Documenti di Didattica
Documenti di Professioni
Documenti di Cultura
Abstract
Objective: The purpose of the present study was to assess
how adolescents with autism who vary in the severity of autistic characteristics judge the emotional state of the speaker
when lexical and prosodic information is congruent or incongruent. Participants: Eighty participants, 24 autistic and
56 typically developing (TD) subjects participated: (a) 11 autistic adolescents between 9.5 and 16.83 years old, studying
at general education settings (AA1), (b) 13 autistic adolescents between 15.91 and 20.33 years old, studying at a special school (AA2), and (c) 56 TD subjects between 6 and 29
years old. Listeners were required to judge the emotional
meaning of words (sad/happy) in congruent conditions and
incongruent conditions. Results: (a) All participants judged
lexical and prosodic meaning separately with high accuracy,
(b) all participants showed prolonged reaction times in the
incongruent compared to the congruent condition, (c) AA1
relied on prosodic information in the incongruent condition
similarly to TD 915 year olds and TD adults, (d) AA2 and TD
68 year olds did not rely on prosodic information in the incongruent condition, and (e) both education placements,
the severity of autistic characteristics and nonverbal IQ con-
Introduction
Autism spectrum disorder (ASD) is a neurodevelopmental disorder expressed in early childhood. The core
deficits in ASD have been defined as a combination of
impairments in social interaction and communication
accompanied with fixated interests, and stereotyped or
repetitive behaviors [1]. Importantly, autism is a continuum disorder, ranging from severe autism to highfunctioning autism (HFA) with normal nonverbal ability
[2, 3].
Individuals with ASD have in many cases fluent speech
[4, 5], but their pragmatic-semantic use and understanding of language in social contexts is impaired. For example, individuals with ASD showed difficulties in turn taking during conversation [6], in taking into consideration
the perspective of another person [7] and in adequately
Osnat Segal
7b Zelig Bas
Petach Tiqwa (Israel)
E-Mail segalll@netvision.net.il
Downloaded by:
Dir.General de Bibliotecas, UNAM
132.248.9.8 - 1/26/2015 4:50:51 PM
Key Words
Emotions Speech stimuli Autism spectrum
26
Downloaded by:
Dir.General de Bibliotecas, UNAM
132.248.9.8 - 1/26/2015 4:50:51 PM
responding to what another person has said [6]. Difficulties in understanding nonliteral language such as irony
[4], figurative expressions and metaphors [8, 9] have also
been reported, suggesting a tendency for literal understanding of language without taking into consideration
the real intention of the speaker and the context. Furthermore, individuals with ASD showed difficulties drawing
appropriate inferences especially in multiple-cue environments [8, 10]. Children with ASD may understand,
for example, the meaning of mental verbs but failed to
infer what these verbs imply in situation [8]. They also
understood the feeling of the character in a story based on
prosody alone, but failed to judge the feeling of this character based on prosody when there is a discrepancy between positive prosody and the context [10]. Finally, individuals with ASD often displayed difficulties in the production and comprehension of speech prosody which is
known to carry the pragmatic information concerning
the intention of the speaker [1118].
Pragmatic impairments in ASD are usually explained
by the inability to understand the state of mind of other
people in the interaction [19, 20]. One of the major cues
for understanding the state of mind of others is related to
the ability to interpret correctly the prosody on which
their words are carried. Prosody refers to the suprasegmental aspects of speech, including variations in fundamental frequency (F0), intensity and duration [21]. These
prosodic characteristics carry a variety of communicative
meanings including grammatical, pragmatic and affective
purposes [22, 23]. Specifically, affective prosody includes
suprasegmental cues that are used to signal the mental
state and feeling of the speaker as well as the changes in
register used for different social functions [2427].
