc test for least significance in
a series of paired-sample t-tests. The t-tests indicated that
the performance means for the children with ASD were sig-
nificantly higher than for the TD children at the 25-cent (t =
−3.562, p = .05, df = 24) and 35-cent (t = −2.352, p = .05,
df = 24) levels in the melodic task (see Table 3) and
approached significance at the 45-cent level (t = −1.971, p
< .06, df = 24). The means of the children with ASD were
significantly higher at the 45-cent level in the single-tone
task (t = −2.091, p < .05, df = 24; see Table 2). The ASD
children outperformed the TD children in the 45-cent paired
single-tone discrimination task and the 25- and 35-cent
melodic discrimination task and were also more accurate at
every level except same when discriminating pitch in the
context of a melody (see Table 3).
Comparison of performance means: melodic
memory task
The data of two participants, one from each group, were
removed from the scores for melodic memory, as their
Table 2. Pitch discrimination in the single-tone task.
Group (n = 25) Same 25 cents 35 cents 45 cents
M SD M SD M SD M SD
ASD 10.88 1.62 7.20 2.97 9.30 2.29 10.68* 1.58
Typical 10.68 1.60 6.42 2.95 8.28 2.74 9.66 2.25
ASD: autism spectrum disorders; SD: standard deviation.
Maximum score = 12. The means reflect the correct response out of a maximum of 12. Twenty-five participants in each group.
*p < .05.
Table 3. Pitch discrimination in the melodic task.
Group (n = 25) Same 25 cents 35 cents 45 cents
M SD M SD M SD M SD
ASD 9.80 1.98 9.18* 2.38 9.66* 2.21 10.44 2.10
Typical 9.92 1.91 6.66 3.60 8.28 2.71 9.42 2.39
ASD: autism spectrum disorders; SD: standard deviation.
Maximum score = 12. Twenty-five participants in each group.
*p < .05.

Stanutz et al. 143
scores were below chance. An independent t-test deter-
mined that the group means (children with ASD: M = 12.25;
TD children: M = 10.17), excluding outliers, were statisti-
cally significant from each other (t = −2.26, p < .05, df =
24). A box plot comparing the spread of scores for the two
groups is shown in Figure 2. One of the TD children per-
formed at ceiling levels with a score of 16 correct (100%),
while eight of the children with ASD performed at or near
ceiling levels, five with scores of 16 correct (100%), and
three with scores of 15 correct (94%). Two of the children
with ASD who scored 100% were 7 and 8 years of age. The
7-year-old child was the youngest participant in the study.
Correlations between pitch discrimination
and melodic memory and nonverbal fluid
reasoning ability
Pearson product-moment correlations were calculated
between the overall scores from the Leiter-R Brief IQ and
the various experimental tasks. Positive correlations were
found between scores on the Leiter-R and scores on the
pitch discrimination tasks in the single-tone context (r =
.38, p < .001) and on the melodic memory task (r = .31, p <
.05) but not between the scores on the Leiter-R and those on
the pitch discrimination tasks in the melodic context.
Correlations were then calculated between the TD Leiter-R
Brief IQ score and the TD scores on the various experimen-
tal tasks. The same procedure was used for the ASD scores.
A positive correlation was found between the scores on the
Leiter-R Brief IQ and the single-tone discrimination task
for both the children with ASD (r = .48, p < .05) and the TD
children (r = .45, p < .05). In the melodic memory task,
however, only the scores of the ASD children correlated
significantly with the scores on the Leiter-R Brief IQ (r =
.40, p < .05), TD children (r =.34, p < .05).
Correlations between pitch discrimination,
melodic memory and developmental
sensitivity to sound in children with ASD
Spearman rank correlations were performed between
melodic memory, paired single-tone pitch discrimination,
melodic pitch discrimination, and developmental sensitiv-
ity to sound, as reported by parental account. Each child
was ranked with either having or not having an auditory
sensitivity to noise during their developmental history. No
statistically significant relationships were found.
Analysis of variance: discrimination of sharp
versus flat notes in single-tone pairs by
diagnosis
A three-way mixed design analysis of variance was per-
formed to determine differences in performance between
the children with ASD and the TD children on their dis-
crimination of flat versus sharp notes in the single-tone
context. The between-group factor was diagnosis (children
with ASD and TD children), and the within-group factors
were levels (25, 35, and 45 cents) and direction (flat vs
sharp). Main effects were found for levels (p < .01) and
Figure 2. Melodic memory performance. Melodic memory: the total number of correctly identified melodies.

