behavioral

sciences
Article
The Cognitive Profile of People with
High-Functioning Autism Spectrum Disorders
Atusa Rabiee 1, * , Sayyed Ali Samadi 2 , Behnoosh Vasaghi-Gharamaleki 3 , Soode Hosseini 4 ,
Saba Seyedin 5 , Mohammadreza Keyhani 3 , Ameneh Mahmoodizadeh 6 and
Fatemeh Ranjbar Kermani 7
1 Department of Speech and Language Pathology, School of Rehabilitation Sciences, Iran University of
Medical Sciences, Tehran 15459-13487, Iran
2 Institute of Nursing and Health Research, University of Ulster, Northern Ireland BT37 0QB, UK;
s.samadi@ulster.ac.uk
3 Department of Basic Sciences, School of Rehabilitation Sciences, Iran University of Medical Sciences,
Tehran 15459-13487, Iran; vasaghi.b@iums.ac.ir (B.V.-G.); more_keyhani@yahoo.com (M.K.)
4 Department of Psychology, School of Education and Psychology, Alzahra University,
Tehran 1993893973, Iran; hoseini.soodeh@gmail.com
5 Department of Speech and Language Pathology, School of Rehabilitation Sciences, Tehran University of
Medical Sciences, Tehran 11489-65111, Iran; s.seyedin86@gmail.com
6 Division of Diagnosis and Prevention, Iranian Special Education Organization, Tehran 1416935684, Iran;
nmahmoudizadeh@yahoo.com
7 Autism Rehabilitation Center, Roozbeh Hospital, Tehran University of Medical Sciences, Tehran 1514945311,
Iran; dr.ranjbar135@yahoo.com
* Correspondence: rabiee.a@tak.iums.ac.ir




Received: 7 January 2019; Accepted: 15 February 2019; Published: 20 February 2019


Abstract: Several studies have examined the cognitive profile of people with high-functioning
autism spectrum disorders (ASD) (IQ > 70), and its relationship with the symptoms of ASD
and attention deficit hyperactivity disorder (ADHD), using the Wechsler Intelligence Scale for
Children-IV (WISC-IV). However, no data exist on the similarities or differences in this profile
in less affluent countries. The present study examined the cognitive profile and its relationship with
the symptoms of ASD and ADHD in 30 subjects aged 6–16 years with high-functioning ASD and
compared the results with those of 30 typically developing (TD) subjects. In line with previous
research findings, the WISC-IV cognitive profile analysis of subjects with high-functioning ASD
showed a good competence in Matrix Reasoning and weaknesses in Comprehension, but the main
distinguishing point was the competence in processing speed in both groups. In the present study,
the Verbal Comprehension Index correlated negatively with the communication symptoms, and the
Working Memory Index correlated positively with the social symptoms in the ASD group. Given
the similarities that exist between the results of the present research and previous studies, it may be
concluded that there are similarities in the cognitive profile of individuals with ASD.

Keywords: autism spectrum disorders; ASD; intelligence; WISC-IV; ADHD; Cognition

1. Introduction
The Diagnostic and Statistical Manual of Mental Disorders, Fifth Edition (DSM-5) defines
autism spectrum disorders (ASD) as neurodevelopmental disorders [1]. ASD is observed in different
cultures [2]. In Iran, an autism prevalence rate of 95.2 per 10,000 has been reported [3]. ASD can
be accompanied by intellectual disability. Baio et al. [4] reported that 31% of people with ASD
have an intellectual disability (full-scale IQ, FSIQ < 70). Individuals with ASD without intellectual

Behav. Sci. 2019, 9, 20; doi:10.3390/bs9020020 www.mdpi.com/journal/behavsci

Behav. Sci. 2019, 9, 20 2 of 16

disability (FSIQ > 70) are typically referred to as having high-functioning ASD [5]. Intelligence tests
can provide interesting information about the cognitive strengths and weaknesses of people [6].
The Wechsler Intelligence Scale for Children (WISC) is one of the most commonly used tests for
measuring intelligence in ASD individuals, and there is a wealth of literature on this tool. The test
was updated in 2003, and its fourth edition (WISC-IV) is available with some changes made to the
previous version. The WISC-IV has four indices, including the Verbal Comprehension Index (VCI),
the Perceptual Reasoning Index (PRI), the Working Memory Index (WMI), and the Processing Speed
Index (PSI). It also contains 10 core subtests, including Similarities, Vocabulary, Comprehension,
Block Design, Picture Concepts, Matrix Reasoning, Digit Span, Letter-Number Sequencing, Coding,
and Symbol Search, and five complementary subtests, including Information, Word Reasoning, Picture
Completion, Arithmetic, and Cancellation [7]. One of the main points about intelligence tests, especially
the WISC-IV, is that they provide valuable information about a subject’s cognitive strengths and
weaknesses [6].
The first study of the cognitive profile of people with ASD using the WISC-IV was performed
by Wechsler in the WISC-IV manual on 27 individuals aged 9 to 15 who had ASD without comorbid
intellectual disability, and 19 ASD people aged 7 to 16, His results revealed an average cognitive profile
in the individuals with ASD without comorbid intellectual disability and a low average profile in
the ASD group. The PSI was the weakest index, and Coding and Symbol Search were the weakest
subtests. The Comprehension subtest was also one of the weakest in the ASD group. The VCI was
higher than the other indices in the individuals with ASD without comorbid intellectual disability, and
the PRI was the highest in the ASD group [7]. Mayes and Calhoun [8] studied 54 children with ASD,
aged 6 to 14, with high-functioning autism (HFA), and revealed the lowest scores on the WMI and PSI
indices. Among the nonverbal subtests, Matrix Reasoning and Picture Concepts received the highest
scores, and the subtests Coding, Symbol Search, Letter-Number Sequencing, and Digit Span received
significantly lower scores than the norms specified for the tests. Oliveras-Rentas et al. [9]) conducted
the WISC-IV test on 51 ASD people and found that the PSI was the most significantly low index,
and the PRI had significantly higher scores than the WMI. Matrix Reasoning and Similarities were the
strongest subtests, and Comprehension, Coding, and Symbol Search were the weakest. Nader, Jelenic,
and Soulieres [10] examined the Wechsler cognitive profiles of 51 ASD people, 15 individuals with
ASD without comorbid intellectual disability, and 42 typical development (TD) people. Their results
showed that, in the ASD group, the PRI was higher than the full-scale IQ (FSIQ) and the scores of
the other indices. Block Design, Matrix Reasoning, and Picture Concepts were the strongest subtests,
while Comprehension, Digit Span, Letter-Number Sequencing, and Coding were the weakest. In the
group of individuals with ASD without comorbid intellectual disability, the VCI score was higher than
the scores of the other indices, and the PSI had the lowest score. Vocabulary and Similarities were
the strongest subtests, and Digit Span and Coding were the weakest. In the TD group, there were no
significant differences between the indices, and Picture Concepts and Vocabulary were the strongest
subtests. Nader et al. [11] compared the Wechsler cognitive profile of 25 ASD people aged 6 to 16 to
that of 22 TD people. In the ASD group, the PRI received significantly the highest score, and in the TD
group, the WMI received the lowest score significantly.
Although this is a strong overselling of the power of IQ score, the WISC-IV provides useful
information about the effect of people’s culture, biology, maturation, and differences in interventions
on cognitive functioning [6]. All of the cited literature involves research conducted in Western countries,
and the effect of culture on this issue has not been seriously taken into consideration, so it is difficult to
determine whether the behavior and cognition of people with ASD vary among different cultures [12].
Intercultural studies allow researchers to recognize the common and distinctive characteristics of
the disorder between different cultures, and provide hypotheses about the cognitive nature of
developmental disorders. No studies have yet directly compared the phenotypes of ASD children
from different cultural backgrounds [13]. One of the aims of this study is to provide information that
can help compare different communities.
Behav. Sci. 2019, 9, 20 3 of 16

Assessing the cognitive profile in the ASD group poses the question of whether there is a
relationship between ASD symptoms and the cognitive profile. Most studies on the possibility
of a relationship between ASD symptoms and performance on cognitive tests have been conducted
using previous versions of the WISC or other intelligence assessment tests. Oliveras-Rentas et al. [9]
investigated the relationship between ASD symptoms and the cognitive profile using the Autism
Diagnostic Observation Schedule (ADOS) [14] and the WISC-IV, and concluded that there was a
negative relationship between the scores of the Communication subtest in the ADOS, and the VCI
and the PSI and their subtests in the WISC-IV. There was also a negative relationship between the
Vocabulary and Comprehension subtests of the WISC-IV and the scores of the Reciprocal Social
Interaction subtest of the ADOS.
Due to the reported 37–85% prevalence of simultaneous attention deficit hyperactivity disorder
(ADHD) and ASD symptoms [15], investigating the relationship between ADHD symptoms and the
cognitive profile of ASD people is important. Measuring the cognitive profile of those with ADHD
using the WISC-IV showed weaknesses in the WMI and PSI [16,17], but Oliveras-Rentas et al. did not
confirm the significance of this relationship in the ASD group with ADHD symptoms [9]. Another aim
of the present study was to evaluate the cognitive profiles of people with ASD and ADHD symptoms
using the WISC-IV in a developing society.

Aims
The general aim of this study was to evaluate the cognitive profile in Iranian people with
high-functioning ASD and compare it to the profile of TD people, and then investigate the relationship
between the cognitive profile in those with high-functioning ASD with ASD and ADHD symptoms
through three sub-aims:

(1) Investigating the cognitive profile of those with ASD and TD using the Persian version of the
WISC-IV, and comparing it to previous research findings;
(2) Investigating the possible relationship between ASD symptoms and performance on the WISC-IV;
(3) Investigating the relationship between ADHD symptoms and the cognitive profile of people with
high-functioning ASD.

These goals offer data for comparing different communities from varying cultural backgrounds.
Thus, the relationship between the WISC-IV’s cognitive profile and ASD and ADHD symptoms can be
assessed for the first time in a developing country.

2. Materials and Methods

2.1. Participants
Two groups of people participated in this study, those with high-functioning ASD (n = 30), and the
typical development or TD group (n = 30).

2.1.1. Participants with High-Functioning ASD

The participants of this study consisted of 27 boys (90%) and 3 girls (10%) with high-functioning
ASD, aged 6 to 16 (mean = 11 years and 1 month, SD = 2 years and 9 months). In terms of family
income, 14 participants had a high socioeconomic status, seven had a moderate status, and nine had
a low socioeconomic status. A total of 13 participants had parents with associate degrees or lower,
and 17 had parents with bachelor’s degrees or higher. All of the participants were Iranian and lived in
Tehran, Iran.
The samples were selected through convenience sampling. A total of 121 children and adolescents
aged 6 to 16 with a diagnosis of high-functioning ASD were referred by the Iranian Special Education
Organization, the Autism Charity, Roozbeh Psychiatric Hospital, the specialized clinics of schools
of rehabilitation sciences, and 10 specialized clinics for children with autism in Tehran, Iran. All of
Behav. Sci. 2019, 9, 20 4 of 16

the participants had previously been evaluated by psychiatrists, psychologists, and speech therapists.
A trained evaluator from the Special Education Organization had administered the Autism Diagnostic
Interview-Revised (ADI-R) to 51 (42.14%) of the parents of participants. A total of 15 families (12.39%)
were unwilling to participate, and eight (6.61%) of the children were outside of the study’s age range.
To review the inclusion criteria, the children’s medical records were studied at a meeting with their
families and the following items were assessed: comorbid conditions such as metabolic disorders and
genetic syndromes, history of neurological diseases (such as trauma, brain lesions, tumors, stroke,
epilepsy, and Tourette’s syndrome), and other medical issues that could affect cognition, comorbid
mental disorders such as schizophrenia or bipolar disorder, and special visual, hearing, or motor
problems that could affect their performance on the test. If any of these conditions were present,
the child was excluded from the study. At this stage, eight (6.61%) children were excluded. To ensure
the presence of ASD symptoms at the time of the study and estimate their severity, the second version
of the Gilliam Autism Rating Scale (GARS-2) was conducted by a trained person. The participants
who obtained the cut-off score entered the next stage. At this stage, three (2.48%) children did not
obtain the minimum cut-off score of GARS-2 and were excluded.
At the next step, a well-trained and experienced psychologist, with certification for implementing
the Wechsler test in ASD people and who is currently active in the implementation of the test, conducted
it in a quiet room with proper light and temperature. This person was not informed about the aims and
outcomes of the study. The participants entered the study if they had an FSIQ over 70. At this stage,
57 (47.1%) children were excluded and 30 were enrolled. The parents completed the Conners’ Parent
Rating Scale-Revised (Short) (CPRS-RS). Table 1 show the GARS-2 and CPRS-RS scores for this group.