The ability of autistic children, adolescents and adults
to perceive and process affective prosody has been mainly studied in the perspective of the theory of mind (ToM)
hypothesis [19]. The ToM suggests that the basic deficit
of autism involves a cognitive deficit in inferring the mental states of others. In agreement with the ToM hypothesis, studies suggest that individuals with HFA have difficulties in recognizing complex emotions that involve
advanced understanding of social and interpersonal situations from prosodic information. Rutherford et al. [28],
for example, investigated the ability of adults with HFA
or Aspergers syndrome (AS) to understand vocally expressed complex emotions (e.g. suspicious, worried, nervous) presented in spoken phrases. Their results showed
that participants with HFA and AS were impaired in this
task compared to typically developing (TD) adults. Golan
et al. [29] used revised tasks from Rutherford et al. [28] to
27
Downloaded by:
Dir.General de Bibliotecas, UNAM
132.248.9.8 - 1/26/2015 4:50:51 PM
Characteristics
AA1 (n = 11;
9 males, 2 females)
Age, years
AQ scores
Raven scores in percentiles
for age norms
Raven standard scores
range
mean
SD
12.75
29.64
63.90
110.45
45.85
34.96
0.1 99.9
55 145
AA2 (n = 13;
10 males, 3 females)
Significance
mean
SD
range
16.95
36
1.64
5.99
15.91 20.33
28 45
54.79
98.07
42.47
32.63
0.1 99.9
55 145
Stimuli
Stimuli consisted of a total of 40 words: 15 words associated
with happy meaning, 15 words associated with sad meaning and
10 neutral words with no emotional intent (see Appendix 1 for the
list of words). The words were judged on their association with the
meaning sad/happy/neutral with separate scaling. Eight children
with a mean age of 7.7 years (SD = 1.17) and 15 adults with a mean
age of 29.8 years (SD = 9.96) judged the words. Sad and happy
words chosen for the test from a total of 69 words scored above 5
points on a scale from 1 to 7 by both adults and children (mean =
6.06, 6.29, SD = 0.32, 0.37, for sad and happy words, respectively)
suggesting that all judges graded them as highly associated with
either sad or happy meaning. The mean scores for sad and happy
judgment did not differ significantly (p > 0.05). Neutral words
were scored between 1 and 3 points on a 17 scale suggesting that
all judges graded them as not associated with either sad or happy
meaning (mean = 2.36, SD = 0.27). Additionally, words were
judged by adults on a 17 familiarity scale and were perceived as
familiar (mean = 4.98, 5.45, 4.72, SD = 0.66, 1.1, 0.98, for sad, happy and neutral words, respectively). The words are shown in Appendix 1. Note that words with happy and sad meaning were
matched by number of syllables.
Recording of Stimuli
Stimuli were produced and recorded by a female, native Hebrew speaker in a soundproof room via a JVC MV 40 microphone
using Sound Forge software (version 4.5a), at a sampling rate of
44,100 Hz and with 16-bit quantization. Ten sad/happy words (5
28
words with happy meaning and 5 words with sad meaning) were
recorded with neutral prosody, 10 neutral words were recorded
with happy/sad prosody (5 words with happy prosody and 5 words
with sad prosody). The additional 20 words (10 words with happy
meaning and 10 words with sad meaning) were recorded twice:
once with happy prosody and once with sad prosody, resulting in
20 congruent stimuli (sad or happy meaning with congruent sad
or happy prosody, respectively), and 20 incongruent stimuli (sad
or happy meaning with incongruent happy or sad prosody, respectively). All words were normalized for intensity without changing
the intensity ratios between syllables within the words.
Since F0 is the main acoustic parameter reflecting emotion [52],
we analyzed F0 parameters in each stimulus (word). F0 was extracted by autocorrelation using the speech analysis software
PRAAT. F0 mean, F0 minimum and F0 maximum were computed
for each stimulus. The results are shown in Appendix 2. Note that
words with the same prosody but opposing lexical meaning did not
differ in the measurements of F0 (p > 0.05). This suggests that the
speaker was able to control for the prosody in production regarding the lexical meaning of the word. Overall, test items with happy
prosody had an increased F0 range (mean = 236.88, SD = 119.40)
compared to test items with sad prosody (mean = 146.78, SD =
54.03; t19 = 3.42, p = 0.003).
In order to confirm that the prosody of words was perceived by
listeners as sad, happy or neutral, the words were low-pass filtered
with a 100- to 500-Hz low-pass band with a slope of 96 dB per octave. Six additional adult listeners listened to the low-pass filtered
stimuli and judged them as happy, sad or neutral based on prosody alone. All judges confirmed for all words the intended prosody
of the recorded word.