144 Autism 18(2)
direction (p < .01). No main effect was found for diagnosis.
A significant interaction was found between levels by
direction (p < .05), indicating that both groups were better
able to discriminate sharpness than flatness at every level
(see Table 4). No other significant interactions were found.
Analysis of variance: discrimination of sharp
versus flat notes in the melodic task by
diagnosis
An analysis of variance identical in form to the one used for
single tones was used to examine sharp and flat sensitivities
in the melodic task. The between-group factor was diagno-
sis (children with ASD and TD children), and the within-
group factors were levels (25, 35, and 45 cents) and
direction (flat vs sharp). A main effect for diagnosis (p <
.01) was found, indicating statistically significant higher
scores for the children with ASD (see Table 5). A main
effect was found for levels (p < .01) but not for direction,
suggesting that children with ASD were more accurate than
TD children at discriminating whether a note was sharp or
flat at every level in the melodic context.
Discussion
In the single-tone context, children with ASD judged
whether two notes were different from one another in fre-
quency more accurately than TD children, and at the 45-cent
level, these results were statistically significant. This pitch
discrimination ability was stronger for children with ASD in
the melodic context where, at the 25- and 35-cent levels,
elevated performance was statistically significant. In the
melodic context, the task was to discriminate the same pitch
deviations on the leading tone within the context of 2 two-
bar melodies. The elevated scores of children with ASD on
the paired single-tone context at the 45-cent level was con-
sistent with the findings of Bonnel et al. (2003) who found
that high-functioning young adults and adolescents with
autism show enhanced pitch discrimination ability. The
enhanced scores of children with ASD on the pitch discrimi-
nation tasks in the melodic context of the study were some-
what surprising, as this would seem to be the more difficult
pitch discrimination task. The children needed to hold more
information in memory in order to be able to compare the
two melodies with each other, and they had to discriminate
the detail of the pitch deviation of one note, the leading tone,
within the whole context of a melody.
How do these results fit in with the primary theories of
cognitive function in autism? According to WCC theory
(Happé, 1999; Shah and Frith, 1993), superior discrimina-
tion in the single-tone task was expected. Those with
autism, according to this theory, have a tendency to under-
stand their world in terms of details, with limited ability to
define the details within the context of a whole. While
WCC may be useful in predicting the superior performance
of children with ASD in their discrimination of small dif-
ferences in pitch between two single tones, it fails to explain
their superior performance in discriminating pitch within
the context of the melody. Therefore, enhanced pitch dis-
crimination performance in the melodic task by the chil-
dren with ASD may be best explained by the EPF approach,
according to which, cognitive functioning in autism is
driven by an acute detection of small changes to the envi-
ronment. According to this theory, although there is a
Table 4. Discrimination of sharp versus flat notes in single-tone pairs.
Group (n = 25) 25 cents 35 cents 45 cents
S F S F S F
M SD M SD M SD M SD M SD M SD
ASD 2.68 1.31 2.12 1.01 3.36 0.81 2.76 1.05 3.68 0.56 3.44 0.82
Typical 2.36 1.32 1.92 1.15 3.32 0.90 2.24 1.23 3.40 0.70 3.00 1.00
ASD: autism spectrum disorders; SD: standard deviation; S: sharp, F: flat.
Maximum score = 4. Twenty-five participants in each group.
Table 5. Discrimination of sharp versus flat notes in the melodic task.
Group (n = 25) 25 cents 35 cents 45 cents
S F S F S F
M SD M SD M SD M SD M SD M SD
ASD 2.88 1.24 3.16 0.85 3.16 0.90 3.28 0.79 3.64 0.86 3.32 0.80
Typical 2.08 1.52 2.36 1.15 2.56 1.00 2.96 1.06 3.08 1.04 3.16 1.03
ASD: autism spectrum disorders; SD: standard deviation; S: sharp, F: flat.
Maximum score = 4. Twenty-five participants in each group.

Stanutz et al. 145
preference for processing detail in autism, this strategy
does not create an imbalance between understanding detail
in terms of context (Mottron et al., 2006).