Table 1. The mean (SD) of the indices and the subtests of the GARS-2 and CPRS-RS in the ASD sample
(N = 30).

Scales Mean (SD) 95% CI

GARS Autism Index 74.18 (11.85) [69.73, 78.59]
Stereotyped
6.1 (2.66) [5.10, 7.09]
Behaviors—SS
Stereotyped
15.46 (15.25) [9.77, 21.16]
Behaviors—%ile
Communication—SS 4.26 (1.38) [3.74, 4.78]
GARS-2
Communication—%ile 6.8 (3.41) [5.52, 8.07]
Social interaction-SS 6.1 (1.56) [5.51, 6.68]
Social interaction—%ile 11.7 (5.84) [9.51, 13.88]
Total Standard Score 16.46 (4.42) [14.81, 18.11]
Rank Percent 23.13 (15.59) [17.31, 28.95]
ADHD Index 60.83 (7.28) [58.14, 63.55]
Oppositional 54.97 (9.70) [51.34, 58.59]
CPRS-R:S Cognitive Problems/
60.40 (9.03) [57.03, 63.77]
Inattention
Hyperactivity 64.23 (11.59) [59.91, 68.56]
CI: Confidence Interval; SS: Standard Score; %ile: Percentage Score; GARS-2: Gilliam autism rating scale-2; CPRS-R:S:
Conners’ Parent Rating Scale-Revised, (Short); ADHD: Attention-deficit/hyperactivity disorder.
Behav. Sci. 2019, 9, 20 5 of 16

2.1.2. TD Participants
The TD participants were selected through cluster sampling. Some of the municipal districts of
Tehran were randomly selected, and an all-boys’ and an all-girls’ school were randomly selected from
the primary and lower secondary schools of each district. The TD students were randomly selected
from these schools in proportion to the ASD students’ genders and dates of birth. The Strengths and
Difficulties Questionnaire (SDQ), which will be described in the Research Tools section, was completed
for the students by their teacher. If the subjects’ mental health was approved by the questionnaire,
their physical and mental health status was reconfirmed by reviewing their educational records and
talking to their teachers. The students with a personal or family history of neurological, psychiatric, or
other conditions affecting brain development were excluded. Of the 30 students examined, three were
replaced with three others for these reasons. Ultimately, 30 TD students were selected in proportion to
the ASD participants. None of these students had any academic problems.

2.2. Research Tools

correlation coefficient was used to examine the relationship between the FSIQ, the Wechsler intelligence
The Persian version of the Wechsler Intelligence Scale-Fourth Edition (WISC-IV): The present study
indices and subtests, and the GARS-2 and CPRS-RS subtests. The level of statistical significance was set at
used the Persian version of the WISC-IV, which has had its reliability and validity assessed in 872
P < 0.05. The mean of each index or subtest in the groups was used to draw the group analysis plots of the
Iranian children aged 6 to 16 [18]. The Introduction section discussed the components of this scale.
data.
The Gilliam Autism Rating Scale-Second Edition (GARS-2): The GARS-2 is a behavior scale for use
with people at ages 3 to 22. The scale includes 42 items in three subscales, including stereotyped
2.4. Ethics
behaviors, communication, and social interaction. The GARS-2 contains an additional 14 items that
This study received an ethics code (ir.iums.rec.1394.9211363204) from the Ethics Committee of Iran
examine children’s development in the first three years of their life. These items are answered with
University of Medical Sciences. Written consent was obtained from the parents for their children’s
Yes and No, and provide additional information. This study used a version of the test normalized by
participation in the study, and verbal consent was obtained from the participants themselves.
Samadi and McConky [19] for use among Iranians.
The Strengths and Difficulties Questionnaire (SDQ): The SDQ is a short screening tool that is used to
3. Results
determine behavioral and emotional problems in children and adolescents, and assesses five main
subgroups of psychiatric
3.1. Intergroup symptoms, including conduct problems, hyperactivity/inattention (HI),
Comparison
emotional symptoms, peer relationships, and pro-social behavior. The total score of the first four yields
The two groups were not significantly different in terms of age and gender, but they were in terms of
the overall score in terms of problems. The impact score is also calculated to determine the impact of
the FSIQ. The presence of significant FSIQ differences among the groups presents a methodological
the children’s problems on their own and their family’s daily life [20]. The reliability and validity of
dilemma. Some authors have argued that FSIQ should not be included as a covariate because psychiatric
the Persian version of this questionnaire have been calculated in two separate studies [20,21].
disorders may directly cause mild FSIQ deficits compared to individuals without psychiatric disorders and
Conners’ Parent Rating Scale-Revised (Short) (CPRS-RS): This short 27 item form is suitable for ages
that controlling for the FSIQ removes a portion of the variance that is associated specifically with
3 to 17. The parents use this scale to score their child’s behavior over the prior month on a four-point
psychiatric disorders. In contrast, some authors have argued that the FSIQ should be controlled [27].
scale. The scale has four subscales, including (1) oppositional, (2) cognitive problems/inattention,
Because this issue has not been resolved definitely, the results are reported both with and without
(3) hyperactivity, and (4) the ADHD index [22]. In Iran, the standardization and reliability of the
controlling for the FSIQ between two groups. To match the FSIQ for the reasons mentioned, the researchers
CPRS-R have been assessed in two separate studies [23,24]. The present study used the CPRS-RS to
did not use the FSIQ as a covariate in statistical analysis, but, instead, the participants with an average FSIQ
measurewere the ADHDfrom
selected indexthealong with three other
high-functioning ASD (nscales,
= 13) and
and its
TDgoal wasgroups.
(n = 13) not to The
make a diagnosis.
statistical tests showed
Corbett et al. [25] also recommended this tool for classifying ASD
no significant differences in the FSIQ scores between the two groups. people with ADHD comorbidity.
The self-administered
The comparison demographic and economic
of the WISC-IV questionnaire:
index and Thisbetween
subtest scores questionnaire contains questions
the high-functioning ASD and TD
aboutgroups
the children’s demographics along with parental data, including their employment
without controlling for the FSIQ showed a significant difference on combined dependent status, variable
annual ofincome,
WISC-IVand levelscores
subtest of education.
(F (10, 49) = 14.30, P = 0.001, Wilk’s Lambada = 0.25, Partial ŋ2 = 0.74) and index
scores (F (10, 48) = 3.08, P = 0.005, Wilk’s Lambada = 0.60, Partial ŋ2 = 0.39). Analysis of each of the dependent
2.3. Statistical Analysis
variables, as shown in Table 2, revealed that participants with ASD were weaker than TD participants in
The data were
all WISC-IV statistically
index analyzed
and subtest scores. with SPSSdifference
The mean software using bothmaximum
produced descriptive andsizes
effect analytical
for the VCI and
Picture
measures. TheConcept
mean and subtest, and deviation
standard minimumwere effectused
sizes
forfor
thethe WMI andpart
descriptive Digitof Span subtest. For
the analysis. After
thematching
two part,
analytical groupsthein term of FSIQ, the effect
Kolmogorov–Smirnov of the
test was usedgroups
to ensure werethestill statistically
normal significant
distribution on combined
of the data.
dependent variables
The multivariate analysis of of variance
WISC-IV(MANOVA)
subtest scoreswas(F (10,
used15) to =compare
3.02, P = the
0.02, Wilk’s differences
existing Lambada = in 0.33,
thePartial ŋ2
= 0.66),
Full-Scale but not inQuotient
Intelligence index score.
(FSIQ)Analysis
and theofWechsler
each of the dependent
intelligence variables,
indices, and as shownbetween
subtests in Tablethe
2, revealed
ASD andthat TD
participants with ASD
groups. Repeated were weaker
measure ANOVAs significantly
were used than TD participants
to investigate in the differences
within-group subtests of Picture
among Concepts
the index(F (1,
and24)subtest
= 19.48,scores.
P = 0.001, Partial (ŋ2 ) was
Eta-squared 2 = 0.45), Comprehension
reported as the effect(Fsize
(1, 24) = 12.90, P = 0.001,
in ANOVA.
Partial ŋ2 = 0.35), Vocabulary (F (1, 247) = 6.30, P = 0.01, Partial ŋ2 = 0.20), and Symbol Search (F (1, 24) =
4.32, P = 0.04, Partial ŋ2 = 0.15).