Baseline and Testing Material
The recorded words were divided into two baselines and the
test materials as shown in Appendix 1. Each baseline was aimed to
assess the ability of the participant to recognize the lexical meaning
or prosody of the presented word while the other dimension (lexical meaning or prosody) was neutral. The lexical baseline included 10 words with neutral prosody and lexical meaning that was
associated with sad/happy feeling (5 words with happy meaning
and 5 words with sad meaning). The mean duration of the happy
(mean = 576 ms, SD = 56) and sad words (mean = 627 ms, SD =
66) was not significantly different (p > 0.05). The prosodic baseline
Segal/Kaplan/Patael/Kishon-Rabin
Downloaded by:
Dir.General de Bibliotecas, UNAM
132.248.9.8 - 1/26/2015 4:50:51 PM
**
**
100
90
Correct response (%)
80
70
60
50
40
30
20
10
0
68
915
Years
2629
AA1
AA2
Procedure
Participants were tested individually in a quiet room using a
Dell Precision 4300 computer. The stimuli were presented with the
Superlab 4.5 software. The participants sat facing the monitor, and
the experimenter sat beside the participant. Each participant heard
first the baseline lists and then the test word lists. The 10 stimuli of
each baseline were presented randomly. The 40 stimuli of the test
phase were presented twice in random order.
Participants were told that they were going to play a listening
game. In the lexical baseline the instructions were: My friend,
Danna will say words and youll decide whether the word is connected to something sad or happy. If you choose happy, press the
key with the happy face, if you choose sad, press the key with the
sad face. In the prosodic baseline the instructions were: My friend,
Danna will say words and youll decide whether her voice is sad or
happy. If you choose happy, press the key with the happy face, if
you choose sad, press the key with the sad face. In the test phase
the instructions were: My friend, Danna will say words and youll
decide whether she feels sad or happy. If you choose happy, press
the key with the happy face, if you choose sad, press the key with
the sad face.
The instructions were written on the computers monitor, and
the experimenter read them to the participant out loud and gave
an example in words. Before the beginning of both baselines and
the test, 2 examples were provided through the computer in order
to confirm that the participant understood the task and recognized
the buttons for sad/happy. In the lexical baseline, the examples included 2 words with sad and happy meaning and neutral prosody.
In the prosodic baseline, the examples included 2 words with sad
and happy prosody and neutral meaning. In the test phase, the examples included lexical-prosodic congruent and incongruent
stimuli. Feedback was provided through the computer and by the
experimenter. After the 2 examples, the experimenter confirmed
that the participant understood the task. All participants understood the task within the 2 examples.
The software written for this experiment controlled for presentation of stimuli and collection of responses and reaction
times. Reaction times were defined as the time in milliseconds
from the end of the stimulus to the entry of the keyed response
(happy or sad). The stimuli were presented through the computer at a comfortable level of 65 dB HL. The distance from the
Baselines
The average percentages of correct responses for each
baseline across the 5 groups of participants are summarized in figure 1.
Overall, all participants performed with high recognition scores on both baselines although a drop in performance can be seen for AA2. Repeated measures analysis
of variance following arcsine transformation with correct response scores in each baseline as the dependent
variable and type of baseline and group of participants as
the independent variable was conducted on the data. The
statistical analysis revealed no interaction between type
of baseline and group (p > 0.05) and no main effect for
type of baseline (scores for baseline 1 or 2; p > 0.05). A
main effect was found for group (F4, 78 = 5.09, p = 0.001).
TD participants
Autistics
Fig. 1. Mean lexical and prosodic judgment scores (in percent) and
standard errors for baseline stimuli for the 3 groups of TD participants, 68, 915 and 2629 year olds, and the 2 groups of autistic adolescents who study at regular education placement (AA1)
and at special education placement (AA2). *p < 0.05, **p < 0.01.
loudspeaker was 50 cm (19.5 inches). The procedure was approved by the ethical committees of Tel Aviv University and the
Ministry of Education.
Results
29
Downloaded by:
Dir.General de Bibliotecas, UNAM
132.248.9.8 - 1/26/2015 4:50:51 PM
**
90
80
80
70
70
60
50
40
10
0
TD participants
AA1
AA2
68
915
Years
TD participants
2629
AA1
AA2
Autistics
Autistics
Fig. 2. Mean lexical and prosodic judgment scores (in percent) and
standard errors for congruent stimuli for the 3 groups of TD participants, 68, 915 and 2629 year olds, and the 2 groups of autistic adolescents who study at regular education placement (AA1)
and at special education placement (AA2). *p < 0.05, **p < 0.01.