Children with ASD more accurately noticed when the
leading tone was mistuned, indicating that they understood
the gestalt of the musical phrase. The leading tone was
used within the context of the melody to suggest the V-I
cadence, a prominent harmonic convention in Western cul-
ture. The V-I cadence denotes a finish to the musical
phrase. Their enhanced discrimination of the leading note
within the musical phrase suggests that children with ASD
had an expectation of how the leading tone should sound
within its melodic context. They recognized the pattern
within a larger structure, reinforcing the idea in EPF that
elevated pattern detection mechanisms are active in autism
(Mottron et al., 2009). In fact, the children with ASD were
better able to discriminate pitch within the context of a
melody than to discriminate the pitch of two single tones,
reflecting the advantage of context when discriminating
pitch. These results support the EPF approach, which con-
tends that contextual understanding in autism remains
intact. The EPF theory was also supported by the finding
that the performances of children with ASD and the TD
children did not differ when attempting to identify two
melodies as identical. This was surprising, as this task
would seem to be the easiest of the melodic discrimination
trials. Thus, perhaps recognizing when the two melodies
were exactly the same was more difficult for the children
with ASD because there was no change or difference to be
discriminated. Perception for ASD children may be height-
ened when small changes occur in the environment.
In the melodic memory task, the children with ASD were
better able than the TD children at remembering the two-
measure melodies that were paired with animal pictures
over a period of a week. The recall abilities of a number of
children with ASD was striking, especially considering that
they had heard each of the melodies only four to six times a
week earlier. Of the 25 children with ASD, 8 performed at
ceiling levels, including a 7-year-old boy who was the
youngest participant in the study. Another 8-year-old boy
who also performed with complete accuracy on this task
appeared to be barely paying attention. He found it difficult
to sit still, and his attention seemed to be darting from one
aspect of the room to another. Yet he was able to recall the
melodies instantly and automatically without effort. The
superior long-term melodic memory of these two boys and
of many other participants with ASD was surprising in light
of the problems reported with general memory systems in
autism (Volkmar et al., 2004) and the difficulty with mem-
ory where language learning is concerned. This points to the
possibility that superior memory for music in nonsavant
children with autism has commonalities with the musical
memory of savants (Miller, 1999; Sloboda et al., 1985).
Superior memory for music may be indicative of devel-
opmental differences among children with autism as
compared to TD children, revealing another aspect of their
cognitive style (Happé, 1999). Memory for music among
children with ASD may operate similar to the type of mem-
ory used by AP possessors for whom recognition of single
tones is instant and the updating of working memory is not
necessary (Zatorre, 2003). This process is unlike the longer
strategy used by relative pitch possessors, in which the
memory of a single memorized tone is recalled from long-
term memory in order to make a comparison between two
pitches. Similarly, reports of superior musical memory in a
musical savant with AP would suggest that superior mem-
ory for pitch in the context of melody and also harmony
may draw on similar mechanisms of AP strategy (Sloboda
et al., 1985). In TD populations, familiar overlearned melo-
dies are encoded in memory as a whole (Attneave and
Olsen, 1971; Davies and Jennings, 1976). Some children
with ASD in this study may have encoded melody in mem-
ory systems in a manner similar to that used by TD popula-
tions to encode overlearned melodies.
Generally, high-functioning persons with autism tend to
perform well on measures of nonverbal fluid reasoning
ability, such as the Leiter-R Brief IQ measure (Shah and
Frith, 1993). In order to complete the tasks, one must enlist
a highly sophisticated sense of recognizing and creating
patterns. Since the cognitive processes used for these types
of nonverbal reasoning tasks among persons with autism
have been thought to be analogous to cognitive processes
used for long-term pitch memory, and as the children with
ASD in this study demonstrated cognitive strengths in the
area of pitch memory, we attempted to determine whether
there was a relationship between the scores of the Leiter
Brief IQ measure and the musical tasks of the study.
Positive correlations were found between the combined
overall scores of fluid nonverbal reasoning ability as meas-
ured by Leiter-R performance and the combined overall
scores for pitch discrimination in the single-tone task and in
melodic memory both among TD children and children
with ASD. After separating the scores of the two groups,
the nonverbal reasoning ability scores remained positively
correlated with accuracy in the single-tone task for both
groups. However, nonverbal reasoning was positively cor-
related only with melodic memory ability in children with
ASD. A similar relationship had previously been discov-
ered. Rauscher and colleagues (Rauscher et al., 1994;
Rauscher and Zupanee, 2000) found that both music listen-
ing and music education improved performance on meas-
ures of nonverbal reasoning in both college students and
kindergarten students. Thus, the evidence from this study
and others suggests that processing musical information
and visual patterning ability may either be related to each
other or are affected by each other.