Table 2. Comparison of the mean (SD) of the indices and the subtests of the WISC-4 between the high-
functioning ASD and TD groups with FSIQ matched and not matched (ASD (n = 13) and TD (n = 13)), and
without (ASD (n = 30) and TD (n = 30)) FSIQ matched.
 correlation coefficient was used to examine the relationship between t
correlation coefficient indiceswas and used subtests, and the the
to examine GARS-2 and CPRS-RS
relationship between subtests.
the FSIQ, Thethe levelW
indices and subtests, P < 0.05. andThe themean GARS-2 of each
and indexCPRS-RS or subtest
subtests. in The
the groups
level ofwas used tosig
statistical d
P < 0.05. The mean data. of each index or subtest in the groups was used to draw the group a
Behav. Sci. 2019, 9, 20 data. 6 of 16
2.4. Ethics
2.4. Ethics
This study received an ethics code (ir.iums.rec.1394.9211363204)
The standard difference
correlation coefficient was used to examine is the difference between the two test means, divided byWechsler
the square root
This study received anofbetween
the relationship
University Medical
ethics codetheSciences.
FSIQ, the Written consent
(ir.iums.rec.1394.9211363204) intelligence wasfrom obtained
the Ethicsfrom
ofindices
the pooled variance
and subtests, and the GARS-2[26]. Pearson’s correlation
and CPRS-RS coefficient
subtests. Thewas used
level ofto examine
statistical the relationship
significance was set at
University of Medical participation Sciences. in theWritten study, and verbalwas
consent consentobtained was obtained from thefrom the paf
parents
between
P < 0.05.the TheFSIQ,mean the Wechsler
of each intelligence
index participation
or subtest in the indices groups andwas subtests,
used toanddraw the theGARS-2
group and CPRS-RS
analysis plots ofparticipants
the
in the study, and verbal consent was obtained from the them
ed to examine the subtests. data. Thebetween
relationship level of the
correlation statistical
FSIQ,
coefficient significance
the Wechsler
was used was to set 3.
intelligence at P < the
Results
examine 0.05.relationship
The mean of each index
between the FSIQ, or subtest the Wechslerin intelligence
GARS-2
efficientand usedthe
wasCPRS-RS groups
subtests.was
to examine The
the used
indiceslevelto ofdraw
and
relationship thecorrelation
statistical
subtests,
between group
3.significance
Results
and the analysis
the GARS-2
FSIQ, was
coefficient plots
the set
and of the
atCPRS-RS
Wechsler
was used data. tosubtests.
intelligence examineThe the level of statistical
relationship between significance
the FSIQ, wastheset Wecat
ex or subtest
ubtests, and the in GARS-2 2.4.
the groups Ethics
andwas used
CPRS-RSP < 0.05.tosubtests.
draw
The the mean
The group of
indices
leveleach
analysis
of
andindex plots
statistical
subtests, 3.1.
or subtestof Intergroup
the in
significance
and thethe GARS-2Comparison
groups
was setandwas used to draw
at CPRS-RS subtests. the The group level analysis plots ofsigni
of statistical the
2.4. Ethics 3.1. Intergroup Comparison
mean of each index or subtest in data.
the groups was used P < to0.05.
draw
This study received an ethics code (ir.iums.rec.1394.9211363204) The the mean
group of each
analysis index plots
The two groups were or subtest
of the in
fromnot the groups
thesignificantly was
Ethics Committee used
different to draw
of in Iran the group
terms of age an an
This study received
University of Medical Sciences. Written an ethics data. code (ir.iums.rec.1394.9211363204)
The two consent the was
groups FSIQ.
were obtained
The presence
not significantly from
from the the Ethics
parents in
of different
significant Committee
fortermstheirof
FSIQ of
children’s
differences
age and gender, amongbut thethg
Iran University2.4.
participation inofthe Ethics
Medical
study,Sciences.
and verbal Written
consent consent
was was
obtained
dilemma. obtained from
Some from
the
authors the parents
participants
the FSIQ. The presence of significant FSIQ differences among the groups presents have argued for their
themselves.
that children’s
FSIQ should not be included
participation in theThis study, and 2.4. Ethics
verbal consent wasdisorders
obtained from
may the participants
directly cause mild themselves.
FSIQ
thics code (ir.iums.rec.1394.9211363204) fromstudy thereceived dilemma.
Ethics Committee
an ethics Some authors
code
of Iran have argued
(ir.iums.rec.1394.9211363204) that FSIQ shouldfromdeficits
not
thebe compared
included
Ethics Committeeastoaindividual
covariate
of Iran
dyes.received
Written an consent
ethics 3. code
Results
was obtained University
from the
(ir.iums.rec.1394.9211363204) of parents
Medical disorders
for
ThisSciences.
from their
studymay that controlling
directly
thechildren’s
Written
Ethics
received cause
consent
Committee
an ethics mild for
was
of FSIQ
code the
Iran FSIQ from
deficits
obtained removes
compared the aparents
(ir.iums.rec.1394.9211363204) toportion
individuals for of the
their
from variance
without psych
children’s
the Ethics th
C
3. Results
verbal
Medical consent was obtained
Sciences. Written consent from participation
thewas participants in the
obtained that
themselves.
study,
from and
University controlling verbal
theofparents Medicalpsychiatric
for
consentthe
forSciences. FSIQ
their disorders.
removes
waschildren’s
obtained
Written from In contrast,
a portion
consent some
of
the participants the
was obtained authors
variance have
themselves. thatargued
from the parents foris that
associate th
in the study, and verbal 3.1. Intergroup
consent was Comparison
obtained from psychiatric
participation
the participants in the Because
disorders.
themselves.
study, In this
and contrast,issue
verbal has
some
consent not been
authors
was resolved
have
obtained argued
fromdefinitely,
that
the thetheFSIQ
participants results
should
themseare
3.1. Intergroup Comparison
3. Results Because this controlling
issue has
The two groups were not significantly different in terms of age and gender, but they were in terms of not forbeenthe FSIQ
resolved between two
definitely, groups.
the To
results match are the FSIQ
reported for the
both
The two groups were not3.significantly ResultsFSIQfor different
didFSIQ notin use terms
the FSIQ of ageaand gender, but they were
the FSIQ. The presence of significant controlling the
differences between
among theasgroups.
two covariate
groups To in statistical
match
presents the FSIQ analysis,
a methodological for the reasonsbut, instead,mentio th
in terms of the 3.1. FSIQ. The presence
Intergroup Comparison did not of significant were FSIQselected differences from the among the
high-functioning groups ASDpresents (n = 13)a the participants wg
and TD (n = 13)
dilemma. Some authors have argued thatuse FSIQthe should
FSIQ asnot a covariate
be included in statistical
as a covariateanalysis, but, instead,
because psychiatric
Comparison methodological dilemma. Some3.1. authorsIntergroup have argued
Comparison that
no significant FSIQ should
differences not be
in included as a covariate
significantly different disorders
in terms mayof age directly
The
and two cause
gender, mild
groups but weretheyselected
FSIQ
were deficits
were infrom
compared
not significantlyterms the
of high-functioning
to individuals
different in terms ofthe
ASD
without age(nFSIQ
psychiatric
and scores
= 13)gender,
and TD between
disorders
but = 13)the
(n they and
were twoingroups.
groups. The stat
terms of
because psychiatric
that controlling disorders
for the FSIQ may nodirectly
removes significant cause
asignificant
portion mild
differences FSIQ
The
ofFSIQthe indeficits
comparison
the
variance FSIQ compared
of the
scores to
WISC-IV
between individuals
index
the two without
and subtest
groups. scores between
gnificant
groups FSIQ
were notdifferences
significantly among the the
different FSIQ.
ingroups The
terms presents
presence
of age and
The aof methodological
gender,
two groups but they were were
not in termsthat
differences
significantly ofamong is associated
different the in groups
termsspecifically
presents
of age and with
a gender,
methodological
but they
psychiatric disorders and that controlling The for the
comparison groupsFSIQargued removes
without a the
portion
controlling forofshould
the
the variance
FSIQ showed that is [27].
a significant differenca
eargued
presencethat of
FSIQ psychiatric
should
significant not
FSIQ bedisorders.
included
dilemma.as
differences In acontrast,
Some covariate
among authors some
thetheFSIQ. because
haveThe
groups authors
argued
psychiatric
presents
presence thatof
have aFSIQ
ofthe WISC-IV
should
methodological that
significant notindexFSIQ
be
FSIQ and
included subtest
differences asbe ascores
controlled
covariate
among between
the because
groupsthe psychiatric
high-functi
presents
associated specifically with psychiatric groups disorders. In
of contrast,
WISC-IV some
subtest authors
scores have
(F (10, argued
49) =not that
14.30, the FSIQ
Ppsychiatric
= 0.001, Wilk’s Lambada
mild FSIQ deficits
me authors have argued Because
compared this
thattoFSIQ issue
individuals
disorders
should has not not
without
may been
bedirectly resolved
psychiatric
included
dilemma. cause aswithout
amild definitely,
disorders
Some covariateFSIQ controlling
authors and the results
deficits
because
have for
compared
argued the are
psychiatric FSIQ
that showed
toreported
individuals
FSIQ should a significant
both with
without be and difference
included without as aoncovariate
combined
disorders andbe
should be controlled [27]. Because this issue has not
scores been (F resolved
(10, 48) = definitely,
3.08, P = the
0.005, results
Wilk’s are
Lambada reported = 0.60,
Qy directly
removescause a portion controlling
mild FSIQ of the for
variance
deficits thatthe
compared FSIQ
controlling
that to between
is individuals
associated
for of
two WISC-IV
groups.
the FSIQ
disorders specifically subtest
To
mayremoves
without match
directly scores
with
psychiatric the
a portion
cause (F
FSIQ (10,
disorders
mild FSIQfor 49)
the
of and =
the 14.30,
reasons
variance
deficits P = 0.001,
mentioned,
compared Wilk’s the
that istoassociated Lambada
researchers
individualsspecifically without ŋ2
Partial
= 0.25, =with
Partial
psychia0.39ŋ
rast,forsome
ng FSIQ both
the authors did
have
removes with aand
notargued
use thewithout
FSIQof
psychiatric
portion that ascontrolling
the aFSIQ
the covariate
disorders.
should
variance thatfor
scoresInthe (F
incontrollingFSIQ
is(10,
statistical
becontrast,
that controlled between
48)
analysis,
associated some
for thetwo
variables,
= 3.08,
[27]. but,
authors=as
PFSIQ groups.
shown
0.005,
instead,
specifically Wilk’s
have
removes Toin
the match
Table
Lambada the
participants
argued
with a portionthat FSIQ
2, revealed
=the0.60,
of forPartial
with
FSIQ
the the
that
an reasons
participants
ŋ2 = 0.39).
average
should
variance beFSIQ
that with
Analysis
controlled
is ASD
associated w
of ea
[27].
een resolved
isorders. mentioned,
were
definitely,
In contrast, some selected the
theauthors
results researchers
from
Because thereported
are
have this
argued did
high-functioning
issue not
both use
variables,
hasthe
psychiatric
that with the
notFSIQASDbeenFSIQ
as should
and shown
disorders. all as
= 13)in
(nresolved
without a
WISC-IV
be
Incovariate
and Table TD 2,index
definitely,
contrast,
controlled in and
(n revealed
= some
13)statistical
the subtest
groups.
[27]. results
authors analysis,
scores.
that participants
The are The
statistical
have but,
reported
arguedmean
with instead,
tests difference
ASD
both
that showed were
with
the FSIQ produced
weaker
and without
should m
thanb
en twohas
issue groups.
not To beentheno
match participants
significant
the FSIQ
resolved with
differences
controlling an
for the reasons
definitely, average
in the
theforresults FSIQ
all
FSIQ
mentioned,
FSIQ
Because were
WISC-IV
scores
arebetween
reported
this selected
index
between
the issue Picture
researchers
two both from
and the
groups.
haswith Concept
the
subtest
two
notTo andgroups.
beenmatch subtest,
high-functioning
scores.
without The
the FSIQ
resolved and
mean minimum
ASD
for the reasons
definitely, (n
difference = effect
13)
the results and
producedsizes
mentioned, TD for
are reported the
maximum WMI
the researchers and
effect
both w sD
riate
thein statistical
FSIQ between (n =
analysis, 13)The
two groups. groups.
but,comparison
instead,
Todid The