Further multiple comparisons showed a significant difference between the group of adolescents with ASD attending a special school for ASD (AA2) and the adolescents with ASD attending a regular school (AA1; t78 =
3.04, p = 0.03). Also, a significant difference was found
between the group of adolescents with ASD attending
special school (AA2) and TD children aged 915 years
(t78 = 3.34, p = 0.01) as well as with TD adults (t78 = 3.63,
p = 0.005). No differences were found between the 2
groups of TD children and adults, as well as between adolescents with ASD (AA1) and TD children and adults
(p > 0.05).
Congruent and Incongruent Tests
The groups mean percent correct scores for the congruent and incongruent lexical-prosodic words are shown
in figures 2 and 3, respectively.
Figure 2 shows that all 5 groups of participants received high percent correct scores in the congruent lexical
prosodic condition. One-way analysis of variance with
correct judgments as the dependent variable and group of
participants as the independent variable revealed a main
effect for group (F4, 78 = 3.52, p = 0.01). Multiple comparisons (Ryan-Einot-Gabriel-Welsch) suggested a sig30
30
10
2629
40
20
915
Years
**
50
20
68
**
60
30
Lexical judgment
Prosodic judgment
90
Response (%)
100
100
Fig. 3. Mean lexical and prosodic judgment scores (in percent) and
standard errors for incongruent stimuli for the 3 groups of TD participants, 68, 915 and 2629 year olds, and the 2 groups of autistic adolescents who study at regular education placement (AA1)
and at special education placement (AA2). *p < 0.05, **p < 0.01.
Downloaded by:
Dir.General de Bibliotecas, UNAM
132.248.9.8 - 1/26/2015 4:50:51 PM
**
2,500
Congruent
Incongruent
2,000
1,500
1,000
500
0
68
915
Years
TD participants
2629
AA1
AA2
Autistics
31
Downloaded by:
Dir.General de Bibliotecas, UNAM
132.248.9.8 - 1/26/2015 4:50:51 PM
One-way analysis of variance with the difference between lexical and prosodic judgments as the dependent
variable and group of participants as the independent
variable confirmed that the difference between the 5
groups of participants was statistically significant (F4, 78 =
3.60, p = 0.009). Further multiple comparisons (RyanEinot-Gabriel-Welsch) showed that the difference between prosodic and lexical judgments was significantly
increased in TD adults compared to 6- to 8-year-old TD
children (p < 0.05). No other significant differences were
found between the 5 groups.
The mean reaction time in milliseconds for each condition (congruent and incongruent) and group were calculated and shown for the congruent and incongruent
stimuli in figure 4. The figure shows longer reaction times
for the young TD children compared to the other groups
and for the incongruent stimuli compared to the congruent ones.
Two-way analyses with repeated measures (group,
type of stimuli: congruent/incongruent) revealed a main
effect for type (F1, 78 = 22.53, p = 0.0001). In general, reaction time for incongruent stimuli was significantly longer
(mean = 1,252.09 ms, SD = 801) compared to the congruent condition (mean = 1,082.59 ms, SD = 851). Also, a
main effect for group was found (F4, 78 = 25.25, p < 0.0001).
Multiple contrasts revealed that the group of young
TD children (68 years old) had longer reaction times for
both the congruent and incongruent stimuli compared to
the other groups of participants (t78 = 10.15, 11.54, 9.32,
8.58, p = 0.001, for the comparison to TD 925 years old,
TD adults, AA2 and AA1, respectively). No differences in
reaction time were found between TD 915 years old and
adults and between these two groups and the 2 groups of
adolescents with ASD (p > 0.05). No group type interaction was found (p > 0.05).
Correlations between the AQ scores, the Raven Progressive Matrices scores (for age norms) and the judgments by prosody in the incongruent condition were conducted for the 2 groups of participants with ASD pooled
together. Among the possible correlations, the only significant correlation was found between the Raven scores
and prosodic judgment (r = 0.49, 0.51, p = 0.01, for Raven
results in percentiles and standard scores, respectively).
Thus, the results of the Raven test can explain about 26%
of the variance in prosodic judgment in adolescents with
ASD. No correlations were found between age and judgment by prosody in each one of the ASD groups apart or
when pooled together.
To provide better understanding of the factors that
contributed to prosodic preference in adolescents with
Variable
Step 1
Educational
placement
AQ
Nonverbal IQ
Adjusted R2
F
Step 2
2
0.26 7% 1.28
7%
1.63
r2
0.31 6%
1.34
0.24 3%
1.00
0.41 14%
2.07*
23%
3.23*
autism, multivariate linear regression analysis was computed. Specifically, the regression model included 2 steps.