We compared children’s discrimination of flat versus
sharp notes in both the single-tone context and in the
melodic context. In the single-tone context, both groups
discriminated sharp notes more accurately than flat notes.

146 Autism 18(2)
However, in the melodic context, no differences were found
in accuracy when identifying sharp notes as opposed to flat
notes. Both groups of children were able to identify sharp
or flat notes with equal accuracy, although the children with
ASD were more accurate than the TD children at this task
in the context of melody. One possible explanation is that
when pitches are put into the context of a melody, the dis-
crimination of both sharp and flat notes may be easier than
when they are not part of a melody (Dowling, 1978;
Dowling and Fujitani, 1970). The relationship of the notes
to one another within a melody creates the tonal center or
key. Both groups of children may have found the single
tones more difficult to discriminate because they were pre-
sented out of a tonal context.
Memory for pitch by the children with ASD was
enhanced, regardless of whether the pitch to be discrimi-
nated was in or out of context. These results supported EPF
approach. Nevertheless, an important question remains:
why is memory for musical information preserved in
autism, whereas memory for language is often severely
impaired (Rutter et al., 2005)? One possible answer to this
question is that developmental differences in autism could
dictate a stronger memory for pitched sounds as compared
to speech sounds, if the enhanced pitch memory of nonsa-
vant children with autism is related to the AP ability of
musical savants. Saffran and Griepentrog (2001) demon-
strated that TD infants preferred AP cues to relative ones.
This preference disappeared, however, with increased age.
Saffran and Griepentrog concluded that AP ability might be
part of a developmental disinhibitory process (Bossomaier
and Snyder, 2004) in which relative pitch becomes favored
as language acquisition takes place. If children with autism
retain a level of AP ability, their preference and memory for
the pitched (vowel) sounds of words may interfere with
their memory of speech (consonant) sounds. This idea is
supported by evidence that many children with autism have
an aversion to spectrally complex sounds like noise
(Samson et al., 2006). Speech contains spectrally complex
sounds in the formation of consonants. Studies in backward
masking suggest that the noise created by consonants may
interfere with language comprehension (Marler et al.,
2002). Children with autism may have better memories for
pitches of vowels than for consonant sounds, and this may
affect their comprehension of language (Heaton et al.,
2008a; Järvenen-Pasley and Heaton, 2007).
Although developmental differences in memory for
pitch may affect language acquisition in autism, this ability
could be adaptive in some musical contexts. Pitch memory,
along with nonverbal reasoning ability, may eventually be
regarded as a general strength among persons with autism.
Understanding developmental strengths in autism will be
the first step in designing effective educational systems for
these children.
We propose five areas for further research following
from the findings of this study: (a) to further delineate to
what degree the development of pitch memory in autism dif-
fers from that of both TD populations and populations with
other developmental disabilities, (b) to determine to what
extent visual patterning ability in autism is related to or
coexists with enhanced pitch memory, (c) to define whether
enhanced pitch memory is evident across all forms of autism
and to what degree this strength contributes to the idea of
cognitive style, (d) to answer the question of whether
enhanced pitch memory affects language acquisition in
autism, and (e) to determine how music could be used as a
tool to educate children with autism. Finally, further research
into the relationship between autism, nonverbal reasoning
ability, and pitch memory can lead to important discoveries
regarding cognition among typical populations.
Funding
This research received no specific grant from any funding agency
in the public, commercial, or not-for-profit sectors.
References
Alcantára J, Weisblatt E, Moore B, et al. (2004) Speech-in-
noise perception in high functioning individuals with autism
or Asperger’s syndrome. Journal of Child Psychology and
Psychiatry 45(6): 1107–1114.
Altgassen M, Kliegel M and Williams T (2005) Pitch perception
in children with autistic spectrum disorders. British Journal of
Developmental Psychology 23: 543–558.
American Psychiatric Association (1994). Diagnostic and statis-
tical manual of mental disorders (4th ed.). Washington, DC:
Author.