matchnotstatistical
of
theuse the
participants
the the
FSIQ FSIQtests
WISC-IVfor Picture
asshowed
aindex
with
controlling
the Concept
covariate
anfor
reasons no
and
average
the intwo
significant
subtest
mentioned, groups
subtest,
statistical
FSIQFSIQ scores
between and in
differences
theanalysis, term
minimum
between
researchers
two groups. of
in
but,theFSIQ,
effect
the FSIQ the
sizes
high-functioning
instead,
To match effect
scores for the
the participants
the FSIQ of
between the
WMI
ASD forwithgroups
and
the
and Digit
TD
the reasons were
an average Spanstill
FSIQ
mentione sta
subt
unctioning
e FSIQ as aASD (ntwo
covariate ingroups.
=groups
13) andwithout
TD (n
statistical were=controlling
13)
analysis, selected
groups. for
from
The
but, instead, the
did two
the FSIQ groups
statistical
not showed
high-functioning
the use tests
participants
the in
FSIQ dependent
term
showed aof
a significant
as ASD
with anFSIQ,
covariate variables
(n
average the
=difference
13)inand of TD
effect
FSIQ WISC-IV
on of
statistical (n =the
combined13)subtest
analysis, groups
groups.but,scores
dependent The (F
were
instead, (10,
still
variable
statistical
the 15) =tests
3.02,showed
statistically
participants Psignif
=with
0.02
efrom
FSIQthe scores between
high-functioning Thetwo
of WISC-IV
the comparison
ASD subtest
groups.
no=significant
(n of
13)scores
and theTD WISC-IV
(F (10, = dependent
differences
(n were 13)index
49) =groups.
14.30,
in the
selected andvariables
P
FSIQ
from
The =0.001,
=subtest0.66),
scoresof but
WISC-IV
scores
Wilk’s
statistical
the betweennot
high-functioningbetweenin
Lambada
tests index
subtest
the
showed the
two =score.
scores
ASD0.25, Analysis
=(F13)
high-functioning
(nPartial
groups. (10, ŋ2
and 15)of
=TD each
=0.74)
3.02,
ASD =of
(n and Pand
13)=the0.02,
index dependent
groups. Wilk’sTheLambavaria
statist
ISC-IV indexinand
differences TD
the subtestgroups
scores
FSIQ scores (F
scores without
(10, 48)
between
between = controlling
3.08,
ThethetheP = 0.005,
comparison
two for = 0.66),
the
Wilk’s
high-functioning
groups. FSIQ but
Lambada
of the WISC-IV
no significant ASDnot
showed that
in
=
and
differences a
0.60,
indexparticipants
index
TD in score.
significant
Partial
and ŋ2
thesubtest = with
Analysis
difference
0.39).
FSIQ scores ASD
of each
on
Analysis
scoresbetween were
betweentheof
combined
of weaker
the
each dependent
thetwo of significantly
dependent
the variables,
dependent
high-functioning
groups. than
ASD andpart
as TDshown TD
rparison
the FSIQ showed
of the variable
WISC-IVavariables,
significant
indexof WISC-IV
as
and shown
difference
groups
subtest subtest
in Table
without
on
scores scores
2,
combined that(F
revealed
controlling
between participants
(10,
dependent
The the 49)
that
for
comparison Concepts
= 14.30,
participants
the with
variable
FSIQofshowed
high-functioning Pthe ASD
= with(F
0.001, (1,
WISC-IV
ASD were24)
aASDWilk’s = 19.48,
weaker
TDLambada
were
significant
andindex P
weaker = 0.001,
significantly
difference
and =
subtestthan Partial
0.25,
on TD than
Partial TD ŋ
participants
combined
scores between 2 = 0.45),
participants
2 dependent
= the Compreh
in the
in high-function
variable
10, 49) = 14.30,for
ut controlling 0.74)
P =the allFSIQ
0.001, and
WISC-IV index
Wilk’s
showed ofscores
index
Lambada and(F
WISC-IV
a significant (10,
=subtest
0.25, 48)
subtest =Concepts
scores.
Partial 3.08,
scores
difference
groups ŋ2The =P =mean
(F0.74)
without
on (F
(10, (1,
0.005,and
combined49) Partial
24)
Wilk’s
difference
=index
14.30,
controlling ŋ2
=dependent
19.48, ==P
Lambada 0.35),
Pproduced
for = 0.001,
0.001,
the Vocabulary
=Wilk’s
variable
FSIQ Partial
0.60,
maximum Partial
showed Lambada(Faŋ(1,
effect2 =247)
20.39).
=sizes
0.25,
significant =for
= 0.45),6.30,the
Partial PVCI
difference =ŋ20.01,
Comprehension
Analysis and
= on Partial
0.74) and(Fŋ2
combined (1,= de
index 0.
24
5, Wilk’s
ubtest (F (10,of
Lambada
scores Picture
each
=49)
0.60, ofConcept
thePŋ2
Partial
= 14.30, scores= subtest,
dependent
= 0.001,0.39).
(FWilk’s
(10, and
Analysis
48)Lambada ofPartial
=minimum
variables, 3.08,
of as
P ==shown
each
WISC-IV ŋ2
of effect
0.005,
the
0.25, = 0.35),
in 4.32,
sizes
dependent
Wilk’s
subtest
Partial Table forP 2,
==the
Vocabulary
Lambada
scores
ŋ2 0.04,
0.74)
(F WMI
revealed
(10, =Partial
(F
0.60,
and49) (1,
andthatŋ2
247)
=Partial
index14.30,= =0.15).
Digit 6.30,
Span
participants
ŋ2
P Psubtest.
==0.001,
0.39). = 0.01,
with
Analysis
Wilk’s Partial
After
ASD Lambada ŋ2
matching
ofwereeach = 0.20),
of the and
= 0.25, Symbo
dependent
Partial ŋ2
, revealed
48) = 3.08, Pthat
= 0.005, two groups
weaker
participants
Wilk’s than
with
Lambada TD in =termwere
0.60, of
participants
variables,
ASD FSIQ,
asweaker
Partial inŋ2all
shown the4.32,
WISC-IV
=than
in
scores P
effect
Table
0.39). TD
(F = 0.04,
of
2,index
participants
Analysis
(10, Partial
the=of
revealed
48) groups
and
3.08,
each ŋ2
that
in
P =
subtest 0.15).
were
participants
=of0.005,
the still statistically
scores.
dependent
Wilk’s Thewith
Lambada mean ASD significant
=difference
werePartial
0.60, weaker on
produced combined
ŋ2than= 0.39). TDAnalysis
participants of eachin
dependent variables of Table 2. Comparison of the mean (SD) of the indices and the subtests
scores.in
shown The mean
Table maximum
2, difference
revealed effect
that allsizes
produced WISC-IV
participants forWISC-IV
maximumthe
with VCI
index effect
ASD subtest
and
and sizes
subtest
variables,
were scores
Picture
for
as scores.
the
weaker
shown (F
Concept
VCI (10,
thanThe
inand15)
mean
Table
TD =participants
subtest, 3.02,and P =minimum
2,difference
revealed 0.02, Wilk’s
inproduced
that Lambada
effect sizes for
maximum
participants = 0.33,
with the
effect
ASD Partial
WMI wereŋ2
sizes for the VCI
weaker andT
than
Table 2. Comparison functioning
of the mean ASD and (SD)TD groups
ofasthe with and
indices FSIQthe matched
subtests andof not
the matched
WISC-4 be (A
minimum effect
ndex and subtest scores. =
sizes 0.66),
for
and Digit thebut
The mean WMInot in
Picture
andindex
Span difference Digitscore.
Concept
subtest. After Span
produced Analysis
subtest,
subtest.
matching
all WISC-IV
maximum of
and
two each
After
minimum
groups
index of
matching
effect andthe in
sizesdependent
effect
term
subtest sizes
for theof VCI
scores. variables,
for
FSIQ, the
The
and theWMI
mean shown
effectand in
Digit Table
of the groups
difference Span
produced 2, revealed
subtest.
were After matching
maximum effect size
functioning ASD without
and TD (ASD
groups (nwith
= 30)FSIQ
and TD (n = 30))
matched and FSIQ
not matched.
matched (ASD (n = 13) and T
, the
ept effect and
subtest, of the thatstatistically
groups
minimum
still participants
were two
effect still
sizes with
groups ASD
for the
significant onwere
statistically
in
WMI term
Picture
and
combined weaker
significant
of FSIQ,
Digit on
Conceptsignificantly
the
Span
dependentcombined
effect
subtest,
subtest. of than
and
After
variables TD
theminimum
groups participants
matching
of WISC-IV were in statistically
still
effectsubtest
sizes the
for subtests
the WMI
scores ofandPicture
significant
(F (10, 15) DigitonSpancombined
subtes
without (ASD (n = 30) and TD (n = 30)) FSIQ matched.
IV subtest
n term of scores
FSIQ, (FtheConcepts
=(10, 15)P ==of
effect
3.02, (F
3.02,(1,
the
0.02, 24)
dependent
PWilk’s = 19.48,
=groups
0.02, Wilk’s
were
LambadaP =Lambada
variables 0.001,
still
two Partial
=ofstatistically
WISC-IV
groups
0.33, =Partial
0.33, ŋ2 =2of
in subtest
Partial
term
significant = FSIQ,
ŋ2 0.45),
scores
0.66), on(FComprehension
but (10,
the
combined15)
in =index
noteffect 3.02,
of theP (F (1, Analysis
24)
= groups
0.02,
score. =were
Wilk’s 12.90, P
Lambada
still
of =statistically
each
ASD
0.001,
= 0.33, Partial ŋ2
significa
Partial ŋ2 = 0.35), Vocabulary (F (1, 247) = 6.30, P =Indices
0.01, and the Subtests
Partial ŋ2 = of and Symbol Search (F (1, 24) =
0.20),
Analysis
riables of each of
of WISC-IV the
subtest dependent
scores =(F0.66),
variables,
(10, but
15) =not
as
3.02,in
shown
Pindex
dependent
= in
0.02,
of the dependent variables, as shown in Table 2, revealed score.
Table
Wilk’s Analysis
2,
variables revealed
Lambada of of each
WISC-IV
= 0.33,of the
Partial dependent
subtest ŋ2
scores
that participants with (Fvariables,
(10,
FSIQ 15) = as shown
3.02,
ASD were weaker
ASD P = in
0.02,Table
Wilk’s
TD 2, revealed
Lambada
Indices and the WISC-4 Mean
were weaker
ot in index score. 4.32, P of
significantly
Analysis =than
significantly 0.04,
each TD
thanPartial
that
ofTD the ŋ2 = 0.15).
participants
participants
dependent
participants in=with
thethe ASD
subtests
variables,
0.66),
in but as the
were
not
subtests ofinofSubtests
shown weaker
Picture
index ofsignificantly
in score.
Picture Table 2,
Analysis
Concepts revealed
FSIQ (F than
of24)
(1, TD
each= ofparticipants
thePdependent
19.48, = 0.001,
(SD)in the subtests
95% CI asof
variables,
Partial Picture
Mean
shown (SD
in
the WISC-4 Mean
=nts
0.001,
withPartial
ASD were ŋ2 =weaker
20.45),
= 0.45), Comprehension
Concepts (F
significantly
Comprehension (1,
than 24)
(F
TD
(F (1, (1,
that=participants
24) 19.48,
24) = 12.90,
P =P0.001,
participants
= 12.90, P0.001,
=in =with
0.001,
thePartial
subtests
ASD were
Partial ŋof
2 = 2weaker
= 0.45),
Picture
0.35), Comprehension
significantly
Vocabulary
(SD) Not 95%
(F than
(1, CI (F
247)
86.70TD(1,
= 24) = 12.90,
participants
Mean
6.30, (SD)
[79.53, Pin
95% = 0.001,
the
CI su
118.27
Table 2. Comparison VCI
1,(F24)
(1, =247) = 6.30,
19.48, PP == 0.01,
P = 0.001, Partial
Partial
0.01, Partial Partial
ŋ2
ŋ2 ==20.20),
ŋ2 = of
= 0.45),
0.20),
theSymbol
0.35),
and mean
Vocabulary
andComprehension
Concepts
Symbol
(SD) of (F
Search
(F
Search
the
(1,(F(1,indices
24)
(F (1,247)
(1,=24) =and
19.48,
24) Pthe
==6.30,
12.90,
4.32,
subtests
=P0.001,
PP=Not
=0.01,
= 0.04, of the WISC-4
Partial
0.001,
Partial
Partial
86.70ŋ2ŋ2= =
Matched between
0.20),
20.15). and
=[79.53,
0.45), the
Symbolhigh- Search
93.86] (F(F
Comprehension
(19.18) 118.27 (1,
(1,24)
24)=
(10.99)
[114.16,
functioning ASD and TD groups with FSIQ VCI
matched and not matched (ASD (n = 13) and TD (n = 13)), and
.35), Vocabulary (F (1, 247) 4.32, P = 0.04, Partial
= 6.30, P = 0.01, Partial Partial ŋ2 =
ŋ2 = 0.20), 0.15).
ŋ2 =and0.35),SymbolVocabulary Search(F (F(1,(1,247)
Matched 24) == 6.30, P = 0.01, Partial ŋ2 = 0.20), and Symbol S
(19.18) 93.86] (10.99) 122.37]
Partial ŋ2 = 0.15). without (ASD (n = 30) and TD4.32, (n = P30)) FSIQ Partial
= 0.04, matched. ŋ2 = 0.15).
mean (SD) of the indices and the subtests Table
of the
2. Comparison
WISC-4 between of thethemean
high-(SD) of the indices and the subtests of the WISC-4 between the high-
Intergroup
ASD TD
oups with FSIQ
Comparison matched
of the and
(SD)not
meanIndices of
andmatched
the functioning
(ASD
theindices
Subtests of (nthe
and = 13)
ASD and
and
subtestsTD TD
Table
of (n groups
=2.WISC-4
the 13)), and
with
Comparison FSIQ
of matched
between the
themean and
high- notofmatched
(SD) (ASD
the indices (n the
and =Comparisons
13)subtests
and TDof(nthe
= 13)), and betw
WISC-4
FSIQ
D (n
ng ASD= 30))
andFSIQ matched.
TD groups with FSIQthe matched
WISC-4without
and not (ASD (n = 30)
matched and(n
(ASD Mean
TD
functioning (nASD
= 13) =and
30)) FSIQ
and
TD TD
(n matched.
=groups
13)), and
with FSIQ matched and not matched (ASD (n = 13) and TD
95% CI Mean (SD) 95% CI P-Value Cohens d
(SD)
ASD (n = 30) and TD (n = 30)) FSIQ matched. without (ASD
Intergroup
(n = 30) and TD (n = 30)) FSIQ matched. Intergroup
ASD TD Not 86.70 [79.53, ASD118.27 [114.16, TD
VCI Indices and the Subtests Comparisons 0.001 2.01 Comparisons
FSIQ Matchedof (19.18)
FSIQ 93.86] Intergroup
(10.99) 122.37]
ASD the WISC-4 IndicesTD ASD TD
e Subtests of Mean 95% CI Mean (SD) 95% CI and the Subtests
P-Value Cohens dofMean
Comparisons
95% CI Mean (SD) 95% CI P-Value Cohens d
ISC-4 (SD) FSIQ Mean the WISC-4 (SD) FSIQ Mean
Not 86.70 [79.53, (SD) 95%
118.27 CI Mean
[114.16, (SD) 95%
Not CI P-Value
86.70 Cohens d
[79.53, 95% CI Mean (SD) 95% CI P
VCI 0.001 2.01 (SD) 118.27 [114.16,
0.001 2.01
Matched (19.18) Not 93.86] 86.70 (10.99) [79.53, 122.37]
118.27 Matched
[114.16, (19.18) Not 93.86] 86.70 (10.99) [79.53, 122.37]118.27 [114.16,
CI VCI 0.001 2.01
Matched (19.18) 93.86] (10.99) 122.37] Matched (19.18) 93.86] (10.99) 122.37]
Behav. Sci. 2019, 9, 20 7 of 16