First, the covariate of school placement as a dummy variable was entered into step 1. In step 2, both nonverbal IQ
(Raven) and AQ scores were added to the models as predictors. Table 2 presents the results for these analyses.
The variables included in the regression predicted a significant proportion of the variability in prosodic preference, with R2 = 0.33, F3, 20 = 3.23, p < 0.05. As shown in
table2, only nonverbal IQ was a significant predictor of
prosodic preference after controlling for school placement variables ( = 0.41, p = 0.05).
Discussion
tal trends exist for processing both types of stimuli. Furthermore, the evidence demonstrated in the present study
for a developmental change in judging emotions by prosody provides the platform for further interpretation of
the results of the participants with ASD.
Adolescents with ASD were able to accurately perceive
the emotions of the speaker based on lexical or prosodic
information alone. Moreover, when provided simultaneously, in the incongruent condition, both types of information, lexical and prosodic, were processed as evident
by the prolonged reaction time. However, the ability to
give more weight to the prosodic over the lexical information was reduced in the group with severer autistic characteristics studying at a special education placement
(AA2). The present findings will be discussed and elaborated in the remainder of the discussion.
One main finding of the present study suggests that
there is a difference in the way by which the 2 groups of
adolescents with ASD judged the emotional intent of the
speaker in incongruent lexical-prosodic stimuli. Adolescents with ASD who studied at regular schools with less
autistic characteristics (AA1) gave priority to prosodic
over lexical information, whereas adolescents with ASD
who studied at special schools with severer autistic characteristics (AA2) did not show such a clear preference.
This finding is further strengthened by the fact that the
AA1 participants were younger than the AA2 participants but nonetheless showed better performance. The
only study, to our knowledge, that is somewhat comparable to our study is that by Stewart et al. [36]. They
showed reduced ability to rely on prosody for emotional
judgment in HFA adults. In contrast, our study showed
that HFA adolescents at regular schools relied on prosody. The discrepancy between the two studies can be attributed primarily to the differences in stimuli [36]. Stewart et al. [36] used complex emotions in sentences whereas we chose familiar words with 2 basic emotions in the
present study. This allowed us to test adolescents with
ASD with severer autistic characteristics. Indeed, we have
managed to obtain judgment in incongruent conditions
in a group of adolescents with ASD that has not been tested so far with this type of paradigm.
The finding of reduced ability to judge the emotional
state of the speaker by prosody (lexical bias) in the group
of adolescents with ASD who study at special education
placement (AA2) might be clarified by either a domainspecific explanation (ToM) or by a more domain-general
explanation (executive dysfunctions) [54, 55]. The ToM
theory predicts limited ability to recognize and understand the emotional state of the speaker in individuals
Segal/Kaplan/Patael/Kishon-Rabin
Downloaded by:
Dir.General de Bibliotecas, UNAM
132.248.9.8 - 1/26/2015 4:50:51 PM
Table 2. Summary of the linear regression model predicting prosodic preference (n = 24)
33
Downloaded by:
Dir.General de Bibliotecas, UNAM
132.248.9.8 - 1/26/2015 4:50:51 PM
with ASD because of a basic deficit in the ability to understand emotions, intents and mental states of others [29].
According to this view, the reduced ability observed in
AA2 to accurately judge the intent of the speaker by prosody reflects a specific insufficiency of the neurocomputational system that is associated with ToM. In this view, it
is also possible that young TD children show a lexical bias
in incongruent conditions because of less mature ToM
abilities.
The executive function explanation suggests that individuals with ASD have difficulties in general cognitive
domains including planning, reasoning, self-monitoring,
flexibility of thought and action, holding mental representations in working memory and inhibition [55, 56].
According to this view, difficulties in executive control
and specifically difficulties in inhibition of lexical information may explain the lexical bias of both autistic adolescents (AA2) as well as young TD children [53]. Furthermore, the association found in the present study between nonverbal reasoning (Raven test) and the tendency
for prosodic judgments in adolescents with ASD supports
the notion that domain-general abilities play a role in understanding emotions in incongruent conditions. It is
possible that preference for prosodic judgment in incongruent conditions requires abstract reasoning and problem-solving strategies that are less mature in some adolescents with ASD. Interestingly, cognitive ability played
a role in understanding emotions in incongruent conditions, in spite of the fact that only two basic emotions
were utilized in single words that reduced the burden on
working memory.