Applebaum E, Egel A and Koegel R (1979) Measuring musi-
cal abilities of autistic children. Journal of Autism and
Developmental Disorders 9(3): 279–285.
Armstrong T and Darrow A (1999) Research on music and autism:
implications for music educators. Update-Applications of
Research in Music Education 18(1): 15–20.
Attneave F and Olsen RK (1971) Pitch as a medium: a new
approach to psychophysical scaling. American Journal of
Psychology 64: 147–166.
Baharloo S, Johnston PA, Service SK, et al. (1998) Absolute
pitch: an approach for identification of genetic and non-
genetic components. American Journal of Human Genetics
62: 224–231.
Bentley A (1966) Musical Ability in Children and Its Measurement.
New York: October House Inc.
Bermudez P and Zatorre R (2009) A distribution of absolute pitch
ability as revealed by computerized testing. Music Perception
27(2): 89–101.
Boddaert N, Chabane N, Belin P, et al. (2004) Perception of
complex sounds in autism: abnormal auditory cortical pro-
cessing in children. The American Journal of Psychiatry 161:
2117–2120.
Bonnel A, McAdams S, Smith B, et al. (2010) Enhanced pure-
tone pitch discrimination among persons with autism but not
Asperger syndrome. Neuropsychologia 48(2010): 2465–2475.
Bonnel A, Mottron L, Peretz I, et al. (2003) Enhanced pitch sen-
sitivity in individuals with autism: a signal detection analysis.
Journal of Cognitive Neuroscience 15(2): 226–235.

Stanutz et al. 147
Bossomaier T and Snyder A (2004) Is absolute pitch accessible
to everyone by turning off part of the brain? Organized Sound
9(2): 181–189.
Brown W, Cammuso K, Sachs H, et al. (2003) Autism-related
language, personality and cognition in people with absolute
pitch: results of a preliminary study. Journal of Autism and
Developmental Disorders 33: 163–167.
Davies JB and Jennings J (1976) Reproduction of familiar mel-
odies and the perception of tonal sequences. Journal of the
Acoustical Society of America 61(2): 534–541.
Dowling WJ (1978) Scale and contour: two components of a theory
of memory for melodies. Psychological Review 85(4): 341–354.
Dowling WJ and Fujitani DS (1970) Contour, interval, and pitch
recognition in memory for melodies. The Journal of the
Acoustical Society of America 49(2): 524–531.
Flowers P and Dunne-Sousa D (1990) Pitch pattern accuracy,
tonality, and vocal range in preschool children’s singing.
Journal of Research in Music Education 38(3): 102–114.
Gage NM, Siegel B, Callen M, et al. (2003) Cortical sound
processing in children with autism: an MEG investigation.
Auditory and Vestibular Systems 14(16): 2047–2051.
Gomot M, Giard M, Adrien J, et al. (2002) Hypersensitivity
to acoustic change in children with autism: electrophysi-
ological evidence of left frontal cortex dysfunctioning.
Psychophysiology 39: 577–584.
Happé F (1999) Autism: cognitive deficit or cognitive style?
Trends in Cognitive Sciences 3: 216–222.
Heaton P (2003) Pitch memory, labeling and disembedding in autism.
Journal of Child Psychology and Psychiatry 44(4): 543–551.
Heaton P (2005) Interval and contour processing in autism. Journal
of Autism and Developmental Disorders 35(6): 787–792.
Heaton P, Hermelin B and Pring L (1998) Autism and pitch
processing: a precursor for savant musical ability? Music
Perception 154: 291–305.
Heaton P, Hudry K, Ludlow A, et al. (2008a) Superior discrimina-
tion of speech pitch and its relationship to verbal ability in autism
spectrum disorders. Cognitive Neuropsychology 25(6): 771–782.
Heaton P, Williams K, Cummins O, et al. (2008b) Autism and
pitch processing splinter skills. Autism 12(1): 21–37.
Hermelin B (2001) Bright Splinters of the Mind: A Personal Story
of Research with Autistic Savants. London: J. Kingsley.
Järvenen-Pasley A and Heaton P (2007) Evidence for
reduced domain-specificity in auditory processing in
autism. Developmental Science 10(6): 786–793.
Khalfa S, Bruneau N, Rogé B, et al. (2004) Increased perception
of loudness in autism. Hearing Research 198: 87–92.