Table 2. Comparison of the mean (SD) of the indices and the subtests of the WISC-4 between the
high-functioning ASD and TD groups with FSIQ matched and not matched (ASD (n = 13) and TD
(n = 13)), and without (ASD (n = 30) and TD (n = 30)) FSIQ matched.

Indices and the Intergroup

ASD TD
Subtests of the FSIQ Comparisons
WISC-4 Mean (SD) 95% CI Mean (SD) 95% CI P-Value Cohens d
[114.16,
Not Matched 86.70 (19.18) [79.53, 93.86] 118.27 (10.99) 0.001 2.01
VCI 122.37]
[106.11,
Matched 100.31 (14.98) [91.25, 109.36] 111.38 (8.71) 0.030 0.90
116.65]
Not Matched correlation
8.20 (4.02) coefficient
[6.69, 9.70] was
11.96used
(3.59) to examine
[10.62, 13.30]the relationship
0.001 between the FSIQ, the Wec
0.98
Similarities
Matched indices
9.92 (4.36)and subtests,
[7.28, 12.56] and 9.53
the(2.84)
GARS-2[7.81,and11.25]
CPRS-RS subtests. -The level of statistical signi
P > 0.05

Vocabulary
Not Matched P8.43
< 0.05.
(4.44) The[6.77,
mean of each
10.09] index or subtest
15.83b(1.93) in the groups
[15.11, 16.55] 0.001 was used to draw the group an
2.16
Matched 12.15 (3.13)
data. [10.26, 14.04] 14.84 (2.26) [13.47, 16.21] 0.019 0.98
Not Matched 5.03 (3.46) [3.73, 6.32] 11.26 (3.00) [10.14, 12.38] 0.001 2.23
Comprehension
Matched 7.00 (3.26)
correlation2.4. Ethics [5.02,
coefficient
8.97] 10.69 (1.75)
was used to examine the[109.07,
[9.63, 11.75] 0.001 1.41
relationship between the FSIQ, the Wechsler inte
Not Matched 93.20 (11.87) [88.76, 97.63] 113.23 (11.14) 0.001
PRI indices and subtests,
This studyand the GARS-2
received an and
ethics CPRS-RS
code 117.39]subtests. The level1.74of statistical
(ir.iums.rec.1394.9211363204) fromsignificance
the Ethics wC
P < 0.05. The
Matched 100.23mean
(11.30) of each
[93.39, index
107.06] or subtest
103.54 (7.25) in the
[99.15,groups
107.92] was
P > used
0.05
University of Medical Sciences. Written consent was obtained from the parents for to draw
- the group analysis plo
Not data.
Matched 8.93 (3.70) [7.55, 10.31] 12.10 (3.26) [10.88, 13.31] 0.001 0.90
Block Design participation in the study, and verbal consent was obtained from the participants themse
Matched 9.61 (3.61) [7.42, 11.80] 9.46 (2.22) [8.11, 10.80] P > 0.05 -
Not Matched 5.66 (2.30) [4.80, 6.52] 10.86 (1.79) [10.19, 11.53] 0.001 2.52
Picture Concepts 2.4. Ethics3. Results
Matched 7.15 (1.51) [6.23, 8.07] 10.00 (1.73) [8.95, 11.04] 0.001 1.75

Matrix Reasoning
This study
Not Matched received
10.76 (2.67) [9.76,an ethics12.73
11.76] code (ir.iums.rec.1394.9211363204)
(2.65) [11.74, 13.72] 0.006 from
0.74 the Ethics Committee
3.1. Intergroup Comparison
University12.15
Matched of (2.64)
Medical Sciences.
[10.55, 13.74] Written
11.23 (2.27) consent was obtained
[9.85, 12.60] P > 0.05 from - the parents for their ch
Not participation
Matched 85.20 in thetwo
(15.43)
The study,
[79.43,
groups and were
90.96] verbal
99.56
notconsent
(9.78) was 103.21]
[95.91,
significantly obtained
different from
0.001 the participants
in terms of age and themselves.
1.11 gender, but they
WMI
Matched 93.61 (15.58) [84.20, 103.03] 95.61 (7.77) [90.91, 100.31] P > 0.05 -
the FSIQ. The presence of significant FSIQ differences among the groups presents
Not 3.
Matched 7.60 (4.06)
Resultsdilemma. [6.08, 9.11] 9.46 (2.44) [8.55, 10.37] 0.035 0.55
Digit Span Some authors have argued that FSIQ should not be included as a covariate be
Matched 9.76 (3.44) [7.68, 11.85] 8.92 (2.39) [7.47, 10.37] P > 0.05 -
disorders may directly cause mild FSIQ deficits compared to individuals without psychia
Letter–Number Not Matched 6.10 (2.72) [5.08, 7.11] 9.30 (2.33) [8.42, 10.17] 0.001 1.25
Sequencing
3.1. Intergroup
that Comparison
controlling for the 8.46 FSIQ removes a portion of the variance that is associated
Matched 6.92 (3.30) [4.92, 8.91] (2.02) [7.23, 9.68] P > 0.05 -
psychiatric
The two
Not Matched 93.00 groups
(14.49) disorders.
were In contrast,
not significantly
[87.58, 98.41] some
115.43 (9.71) different authors
in terms
[11.80, 119.06] have argued
of age and
0.001 that the
1.81 gender, but FSIQ
theyshould
were inb
PSI
Matched Because
the FSIQ.102.62
The(13.89) this[94.21,
presence issue
of11.01] has 108.08
not been
significant FSIQresolved
(5.88) differences
[104.52, definitely,
11.63] P > 0.05 the
among theresults
groups
- are reported
presents both w
a method
Coding
Not dilemma.
Matched controlling
Some authors
7.53 (3.49) for the
have
[6.22, FSIQ
8.83] between
argued (2.23) two
11.20that FSIQ groups.12.03]To
should
[10.36, notmatch
be
0.001 the
included FSIQ
1.25asfor
a the reasons
covariate mentione
because psy
Matched did
9.53 not
(3.59)use the FSIQ
[7.36, 11.71] as a covariate
10.07 (1.75) in statistical
[9.01, 11.13] analysis,
P
disorders may directly cause mild FSIQ deficits compared to individuals without psychiatric disord> 0.05 but, instead,
- the participants with
Not Matched 7.86 (2.37)
were selected [6.98,
from 8.75]the high-functioning
12.33 (2.23) [11.49,ASD
13.16] (n =0.001
13) and TD 1.94(n = 13) groups. The statis
Symbol Search that controlling for the FSIQ removes a portion of the variance that is associated specifical
Matched 9.53 (3.59) [7.36, 11.71] 10.76 (1.42) [9.90, 11.62] 0.048 0.45
psychiatric nodisorders.
significantIn differences
contrast, in somethe FSIQauthors scoreshave between
arguedthe thattwothegroups.
FSIQ should be controll
[108.27,
FSIG
Not Because
Matched 85.63 The
(12.51) comparison
[80.96, 90.31] of the
this issue has not been resolved definitely,11.87WISC-IV
(9.62) index
115,46] and the results are reported the
subtest
0.001 scores 2.35 between bothhigh-function
with and
Matched groups
controlling97.08
for(9.06)without
the FSIQ controlling
between
[91.60, 102.55] two 102.85for
groups.the
(5.04) FSIQ
To showed
match
[99.80, 105.90]the P a
FSIQsignificant difference
> 0.05 for the-reasons mentioned, on combined de
the rese
Not did not use
Matched
of the
WISC-IV
133.17 FSIQ as
(33.12)
subtest
a covariate
[120.80,
scores
14.53] in(Fstatistical
134.43
(10, 49) =analysis,
(32.06)
14.30, P but,
[122.46, = 0.001, Wilk’s
instead,
0.881 the- Lambada
participants = 0.25,
withPartial
an averaŋ2
Age (months) 146.41]
scores
were selected (F (10,
from the48) = 3.08, P = 0.005,ASD
high-functioning Wilk’s (n Lambada
= 13) and =TD 0.60,
(n =Partial ŋ2 = 0.39).
13) groups. TheAnalysis
statisticaloftests
each
[122.46, [103.62,
Matched variables,
146.62
no significant (30.69) as shown
differences in the
151.39] inFSIQ
Table 2, revealed
127.23scores
(39.06) between thatthe
150.84] participants
P > 0.05
two groups. with- ASD were weaker than T