Another explanation for the difficulties in judging incongruent lexical-prosodic stimuli by AA2 is related to
conceptual understanding of mixed emotions (e.g. I felt
happy and excited/sad). TD children are able to recognize and understand mixed emotions in themselves and
in others only by late childhood [57, 58]. It is possible that
both AA2 as well as young TD children in the present
study were less advanced in judging emotions in words
that carry simultaneous and conflict emotional meanings
because of less mature conceptual understanding of
mixed emotions in general. Further research, however, is
needed in order to shed light on this issue.
Importantly, however, the present findings suggest
that the ability of adolescents with ASD to give priority to
prosodic over lexical information in incongruent lexicalprosodic conditions is not always impaired. Nevertheless,
it might be influenced by the cognitive abilities of the individual. One cannot exclude the possibility, however,
that studying in a regular school has allowed autistic in-
emotions in both congruent and incongruent lexicalprosodic conditions, and that a quick response from the
listener is required in order to continue the discourse, it
might be the case that individuals with ASD would have
more difficulties in understanding the real intent (or
emotional state) of the speaker [29, 34]. However, further
studies are needed in order to investigate this issue.
From a clinical point of view, it is highly important
that speech and language pathologists will include emotional judgment in congruent and incongruent lexicalprosodic stimuli as part of both their assessment and
treatment of individuals with ASD. Incorporating training of emotional judgment by prosody within the work of
speech and language pathologists who work with autistic
children and adolescents may improve their understanding of others in everyday social interactions.
Acknowledgments
The authors gratefully acknowledge the ASD students and the
staff at Gil school for their support and participation in this study.
We would also like to thank Adi Luber, Einat Beeri and Rommi
Gan undergraduate and graduate students at the Department of
Communication Disorders, Tel Aviv University, for their assistance with stimulus preparation and data collection, and Mrs. Esther Shabtai for assistance with the statistical analysis. This study
was supported by the Fanny Fannister Award from the Sackler
Faculty of Medicine, Tel Aviv University.
Appendix 1
Test part
Test part
ahav (love)
matan (present)
bdix (joke)
Ser (happiness)
xijx (smile)
kf (fun)
tsxk (laugh)
alz (cheerful)
leitsn (clown)
rikd (dance)
xfeS (freedom)
glda (ice cream)
maksm (charming)
nifl (wonderful)
xibk (hug)
gara (bad)
kiSaln (failure)
pred (separation)
bxi (crying)
tsar (sorrow)
r (bad)
rv (quarrel)
dim (tear)
kev (pain)
yeS (despair)
hefsd (loss)
neS (punishment)
kas (anger)
ptsa (wound)
xol (sick)
baseline 1
baseline 1
baseline 1
baseline 1
baseline 1
test
test
test
test
test
test
test
test
test
test
baseline 2
baseline 2
baseline 2
baseline 2
baseline 2
baseline 2
baseline 2
baseline 2
baseline 2
baseline 2
baseline 2
34
Segal/Kaplan/Patael/Kishon-Rabin
Downloaded by:
Dir.General de Bibliotecas, UNAM
132.248.9.8 - 1/26/2015 4:50:51 PM
Appendix 2
Measurements of F0, minimum F0 and maximum F0
F0 mean
F0 min.
F0 max.
Range
Baseline 1:
words with neutral
prosody
(n = 10)
Baseline 2:
words with happy
prosody
(n = 5)
Baseline 2:
words with
sad prosody
(n = 5)
Test items:
words with happy
meaning and
happy prosody
(n = 10)
Test items:
words with sad
meaning and happy
prosody
(n = 10)
Test items:
words with sad
meaning and
happy prosody
(n = 10)
Test items:
words with
happy meaning
and sad prosody
(n = 10)
166.96
(10.70)
140.92
(15.94)
197.79
(10.80)
56.87
(19.11)
226.44
(15.53)
159.75
(6.98)
312.34
(36.72)
152.59
(40.55)
251.62
(20.60)
195.78
(35.13)
318.53
(37.38)
122.75
(61.57)
271.71
(54.19)
179.24
(32.72)
386.66
(98.87)
207.42
(101.40)
277.47
(67.98)
159.74
(32.33)
426.07
(137.94)
266.34
(133.73)
255.23
(21.28)
188.36
(40.87)
340.07
(37.79)
151.71
(59.85)
231.14
(60.50)
161.32
(48.67)
303.18
(52.21)
141.86
(50.28)
Range = (maximum F0 minimum F0); n = number of words. Measurements in hertz in the stimuli of baseline 1 (lexical judgment),
baseline 2 (prosodic items) and test items. Standard errors are shown in parentheses.