Lepistö T, Kujala T, Vanhala R, et al. (2005) The discriminat-
ing and orienting to speech and non-speech sounds in children
with autism. Brain Research 1066: 147–157.
Marler J, Champlin C and Gilliam R (2002) Auditory memory for
backward masking signals in children with language impair-
ment. Psychophysiology 39: 767–780.
Miller LK (1999) The savant syndrome: intellectual impairment
and exceptional skill. Psychological Bulletin 125: 31–46.
Moog H (1976) The Musical Experience of the Preschool Child
(trans C. Clarke). London: Schott Music.
Moorhead G and Pond D (1978) Music of Young Children. Santa
Barbara, CA: Pillsbury Foundation Studies.
Mottron, L., and Belleville, S. (1993). A study of perceptual anal-
ysis in a high level autistic subject with execeptual graphic
abilities. Brain and Cognition, 23, 279–309.
Mottron L, Dawson M and Soulières I (2010) Enhanced perception in
savant syndrome patterns, structure and creativity. Philosophical
Transactions of the Royal Society B 364: 1385–1391.
Mottron L, Dawson M, Soulières I, et al. (2006) Enhanced per-
ceptual functioning in autism: an update and eight principles
of autistic perception. Journal of Autism and Developmental
Disorders 36(1): 27–40.
Mottron L, Peretz I and Ménard E (2000) Local and global pro-
cessing of music in high functioning persons with autism:
beyond central coherence? Journal of Clinical Psychology
and Psychiatry 41(8): 1057–1065.
Peretz I (2001) Music perception and recognition. In: Rapp B
(ed.) The Handbook of Cognitive Neuropsychology. Hove:
Psychology Press, pp. 519–540.
Profita J and Bidder TG (1988) Perfect pitch. American Journal of
Medical Genetics 29: 763–771.
Rauscher F and Zupanee MA (2000) Classroom keyboard instruc-
tion improves kindergarten children’s spatial temporal per-
formance: a field experiment. Early Children’s Research
Quarterly 15(2): 215–228.
Rauscher F, Shaw G, Levine L, et al. (1994) Music and spatial
task performance: a causal relationship. In: Paper presented at
the 102nd annual convention of the American Psychological
Association August 13, Los Angeles.
Rimland, B., & Fein, D. (1988). Special Talents of Autistic
Savants. In L. K. Obler & D. Fein (Eds.), The exceptional
brain: The neuropsychology of talent and special abilities.
New York: The Guilford Press. 477–480.
Roid G and Miller L (1997) Leiter International Performance
Scale Revised. Wood Dale, IL: Stoelting.
Rutter M, Le Couteur A and Lord C (2005) Autism Diagnostic Interview-
Revised. Los Angeles, CA: Western Psychological Services.
Saffran J and Griepentrog G (2001) Absolute pitch in infant audi-
tory learning: evidence for developmental reorganization.
Developmental Psychology 37(1): 74–85.
Samson F, Mottron L, Jemel B, et al. (2006) Can spectro-temporal
complexity explain the autistic pattern of performance on audi-
tory tasks. Journal of Autism and Developmental Disorders
36(1): 65–76.
Schlaug G, Jancke L, Huang Y, et al. (1995) In vivo evidence of
structural brain asymmetry in musicians. Science 267: 699–701.
Shah A and Frith U (1993) Why do autistic individuals show
superior performance on the block design task? Journal of
Child Psychology and Psychiatry 34: 1351–1364.
Siegal M and Blades M (2003) Language and auditory processing
in autism. Trends in Cognitive Sciences 7(1): 378–380.
Sloboda J, Hermelin B and O’ Connor N (1985) An exceptional
musical memory. Music Perception 3: 155–170.
Thaut M (1988) Measuring musical responsiveness in autistic
children: a comparative analysis of improvised musical tone
sequences of autistic normal, and mentally retarded individuals.
Journal of Autism and Developmental Disorders 18(4): 561–571.
Volkmar FR, Lord C, Bailey A, et al. (2004) Autism and perva-
sive developmental disorders. Journal of Child Psychology
and Psychiatry 45: 135–187.
Zatorre RJ (1988) Pitch perception of complex tones and human
temporal-lobe function. Journal of the Acoustical Society of
America 84: 566–572.
Zatorre RJ (2003) Absolute pitch: a model for understanding the
influence of genes and development on neural and cognitive
function. Nature Neuroscience 6: 692–695.