CI: Confidence Interval, ASD:The all

Autism WISC-IV
comparison
Spectrum of index and
the WISC-IV
Disorder, subtest
TD: Typical scores.
index The
and subtest
Development, mean difference
FSIQ:scores
Full-Scalebetween produced maximum effect
the high-functioning
Intelligence siz
ASD
quotient, VCI: Verbal ComprehensionPicturecontrolling
groups without Index, Concept subtest,
PRI: Perceptualfor the and FSIQminimum
Reasoning showed aeffect
Index, WMI: Working sizes for
significant Memory the WMI
Index,
difference on and
PSI: Digit Span
combined subtes
dependent
Processing Speed Index.
of WISC-IV two groups
subtest scoresin term (F (10, of49)FSIQ, the Peffect
= 14.30, = 0.001, of Wilk’s
the groups Lambada were= still
0.25,statistically
Partial ŋ2 = signific
0.74) an
dependent
scores (F (10, 48) = 3.08, variables
P = 0.005, ofWilk’s
WISC-IV subtest=scores
Lambada (F (10,ŋ2
0.60, Partial 15)= =0.39).
3.02,Analysis
P = 0.02, ofWilk’s Lambada
each of the de
3.2. Intragroup Comparison
variables,=as0.66),
shown butin not in index
Table score. Analysis
2, revealed of each of
that participants theASD
with dependent variables,
were weaker thanasTD
shown in
partici
The cognitive profile of those
all WISC-IV that
with participants
index high-functioning
and subtest with ASD
ASD
scores. were
The showed
meanweaker significantly
a significant
difference difference
produced than TD participants
between
maximum effect sizesinforthe
thesuV
the different indices ofPicture Concepts
Concept
intelligence (F (1,
andsubtest,
the FSIQ 24)
and(F =minimum
19.48,
(4, 116) P ==effect
0.001, Partial
3.92, sizes
P for the
= 0.005, ŋ2WMI=20.119).
= 0.45),
and Digit Comprehension
The Span subtest.
post (F After
(1, 24)m
hoc tests revealed the two mean groupsPartial
scoresin ŋ2
of term =
the FSIQ 0.35),
of FSIQ, Vocabulary
(mean the= effect (F
85.63, of (1, 247)
SDthe = 6.30,
groups
= 12.51) andP =
were
WMI0.01,
stillPartial = 85.2, significant on coS
ŋ2
statistically
(mean = 0.20), and Symbol
SD = 15.43) to be at thedependent
same level,4.32,andP =to
variables 0.04,
be Partial
of significantly
WISC-IV ŋ2subtest
= 0.15).
lower scores
than(Fthe (10,PSI15)(mean
= 3.02,=P93, = 0.02,
SD =Wilk’s
14.49)Lambada = 0.33, P
and PRI scores (mean == 93.2,
0.66),SD
but=not in index score. Analysis of each of the dependent variables, as shown in Table 2, r
11.87).
that participantsTable
with2. Comparison
were of the mean (SD) of the than
indices and the subtests of the WISC-4 betw
There was a significant difference in the ASD
TD group weaker
with significantly
varying TD
profiles between participants
the differentin the subtests of
Concepts functioning
(F score
(1, 24)(F= (4, ASD
19.48, and TD groups with FSIQ matched and not matched (ASD (n = 13) and TD
intelligence indices and the FSIQ 116)P== 27.96,
0.001, PPartial
= 0.001, ŋ2 =20.491).
= 0.45),The
Comprehension
post hoc tests(F (1, 24) = 12.90, P
Partial without (ASD (n = 30) and TD (n = 30)) FSIQ matched.
revealed the mean WMI (meanŋ2= =99.56,
0.35),SDVocabulary
= 9.78) to (Fbe (1, 247) = 6.30,lower
significantly P = 0.01,
thanPartial ŋ2 score
the FSIQ = 0.20),
andand
allSymbol Search (F
4.32, P = 0.04, Partial ŋ2 = 0.15).
ASD TD
Indices and the Subtests of
FSIQ
the WISC-4
Table 2. Comparison Meanand the subtests of the WISC-4 between the hi
of the mean (SD) of the indices 95% CI Mean (SD) 95% CI P
(SD)
functioning ASD and TD groups with FSIQ matched and not matched (ASD (n = 13) and TD (n = 13)),
Not 86.70 [79.53, 118.27 [114.16,
without (ASD (n =VCI
30) and TD (n = 30)) FSIQ matched.
Matched (19.18) 93.86] (10.99) 122.37]
Intergro
 2.4. Ethics
This study received an ethics code (ir.iums.rec.1394.9211363204) from the Ethics Committee of Iran
University of Medical Sciences. Written consent was obtained from the parents for their children’s
ent was used to examine the relationship
participation
Behav. Sci. 2019, between
9, 20in the study, andthe FSIQ,consent
verbal the Wechsler intelligence
was obtained from the participants themselves. 8 of 16
s, and the GARS-2 and CPRS-RS subtests. The level of statistical significance was set at
of each index or subtest 3. Results
in the groups was used to draw the group analysis plots of the
of the intelligence indices. The FSIQ (mean = 111.87, SD = 9.62) and PRI (mean = 113.23, SD = 11.14)
3.1.were
scores Intergroup Comparison
at a similar level, and were significantly lower than the VCI (mean = 118.27, SD = 10.99)
score. The PSI scores (mean = 115.43, SD = 9.71) were significantly higher than the FSIQ scores. Table 3
The two groups were not significantly different in terms of age and gender, but they were in terms of
presents the p-values for the comparison of the indices between the ASD and TD groups.
the FSIQ. The presence of significant
eived an ethics code (ir.iums.rec.1394.9211363204) from the EthicsFSIQ Committee
differences of among
Iran the groups presents a methodological
ical Sciences. Writtendilemma.
consent Some authors have
was obtained fromargued that FSIQ
the parents for should not be included as a covariate because psychiatric
their children’s
Table 3. P-value of the repeated measures analysis for the comparison of the WISC-IV indices between
disorders
study, and verbal consent may directly
was obtained fromcause mild FSIQ deficits
the participants themselves.compared to individuals without psychiatric disorders and
the ASD (n = 30) and TD groups (n = 30).
that controlling for the FSIQ removes a portion of the variance that is associated specifically with
psychiatric Index disorders. Group In contrast,FSIQ some authors VCI have argued PRI thatWMIthe FSIQ should
PSI be controlled [27].
Because this issue hasASD not been resolved definitely, the results are reported both with and without
parison controlling for FSIQ the FSIQ between two groups. To match the FSIQ for the reasons mentioned, the researchers
TD
s were not significantlydiddifferent
not use the FSIQ of
in terms as age
a covariate
and gender,in statistical
but they analysis,
were inbut,terms instead,
of the participants with an average FSIQ
were differences
selected from ASD
the high-functioning 0.633 ASD (n
sence of significant FSIQ VCI among the groups presents a =methodological
13) and TD (n = 13) groups. The statistical tests showed
no significant differences TD
hors have argued that FSIQ should not be included as a covariate becausethe
in the FSIQ0.001
scores between two groups.
psychiatric
The
ctly cause mild FSIQ deficits compared comparison of the
ASD
to individuals WISC-IV index and
0.001 psychiatric
without subtest scores
0.068 disorders and between the high-functioning ASD and TD
PRI
groups without
r the FSIQ removes a portion of the variance controlling for the
TD that is0.236 FSIQ showed
associated a significant
specifically with
0.022 difference on combined dependent variable
rs. In contrast, some of WISC-IV
authors have subtest
argued scores
that(Fthe
ASD
(10,FSIQ
49) = should
0.839
14.30, Pbe = 0.001, Wilk’s[27].
0.682controlled 0.006
Lambada = 0.25, Partial ŋ2 = 0.74) and index
scores (F (10, 48)
WMI = 3.08, P = 0.005, Wilk’s
has not been resolved definitely, the results are reported both with and without Lambada = 0.60, Partial ŋ2 = 0.39). Analysis of each of the dependent
variables, as shown in TD 2, revealed
Table 0.001 that participants
0.001 0.001
with ASD were weaker than TD participants in
SIQ between two groups. To match the FSIQ for the reasons mentioned, the researchers
all WISC-IV
Q as a covariate in statistical analysis, index and subtest scores. The mean difference produced 0.018
ASD 0.003 0.119 0.943 maximum effect sizes for the VCI and
PSIbut, instead, the participants with an average FSIQ
Picture Concept subtest, TD and minimum 0.013
the high-functioning ASD (n = 13) and TD (n = 13) groups. The statistical tests showedeffect sizes
0.247 for the
0.191WMI and Digit Span subtest. After matching
0.001
two
ences in the FSIQ scoresFSIQ: groups
between in
theIQ,
Full-Scale term
two of
groups.
VCI: FSIQ, the effect of the groups were still statistically
Verbal Comprehension Index, PRI: Perceptual Reasoning Index, WMI: Working significant
Memoryon combined
dependent
Index, PSI: variables
Processing of
Speed WISC-IV
Index, ASD:subtest scores
Autism
n of the WISC-IV index and subtest scores between the high-functioning ASD and TD (F
Spectrum (10, 15)
Disorder, = 3.02,
TD: P =
Typical 0.02, Wilk’s
Development. Lambada = 0.33, Partial ŋ2
= 0.66), but
ntrolling for the FSIQ showed not in index
a significant score. Analysis
difference on combinedof each of the dependent
dependent variable variables, as shown in Table 2, revealed
that
The participants
analysis of with
the ASD were
intelligence weaker
subtests
scores (F (10, 49) = 14.30, P = 0.001, Wilk’s Lambada = 0.25, Partial ŋ2 = 0.74) and significantly
in the ASD than
group TD showed
index participants in the subtests
a significant of Picture
difference
Concepts
between
08, P = 0.005, Wilk’s Lambada the (F (1,
subtests
= 0.60, 24)
Partial = 19.48,
(F ŋ2
(9, =261)P = 0.001,
= Analysis
0.39). Partial
9.90, P =of0.001, ŋ
each of2the 2 = 0.45),
= 0.255).
dependent Comprehension (F (1, 24) = 12.90,
The post hoc tests revealed the most P = 0.001,
n in Table 2, revealed Partial
statistically
that ŋ2 = 0.35),with
significant
participants Vocabulary
highASD scores (Fin(1,
were the247)
weaker =than
Matrix 6.30, P =participants
Reasoning
TD 0.01,subtest
Partialin ŋ2 = 0.20),
(mean andSD
= 10.76, Symbol Search
= 2.67), while(F (1, 24) =
the
4.32,
nd subtest scores. ComprehensionP = 0.04,
The mean difference Partial
(mean ŋ2 = 0.15).
= 5.03, SD
produced = 3.46) and
maximum effectPicture
sizes for Concepts
the VCIsubtests
and (mean = 5.66, SD = 2.30) were
btest, and minimum effect
at the samesizes forand
level the had
WMIsignificantly
and Digit Span
lower subtest.
scores After
than thematching
other subtests, with the exception of the
Table 2. Comparison of the mean (SD) of the indices and the subtests of the WISC-4 between the high-
m of FSIQ, the effect of the
Letter–Number groups were still
Sequencing statistically
subtest (mean significant
= 6.10, SD =on2.72).
combined
The subtests Block Design (mean = 8.93,
functioning ASD and TD groups with FSIQ matched and not matched (ASD (n = 13) and TD (n = 13)), and
s of WISC-IV subtest scores (F (10, 15) = 3.02, P = 0.02, Wilk’s Lambada
SD = 3.70), Vocabulary (mean = 8.43, SD = 4.44), Similarities (mean= 0.33, Partial ŋ2 = 8.20, SD = 4.02), and Symbol
without (ASD (n = 30) and TD (n = 30)) FSIQ matched.
ndex score. Analysis of each
Search (meanof the dependent
= 7.86, SD = 2.37)variables, as shown inhigher
had significantly Table scores
2, revealed
than Letter–Number Sequencing.
with ASD were weakerThere significantly
was alsothan TD participants
a significant differencein the subtests
between ASD the of Picture
subtests in the TD Intergroup
TDgroup (F (9, 261) = 20.69,
= 19.48, P = 0.001, P
PartialIndicesŋ and2the
= Subtests
0.45), of
Comprehension (F (1, 24) = 12.90, P = 0.001, Comparisons
= 0.001, 2the=WISC-4
0.416). The post hoc FSIQtests revealed that the Vocabulary subtest (mean = 15.83, SD =
Mean
Vocabulary (F (1, 247)
1.93)= had
6.30,the
P =highest
0.01, Partial
meanŋ2 = 0.20),
scores by and Symbol
a significant Search95%(FCI(1, 24)Mean
= (SD) 95% CI
(SD) margin and the Digit Span (mean = 9.46, SD = 2.44)
P-Value Cohens d
al ŋ2 = 0.15). and Letter–Number Sequencing subtests Not (mean86.70= 9.30, SD = 2.33) had
[79.53, the most statistically
118.27 [114.16, significant
VCI 0.001 2.01
low mean scores. The subtests Matched (19.18) (mean
of Symbol Search 93.86] (10.99)
= 12.33, SD 122.37]
= 2.23) and Matrix Reasoning
rison of the mean (SD) of the indices and the subtests of the WISC-4 between the high-
(mean = 12.73, SD = 2.65) had significantly higher scores than the Coding (mean = 11.20, SD = 2.23) and
D and TD groups with FSIQ matched and not matched (ASD (n = 13) and TD (n = 13)), and
Picture Concepts subtests (mean = 10.86, SD = 1.79). Table 4 presents the p-value for the comparison of
= 30) and TD (n = 30)) FSIQ matched.
the subtests between the two groups.
Intergroup
ASD TD
ests of Comparisons
FSIQ
Mean
95% CI Mean (SD) 95% CI P-Value Cohens d
(SD)
Not 86.70 [79.53, 118.27 [114.16,
0.001 2.01
Matched (19.18) 93.86] (10.99) 122.37]
Behav. Sci. 2019, 9, 20 9 of 16

Table 4. P-value of the repeated measures analysis for the comparison of the WISC-IV subtest score
between the ASD (n = 30) and TD groups (n = 30).

Letter–Number Sequencing
Matrix Reasoning
Picture Concepts
Comprehension

Symbol Search
Block Design
Similarities

Digit Span
Vocabulary
Subtest

Coding
Group

ASD
Similarities
TD
ASD 0.734
Vocabulary
TD 0.001
ASD 0.001 0.001
Comprehension
TD 0.507 0.001
ASD 0.434 0.611 0.001
Block Design
TD 0.828 0.001 0.456
ASD 0.004 0.001 0.314 0.001
Picture Concepts
TD 0.111 0.001 0.148 0.061
Matrix ASD 0.005 0.005 0.001 0.006 0.001
Reasoning
TD 0.250 0.001 0.069 0.279 0.001
ASD 0.478 0.372 0.007 0.113 0.009 0.001
Digit Span
TD 0.004 0.001 0.003 0.001 0.015 0.001
Letter–Number ASD 0.005 0.013 0.157 0.003 0.447 0.001 0.061
Sequencing
TD 0.001 0.001 0.001 0.001 0.005 0.001 0.780
ASD 0.515 0.322 0.007 0.139 0.004 0.001 0.940 0.064
Coding
TD 0.260 0.001 0.475 0.094 0.473 0.005 0.004 0.002
ASD 0.660 0.474 0.001 0.128 0.001 0.001 0.708 0.006 0.533
Symbol Search
TD 0.603 0.001 0.192 0.667 0.003 0.475 0.001 0.001 0.029
ASD: Autism Spectrum Disorder, TD: Typical Development.

Assessing the correlation between the intelligence indices and the subtests of the GARS-2 and
CPRS-RS using Pearson’s correlation coefficient showed a significant negative correlation between
the Communication subtest of the GARS-2 and the VCI, and a positive correlation between the Social
Interaction subtest of the GARS-2 and the WMI. There was also a significant negative correlation
between the cognitive problems/inattention subtest of the CPRS-RS and the VCI and WMI. The FSIQ
showed a significant negative correlation with the Communication subtest of the GARS-2 and the
cognitive problems/inattention subtest of the CPRS-RS.
Assessing the correlation between the intelligence subtests and the GARS-2 and CPRS-RS
subtests using Pearson’s correlation coefficient showed a significant negative correlation between the
Communication subtest of the GARS-2 and the Vocabulary and Comprehension subtests. The Social
Interaction subtest of the GARS-2 also had a significant positive correlation with the Digit Span and
Picture Concepts subtests. The cognitive problems/inattention index of the CPRS-RS had a significant
negative correlation with the Vocabulary, Comprehension, and Digit Span subtests. In addition, the
hyperactivity and ADHD index subscales of the CPRS-RS had a significant negative correlation with
the Comprehension subtest of the WISC-IV (Table 5).
Behav. Sci. 2019, 9, 20 10 of 16

Table 5. Pearson’s correlations between the WISC-IV IQ index and the subtest score and ASD and ADHD symptomatology.