References
11
12
13
14
15
16
17
18
35
Downloaded by:
Dir.General de Bibliotecas, UNAM
132.248.9.8 - 1/26/2015 4:50:51 PM
1 American Psychiatric Association: Diagnostic and Statistical Manual of Mental Disorders, ed 5. Washington, American Psychiatric
Publishing, 2013.
2 Wing L, Gould J: Severe impairments of social
interaction and associated abnormalities in
children: epidemiology and classification. J
Autism Dev Disord 1979;9:1129.
3 Gillberg C: Asperger syndrome and highfunctioning autism. Br J Psychiatry 1998;172:
200209.
4 Martin I, McDonald S: An exploration of
causes of non-literal language problems in individuals with Asperger syndrome. J Autism
Dev Disord 2004;34:311328.
5 Saalasti S, Lepist T, Toppila E, Kujala T,
Laakso M, Nieminen- von Wendt T, et al:
Language abilities of children with Asperger
syndrome. J Autism Dev Disord 2008; 38:
15741580.
6 Capps L, Kehres J, Sigman M: Conversational
abilities among children with autism and children with developmental delays. Autism
1998;2:325344.
7 Loveland KA, Tunali B, McEvoy RE, Kelley
ML: Referential communication and response
adequacy in autism and Downs syndrome.
Appl Psycholinguist 1989;10:301313.
8 Dennis M, Lazenby AL, Lockyer L: Inferential
language in high-function children with autism. J Autism Dev Disord 2001;31:4754.
9 Gold R, Faust M, Goldstein A: Semantic integration during metaphor comprehension in
Asperger syndrome. Brain Lang 2010; 113:
124134.
10 Le Sourn-Bissaoui S, Aguert M, Girard P,
Chevreuil C, Laval V: Emotional speech comprehension in children and adolescents with
36
38 Morton JB, Trehub S: Childrens understanding of emotion in speech, Child Dev 2001;72:
834843.
39 Morton JB, Trehub S, Zelazo PD: Sources of
inflexibility in 6-year-olds understanding of
emotion in speech. Child Dev 2003;74:1857
1868.
40 Morton JB, Trehub S: Childrens judgements
of emotion in song. Psychol Music 2007; 35:
629639.
41 Friend M, Bryant JB: A developmental lexical
bias in the interpretation of discrepant messages. Merrill-Palmer Q 2000;46:342369.
42 Henkin Y, Yaar-Soffer Y, Gilat S, Muchnik C:
Auditory conflict processing: behavioral and
electrophysiologic manifestations of the
Stroop effect. J Am Acad Audiol 2010;21:474
486.
43 Schweickert R: A critical path generalization
of the additive factor method: analysis of a
Stroop task. J Math Psychol 1978;18:105139.
44 Segall MJ, Campbell JM: Factors influencing
the educational placement of students with
autism spectrum disorders. Res Autism
Spectr Disord 2014;8:3143.
45 American Psychiatric Association: Diagnostic and Statistical Manual of Mental Disorders, ed 4. Washington, American Psychiatric
Association, 1994.
46 Baron-Cohen S, Wheelwright S, Skinner R,
Martin J, Clubley E: The autism spectrum
quotient (AQ): evidence from Asperger syndrome/high functioning autism, males and
females, scientists and mathematicians. J Autism Dev Disord 2001;3:517.
47 Kurita H, Koyama T, Osada H: Autism spectrum quotient Japanese version and its short
forms for screening normally intelligent persons with pervasive developmental disorders.
Psychiatry Clin Neurosci 2005;59:490496.
48 Woodbury-Smith MR, Robinson J, Wheelwright S, Baron-Cohen S: Screening adults for
Asperger syndrome using the AQ: a preliminary study of its diagnostic validity in clinical
practice. J Autism Dev Disord 2005; 35: 331
335.
Segal/Kaplan/Patael/Kishon-Rabin
Downloaded by:
Dir.General de Bibliotecas, UNAM
132.248.9.8 - 1/26/2015 4:50:51 PM