WISC-IV Index and Subtests

Letter–Number Sequencing
Matrix Reasoning
Picture Concepts
Comprehension

Symbol Search
Block Design
Similarities

Vocabulary

Digit Span

Coding
Scale

WMI

FSIQ
VCI

PRI

PSI
GARS Autism Index −0.154 −0.238 −0.320 −0.273 −0.235 0.078 0.010 −0.096 0.243 0.089 0.218 −0.027 −0.044 0.071 −0.118
Stereotyped
−0.191 −0.126 −0.347 −0.252 −0.181 0.045 0.047 −0.057 0.242 0.079 0.209 −0.045 0.062 0.132 −0.131
Behaviors
GARS-2

Communication −0.189 −0.466 ** −0.417 * −0.410* −0.151 −0.230 −0.196 −0.267 −0.072 0.020 −0.041 0.039 −0.345 −0.204 −0.378 *
Social interaction 0.057 −0.071 0.031 −0.001 −0.148 0.363 * 0.113 0.109 0.436 * 0.144 0.393 * −0.020 0.141 0.136 0.224
Total Standard Score −0.154 −0.247 −0.329 −0.281 −0.208 0.083 0.007 −0.080 0.277 −0.105 0.252 −0.022 0.056 0.063 −0.114
Rank Percent −0.149 −0.233 −0.316 −0.264 −0.254 0.053 −0.016 −0.127 0.196 0.076 0.177 −0.032 0.036 0.078 −0.129
Oppositional −0.146 −0.065 −0.191 −0.155 −0.112 0.044 −0.047 −0.071 −0.003 −0.084 −0.044 0.118 0.158 0.215 0.014
CPRS-R:S

Cognitive
−0.345 −0.390 * −0.404 * −0.423 * −0.150 −0.229 −0.334 −0.344 −0.526 ** −0.295 −0.531 ** 0.161 −0.108 −0.068 −0.434 *
Problems/Inattention
Hyperactivity −0.171 −0.147 −0.415 * −0.282 −0.126 −0.015 −0.025 −0.089 −0.089 −0.108 −0.128 0.131 0.134 0.230 −0.097
ADHD Index −0.264 −0.162 −0.383 * −0.296 −0.156 −0.038 −0.196 −0.213 −0.241 −0.192 −0.262 0.331 −0.054 0.207 −0.178
GARS-2: Gilliam Autism Rating Scale-2, CPRS-R:S: Conners’ Parent Rating Scale-Revised (Short), VCI: Verbal Comprehension Index, PRI: Perceptual Reasoning Index, WMI: Working
Memory Index, PSI: Processing Speed Index, FSIQ: Full Scale IQ, ADHD: Attention-deficit/hyperactivity disorder, p < 0.05*, p < 0.01**.
 3.3. Individual Comparison
The profiles of the ASD and TD groups were plotted according to the mean scores (one graph
per group, n = 2). Concomitantly, mean intelligence indices were plotted for every subject in both
groups (n = 60). To show dispersion from the general group pattern, the profile of each subject was
Behav. Sci. 2019,
then9, compared
20 to the profile of his or her own group. This indicated that the profiles of 15 subjects 11 of 16
(50%) in the ASD group and 21 subjects (70%) in the TD group were similar to the profiles obtained
from their own groups (profiles of the indices and the subtests of intelligence in the ASD and TD
3.3. Individual
groupsComparison
are shown in Figures 1a and 1b in the Appendix). Remarkably, the profile of seven subjects
(23%) in the TD group was similar to the profile of the ASD group, but the profile of none of the
The profiles
people inofthetheASDASD and
group TD
was groups
similar were
to that plotted
in TD group.according
A graph of the to the mean and
individual scores
group(one graph
n = 2).was
per group,subtests Concomitantly,
also plotted. The mean
results intelligence
showed that the indices were
dispersion plotted
in the profilesfor every
of the subject
subtests in in both
groups (n those
= 60).inToboth groups
show was so high from
dispersion that nothe
onegeneral
had a similar profile
group to the profile
pattern, in his own
the profile group subject
of each or was
the other group.
then compared to the profile of his or her own group. This indicated that the profiles of 15 subjects
The individual analysis showed that 50% of the subjects in the ASD group were average in terms
(50%) in the ASD
of the PRIgroup
and PSI.and
The 21
WMI subjects
and VCI(70%) in the
scores were TDaverage
below groupinwere70.1%similar
and 56.7%toofthe
theprofiles
subjects, obtained
from theirrespectively
own groups (profiles
(according to theof the analysis
group indicesscores,
and thethemean
subtests
PRI and ofPSI
intelligence in theinASD
scores were average the and TD
groups areASD shown group,
in and the WMI
Figure and
1a,b). VCI scores were
Remarkably, thebelow average).
profile of seven subjects (23%) in the TD group was
The individual analysis of the TD subjects showed that the VCI, PRI, and PSI scores were above
similar to the profile of the ASD group, but the profile of none of the people in the ASD group was
average in 70%, 56.7%, and 73.4% of the subjects, respectively, while the WMI was average and below
similar to that in TD
in 86.7% group.
of them A graph
(according of the
to the groupindividual
analysis ofand group
the TD subtests
people, was
all of the also were
indices plotted.
aboveThe results
showed that the dispersion
average except for thein the which
WMI, profiles
wasof the subtests
average). In otherin thosemore
words, in both groups
than 50% of thewas so high
subjects in that no
both groups
one had a similar showed
profile to athe
similar group
profile inlevel of analysis
his own groupin terms
or theof other
the intelligence
group. indices.
ASD TD

140
A

120

100

80

60

40

20
Behav. Sci. 2018, 8, x FOR PEER REVIEW 2 of 16
0
VCI PRI WMI PSI

ASD TD

B
18

16

14

12

10

8
Behav. Sci. 2018, 8, x; doi: FOR PEER REVIEW www.mdpi.com/journal/behavsci
6

Figure 1. Profiles of the indices and subtests of the WISC-IV in the ASD and TD groups are shown.
(A) Profiles of the
Figure indicesofof
1. Profiles thethe WISC-IV
indices in theofASD
and subtests and TDingroups.
the WISC-IV (B) TD
the ASD and Profiles
groupsof the
are subtests of the
shown.
WISC-IV in(A)the
Profiles
ASDofandthe indices of the WISC-IV in the ASD and TD groups. (B) Profiles of the subtests of
TD groups.
the WISC-IV in the ASD and TD groups.

4. Discussion

4.1. Intergroup Comparison

The comparison of the two groups showed a significant difference in all of the indices and
subtests of the WISC-IV. Although it was not possible to match all the participants of the two groups
Behav. Sci. 2019, 9, 20 12 of 16

The individual analysis showed that 50% of the subjects in the ASD group were average in terms
of the PRI and PSI. The WMI and VCI scores were below average in 70.1% and 56.7% of the subjects,
respectively (according to the group analysis scores, the mean PRI and PSI scores were average in the
ASD group, and the WMI and VCI scores were below average).
The individual analysis of the TD subjects showed that the VCI, PRI, and PSI scores were above
average in 70%, 56.7%, and 73.4% of the subjects, respectively, while the WMI was average and below
in 86.7% of them (according to the group analysis of the TD people, all of the indices were above
average except for the WMI, which was average). In other words, more than 50% of the subjects in
both groups showed a similar group level of analysis in terms of the intelligence indices.

4. Discussion

4.1. Intergroup Comparison

The comparison of the two groups showed a significant difference in all of the indices and
subtests of the WISC-IV. Although it was not possible to match all the participants of the two groups
in terms of their FSIQ, the comparison of the results of 13 ASD subjects were matched for FSIQ with
13 TD individuals showed a difference between the two groups in the VCI and subtests of Picture
Concepts, Comprehension, Vocabulary, and Symbol Search. Although this small sample size was not
representative of all of the differences, those significant differences between the two groups, which are
discernible despite the small sample population, indicate a large difference in the two groups.

4.2. Intragroup Comparison

Based on the findings of this study on the cognitive profile of the people with high-functioning
ASD, the mean PRI and PSI were significantly higher than the mean WMI and FSIQ. A number
of studies have shown the higher competence of the PRI compared to the other indices [8,10,11].
This index was higher than the WMI in a study by Oliveras-Rentas et al. [9], but, in contrast to previous
studies, the PSI was higher than the WMI and the FSIQ in both groups in this study. Weaknesses
have been reported in the PSI in people with high-functioning ASD in studies conducted using the
WISC-IV [7–10] and WISC-III [28]. This study does not clarify why the subjects’ processing speeds
were higher than their FSIQ and WMI scores in both groups, but, considering the impact of culture on
people’s performance in intelligence tests [29,30], background differences may be one of the reasons.
In line with previous research findings, the WISC-IV subtests’ analysis showed a good competence
in Matrix Reasoning [7–11] and weaknesses in Comprehension [7,9,10]. Unlike in Wechsler’s
preliminary study [7], or the studies by Mayes and Calhoun [8] and Nader, Jelenic, and Soulieres [10],
the Picture Concepts subtest was one of the weakest subtests in the present study. In other words,
in this study, of the two motor-free and untimed subtests of the PRI, Matrix Reasoning received
the highest total score and Picture Concepts the lowest. The Matrix Reasoning subtest is relatively
culture-free, and its high scores indicate good processing of visual information and nonverbal abstract
reasoning skills. The weaknesses of the ASD group in the Picture Concepts subtest might represent
a weakness in nonverbal concept forming and rigid thought processes. The ability of ASD people
to engage in abstract reasoning might create innovative and unconventional relationships between
pictures [6], and might have thus reduced the Picture Concepts score.
In this study, the PSI and VCI scores were higher than the FSIQ, the VCI score was higher than the
PRI score, and the WMI score was lower than the scores of all of the indices in the TD group. A study
by Nader [11] found no statistically significant differences between the index scores of TD people,
while in a study by Nader, Jelenic, and Soulieres [10], the WMI score was significantly lower than
the FSIQ, PRI, and VCI scores. In the analysis of the subtests, the Vocabulary subtest had the highest
score and Digit Span and Letter-Number subtests had the lowest scores, just as in a study by Nader et
al. [10].
Behav. Sci. 2019, 9, 20 13 of 16

This study found a negative correlation between the Communication subtest of the GARS-2,
the FSIQ, and the VCI, Vocabulary, and Comprehension subtests of the WISC-IV. That is, the more
the parents reported communication problems, the weaker the children scored on the FSIQ, VCI,
and Vocabulary and Comprehension subtests. Two other studies also found a negative correlation
between verbal skills in the WISC and communication problems in the ADOS [9,31]. There was also
a negative correlation between communication problems in the GARS-2 and FSIQ in the present
study, while the FSIQ had no significant relationship with the Social Interaction score in the GARS-2.
Kenworthy et al. [32] also concluded that the FSIQ can predict more scores in the field of communication
and everyday skills, but there are no relationships between the FSIQ and social skills. Other studies
have also suggested the impossibility of using the FSIQ for predicting adaptive behaviors, especially
social skills, in people with high-functioning ASD [33–35]. According to Klein et al. [36], social skills
have been severely impaired in people with high-functioning ASD, and their scores cannot be predicted
based on the FSIQ.
According to the results, there is a significant positive correlation between the Social Interaction
subscale of the GARS-2 and the WMI and the Picture Concepts and Digit Span subtests, which raises
the question of why a cognitive measure such as the WMI and the subtests of Picture Concepts and
Digit Span have a positive correlation with the Social Interaction subscale of the GARS. According
to a study by Joseph, Tiger-Flusberg, and Lord [37], children with higher nonverbal skills obtain
significantly lower scores in the social function of the ADOS, regardless of their overall abilities and
verbal skills. Another study in 2009 showed that, although verbal skills indicate better performance
in the ASD group, the difference between verbal and nonverbal intelligence is more related to the
manifestation of social interactions [31]. Based on the comparison of the results of this and previous
studies, the emergence of such patterns is not unexpected.
Just as in the study by Oliveras-Rentas et al. [9], there was no significant relationship between the
ADHD index of the CPRS-RS and the indices of the WISC-IV in the ASD people, while there was a
significant negative correlation between the ADHD and hyperactivity indices and the Comprehension
subtest of the WISC-IV. Bruce et al [38] found a negative correlation between verbal Comprehension
and ADHD symptoms in ADHD people. The cognitive problems/inattention index of the CPRS-RS
showed a significant negative correlation with the FSIQ and the Vocabulary, Comprehension, and Digit
Span subtests of the WISC-IV in the ASD group. Naglieri et al. [39] found no relationship between
the cognitive problems/inattention index of Conners’ Parent Rating Scales-Revised (Long-Form)
and the subtests of the WISC-III, but a significant relationship was observed between cognitive
problems/inattention in the Conners’ Teachers Rating Scales-Revised (Long-Form) and the FSIQ and
the VCI and freedom from the distractibility subtests of the WISC-III.

4.3. Individual Comparison

In the analysis of the pattern of the IQ indices, 21% of the subjects in the TD group showed the
same pattern as the subjects in the ASD group. This finding probably confirms the results of studies
conducted to assess the features of ASD in TD people [40], but in the present study, there was no TD
group profile in the ASD subjects. Perhaps plotting profiles might serve as an assistive factor in the
final clinical decision for ruling out ASD. In other words, perhaps ASD might be ruled out in a subject
who is an ASD suspect and has a profile of the TD subjects.

5. Limitation
The findings of the present study should be considered with respect to its limitations. Findings
might be different using a larger sample size with a FSIQ matched TD control group at baseline.
One of the limitations was that lack of prior research and findings on the topic from other developing
countries made it impossible to lay a foundation for understanding the research problem we were
investigating. We propose similar studies be done by other researchers to understand the possibility
Behav. Sci. 2019, 9, 20 14 of 16

that these cognitive profile differences might be attributable to ADHD or other learning differences,
rather than ASD?

6. Conclusions
This study examined the cognitive profile of people with high-functioning ASD and compared
it to that of those with TD, and examined the relationship between this profile and ASD and ADHD
symptoms. The findings can help compare the results between Eastern and Western societies.
There were many similarities between the results of this research and previous studies; however,
a high processing speed in both the ASD and TD groups was the distinguishing point between this
study and similar studies, which may be attributed to cultural factors. Further studies with larger
sample sizes are required for examining this hypothesis. According to the results of this study, it may
be possible to use drawing plots of the intelligence indices as an assistive factor in the final clinical
decision for ruling out the chance of ASD.

Author Contributions: Conceptualization, A.R. and S.A.S.; Methodology, M.K. and A.R. and S.A.S.; Software,
M.K. and A.R.; Validation, A.R. and S.A.S.; Formal Analysis, M.K. and A.R.; Investigation, A.R.; Resources, A.R.
and S.A.S. and A.M. and F.R.K.; Data Curation, M.K. and A.R. and S.A.S.; Writing-Original Draft Preparation, A.R.
and S.A.S.; Writing-Review & Editing, A.R. and S.A.S.; Visualization, A.R., S.A.S.; Supervision, S.A.S.; Project
Administration, A.R. and S.H. and S.S. and A.M.; Funding Acquisition, B.V.-G.
Funding: This article was one part of the first author’s Ph.D. dissertation in speech and language pathology
supported by Iran University of Medical Sciences (ir.iums.rec.1394.9211363204).
Acknowledgments: The authors would like to express their gratitude to all those who helped conduct this
research, especially Ms. Elham Shirazi, Ms. Fatemeh Rezvani and Mr. Daryoosh Nikbakht and all of the children
and families involved in the study.
Conflicts of Interest: The authors report no conflict of interest.

References
1. Association, A.P. Diagnostic and Statistical Manual of Mental Disorders (DSM-5® ); American Psychiatric Pub:
Arlington, VA, USA, 2013.
2. Kim, H.U. Autism across cultures: Rethinking autism. Disabil. Soc. 2012, 27, 535–545. [CrossRef]
3. Samadi, S.A.; McConkey, R. Screening for autism in Iranian preschoolers: Contrasting m-chat and a scale
developed in Iran. J. Autism Dev. Disord. 2015, 45, 2908–2916. [CrossRef] [PubMed]
4. Baio, J.; Wiggins, L.; Christensen, D.L.; Maenner, M.J.; Daniels, J.; Warren, Z.; Kurzius-Spencer, M.;
Zahorodny, W.; Rosenberg, C.R.; White, T. Prevalence of autism spectrum disorder among children aged 8
years—Autism and developmental disabilities monitoring network, 11 sites, United States, 2014. MMWR
Surveill. Summ. 2018, 67, 1–23. [CrossRef] [PubMed]
5. Ghaziuddin, M.; Mountain-Kimchi, K. Defining the intellectual profile of asperger syndrome: Comparison
with high-functioning autism. J. Autism Dev. Disord. 2004, 34, 279–284. [CrossRef] [PubMed]
6. Groth-Marnat, G. Handbook of Psychological Assessment; John Wiley & Sons: Hoboken, NJ, USA, 2009.
7. Wechsler, D. Wechsler Intelligence Scale for Children-WISC-IV; Psychological Corporation: San Antonio, TX,
USA, 2003.
8. Mayes, S.D.; Calhoun, S.L. WISC-IV and WIAT-II profiles in children with high-functioning autism. J. Autism
Dev. Disord. 2008, 38, 428–439. [CrossRef] [PubMed]
9. Oliveras-Rentas, R.E.; Kenworthy, L.; Roberson, R.B.; Martin, A.; Wallace, G.L. WISC-IV profile in
high-functioning autism spectrum disorders: Impaired processing speed is associated with increased
autism communication symptoms and decreased adaptive communication abilities. J. Autism Dev. Disord.
2012, 42, 655–664. [CrossRef] [PubMed]
10. Nader, A.-M.; Jelenic, P.; Soulières, I. Discrepancy between WISC-III and WISC-IV cognitive profile in autism
spectrum: What does it reveal about autistic cognition? PLoS ONE 2015, 10, e0144645. [CrossRef] [PubMed]
11. Nader, A.-M.; Courchesne, V.; Dawson, M.; Soulières, I. Does WISC-IV underestimate the intelligence of
autistic children? J. Autism Dev. Disord. 2016, 46, 1582–1589. [CrossRef] [PubMed]
Behav. Sci. 2019, 9, 20 15 of 16

12. Freeth, M.; Sheppard, E.; Ramachandran, R.; Milne, E. A cross-cultural comparison of autistic traits in the
UK, India and Malaysia. J. Autism Dev. Disord. 2013, 43, 2569–2583. [CrossRef]
13. Norbury, C.F.; Sparks, A. Difference or disorder? Cultural issues in understanding neurodevelopmental
disorders. Dev. Psychol. 2013, 49, 45. [CrossRef]
14. Lord, C.; Rutter, M.; DiLavore, P.C.; Risi, S. Autism Diagnostic Observation Schedule-WPS (ADOS-WPS);
Western Psychological Services: Angeles, CA, USA, 1999.
15. Leitner, Y. The co-occurrence of autism and attention deficit hyperactivity disorder in children—What do we
know? Front. Hum. Neurosci. 2014, 8, 268. [CrossRef] [PubMed]
16. Mayes, S.D.; Calhoun, S.L. WISC-IV and WISC-III profiles in children with ADHD. J. Atten. Disord. 2006, 9,
486–493. [CrossRef] [PubMed]
17. San Miguel Montes, L.E.; Allen, D.N.; Puente, A.E.; Neblina, C. Validity of the WISC–IV spanish for a
clinically referred sample of hispanic children. Psychol. Assess. 2010, 22, 465. [CrossRef] [PubMed]
18. Sadeghi, A.; Rabiee, M.; Abedi, M.R. Validation and reliability of the wechsler intelligence scale for
children-IV. Dev. Pscychol. 2010, 7, 377–386.
19. Samadi, S.A.; McConkey, R. The utility of the gilliam autism rating scale for identifying Iranian children
with autism. Disabil. Rehabil. 2014, 36, 452–456. [CrossRef] [PubMed]
20. Shahrivar, Z.; Tehrani-Doost, M.; Pakbaz, B.; Rezaie, A.; Ahmadi, F. Normative data and psychometric
properties of the parent and teacher versions of the strengths and difficulties questionnaire (SDQ) in an
Iranian community sample. J. Res. Med Sci. 2009, 14, 69–77. [PubMed]
21. Ghanizadeh, A.; Izadpanah, A.; Abdollahi, G. Scale validation of the strengths and difficulties questionnaire
in Iranian children. Iran. J. Psychiatry 2007, 2, 65–71.
22. Conners, C.K. Conners’ Rating Scales—Revised: User’s Manual; Multi-Health Systems, Incorporated: North
Tonawanda, NY, USA, 1997.
23. Khushabi, K.; Pour-Etemad, H.; Mohammadi, M.; Mohammadkhani, P. The prevalence of adhd in primary
school students in Tehran. Med. J. Islamic Repub. Iran (MJIRI) 2006, 20, 147–150.
24. Shahaeian, A.; Shahim, S.; Bashash, L.; Yusefi, F. Standardization, reliability and factor analysis of Iranian
version of the conners’ parent and teacher rating scales—Revised: Short form for children 6 to 11 years old
in Shiraz. Q. J. Psychol. Stud. 2006, 3, 97–120.
25. Corbett, B.A.; Constantine, L.J.; Hendren, R.; Rocke, D.; Ozonoff, S. Examining executive functioning in
children with autism spectrum disorder, attention deficit hyperactivity disorder and typical development.
Psychiatry Res. 2009, 166, 210–222. [CrossRef]
26. Cohen, B.H. Explaining psychological statistics. Psyccritiques 1996, 41, 1224–1225.
27. Willcutt, E.G.; Pennington, B.F.; Boada, R.; Ogline, J.S.; Tunick, R.A.; Chhabildas, N.A.; Olson, R.K. A
comparison of the cognitive deficits in reading disability and attention-deficit/hyperactivity disorder.
J. Abnorm. Psychol. 2001, 110, 157. [CrossRef] [PubMed]
28. Calhoun, S.L.; Mayes, S.D. Processing speed in children with clinical disorders. Psychol. Sch. 2005, 42,
333–343. [CrossRef]
29. Walker, A.J.; Batchelor, J.; Shores, A. Effects of education and cultural background on performance on
WAIS-III, WMS-III, WAIS-R and WMS-R measures: Systematic review. Aust. Psychol. 2009, 44, 216–223.
[CrossRef]
30. Boone, K.B.; Victor, T.L.; Wen, J.; Razani, J.; Pontón, M. The association between neuropsychological scores
and ethnicity, language, and acculturation variables in a large patient population. Arch. Clin. Neuropsychol.
2007, 22, 355–365. [CrossRef] [PubMed]
31. Black, D.O.; Wallace, G.L.; Sokoloff, J.L.; Kenworthy, L. Brief report: IQ split predicts social symptoms and
communication abilities in high-functioning children with autism spectrum disorders. J. Autism Dev. Disord.
2009, 39, 1613–1619. [CrossRef] [PubMed]
32. Kenworthy, L.; Case, L.; Harms, M.B.; Martin, A.; Wallace, G.L. Adaptive behavior ratings correlate with
symptomatology and IQ among individuals with high-functioning autism spectrum disorders. J. Autism
Dev. Disord. 2010, 40, 416–423. [CrossRef] [PubMed]
33. Bölte, S.; Poustka, F. The relation between general cognitive level and adaptive behavior domains in
individuals with autism with and without co-morbid mental retardation. Child Psychiatry Hum. Dev. 2002,
33, 165–172. [CrossRef] [PubMed]
Behav. Sci. 2019, 9, 20 16 of 16

34. Kenworthy, L.E.; Black, D.O.; Wallace, G.L.; Ahluvalia, T.; Wagner, A.E.; Sirian, L.M. Disorganization:
The forgotten executive dysfunction in high-functioning autism (HFA) spectrum disorders. Dev. Neuropsychol.
2005, 28, 809–827. [CrossRef] [PubMed]
35. Liss, M.; Harel, B.; Fein, D.; Allen, D.; Dunn, M.; Feinstein, C.; Morris, R.; Waterhouse, L.; Rapin, I. Predictors
and correlates of adaptive functioning in children with developmental disorders. J. Autism Dev. Disord. 2001,
31, 219–230. [CrossRef] [PubMed]
36. Klin, A.; Saulnier, C.A.; Sparrow, S.S.; Cicchetti, D.V.; Volkmar, F.R.; Lord, C. Social and communication
abilities and disabilities in higher functioning individuals with autism spectrum disorders: The Vineland
and the ADOS. J. Autism Dev. Disord. 2007, 37, 748–759. [CrossRef] [PubMed]
37. Joseph, R.M.; Tager-Flusberg, H.; Lord, C. Cognitive profiles and social-communicative functioning in
children with autism spectrum disorder. J. Child Psychol. Psychiatry 2002, 43, 807–821. [CrossRef] [PubMed]
38. Bruce, B.; Thernlund, G.; Nettelbladt, U. ADHD and language impairment. Eur. Child Adolesc. Psychiatry
2006, 15, 52–60. [CrossRef] [PubMed]
39. Naglieri, J.A.; Goldstein, S.; Delauder, B.Y.; Schwebach, A. Relationships between the WISC-III and the
cognitive assessment system with conners’ rating scales and continuous performance tests. Arch. Clin.
Neuropsychol. 2005, 20, 385–401. [CrossRef] [PubMed]
40. Ronald, A.; Happe, F.; Price, T.S.; Baron-Cohen, S.; Plomin, R. Phenotypic and genetic overlap between
autistic traits at the extremes of the general population. J. Am. Acad. Child Adolesc. Psychiatry 2006, 45,
1206–1214. [CrossRef]

© 2019 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access
article distributed under the terms and conditions of the Creative Commons Attribution
(CC BY) license (http://creativecommons.org/licenses/by/4.0/).