J Autism Dev Disord (2014) 44:1959–1971
DOI 10.1007/s10803-014-2071-4
ORIGINAL PAPER
Age Related Differences of Executive Functioning Problems
in Everyday Life of Children and Adolescents in the Autism
Spectrum
Sanne F. W. M. van den Bergh • Anke M. Scheeren
Sander Begeer • Hans M. Koot • Hilde M. Geurts
•
Published online: 23 February 2014
Ó Springer Science+Business Media New York 2014
Abstract Numerous studies investigated executive functioning (EF) problems in people with autism spectrum disorders (ASD) using laboratory EF tasks. As laboratory task
performances often differ from real life observations, the
current study focused on EF in everyday life of 118 children
and adolescents with ASD (6–18 years). We investigated
age-related and individual differences in EF problems as
reported by parents on the Behavioral Rating Inventory
Executive Functions (BRIEF: Gioia et al. in Behavior rating
inventory of executive function. Psychological Assessment
Resources, Odesse 2000), and examined the association with
autism severity. Inhibition problems were mostly found in
the youngest group (6- to 8-year-olds), whereas problems
with planning where more evident for 12- to 14-year-olds as
compared to 9- to 11-year-olds. In a subsample of
Electronic supplementary material The online version of this
article (doi:10.1007/s10803-014-2071-4) contains supplementary
material, which is available to authorized users.
S. F. W. M. van den Bergh H. M. Geurts (&)
Autism Clinic, Research and Development, Dr. Leo Kannerhuis,
Houtsniplaan 1, 6865 XZ Doorwerth, The Netherlands
e-mail: h.m.geurts@uva.nl
S. F. W. M. van den Bergh
e-mail: s.f.w.m.vandenbergh@uva.nl
S. F. W. M. van den Bergh H. M. Geurts
Department of Psychology, Brain and Cognition, University of
Amsterdam, Amsterdam, The Netherlands
S. F. W. M. van den Bergh H. M. Geurts
Dutch Autism and ADHD Research Center (d’Arc),
Weesperplein 4, 1018 XA Amsterdam, The Netherlands
participants meeting the ADOS ASD cut-off criteria the age
related differences in planning were absent, while problems
with cognitive flexibility were less apparent in 15- to
18-year-olds, compared to 9- to 11-, and 12- to 14-year olds.
EF problems surpassing the clinical cutoff were only
observed in 20 % (planning) to 51 % (cognitive flexibility)
of the children and adolescents, and no relation was found
with ASD symptom severity. This underlines the heterogeneous nature of ASD.
Keywords ASD Autism severity Behavioral Rating
Inventory Executive Functions (BRIEF) Development
Executive functioning
Introduction
The theory of executive dysfunction (Damasio and Maurer
1978; Pennington and Ozonoff 1996) suggests that some
autism symptoms might stem from executive functioning
A. M. Scheeren S. Begeer H. M. Koot
Autism Research Amsterdam (ARA), Amsterdam,
The Netherlands
A. M. Scheeren S. Begeer H. M. Koot
EMGO Institute for Health and Care Research, Van der
Boechorststraat 1, 1081 BT Amsterdam, The Netherlands
S. Begeer
School of Psychology, University of Sydney, Brennan
MacCallum Building (A18), Sydney, NSW 2006, Australia
H. M. Geurts
Cognitive Science Center Amsterdam, Nieuwe Achtergracht
129, 1018 WS Amsterdam, The Netherlands
A. M. Scheeren S. Begeer H. M. Koot
Department of Developmental Psychology, VU University,
Amsterdam, The Netherlands
123
1960
(EF) deficits. EF refers to cognitive skills that serve independent, purposive, goal-directed, and self-serving behavior (Lezak et al. 2012). A plethora of studies has shown that
children, adolescents, and adults with autism spectrum
disorders (ASD) encounter problems in executive functioning (e.g., Bramham et al. 2009; Corbett et al. 2009;
Sinzig et al. 2008, for reviews see Hill 2004; Russo et al.
2007). Also, EF deficits relate positively to certain autism
symptoms (e.g., repetitive behavior: de Vries and Geurts
2012; Yerys et al. 2009). However, children and adolescents with ASD do evolve in their EF skills (Christ et al.
2011; Happé et al. 2006; Luna et al. 2007; Pellicano 2010),
and not all individuals with ASD have clinically significant
EF deficits (Hill and Bird 2006; Pellicano et al. 2006;
Towgood et al. 2009). Hence, within the population of
individuals with ASD the development of EF is highly
heterogeneous. Moreover, it is important to examine the
development of different domains of EF separately,
because research in typical development has shown that the
structure of EF becomes more differentiated with age (see
Hughes et al. 2009; Huizinga et al. 2006; Miyake et al.
2000), and different types of EF develop at a different pace
(Best et al. 2009; De Luca et al. 2003; Hughes 2011). Four
domains of EF that are traditionally referred to are inhibition, working memory, cognitive flexibility (i.e., shifting), and planning (Hill 2004; Hughes 2011; Pennington
and Ozonoff 1996). These four domains have been studied
extensively in psychological laboratories in both typically
developing (TD) children and children with ASD of
varying ages. In the current study we aim to describe
developmental EF profiles of children and adolescents with
ASD, based on everyday life observations by parents.
The first component of EF, inhibition, refers to the
ability to voluntarily and deliberately suppress responses.
There are three different types of inhibition (Friedman and
Miyake 2001). The first type, inhibition of prepotent
responses, refers to the suppression of a dominant
response: for example when a child inhibits the response to
speak before it is his/her turn. Resistance to interference,
the second type, refers to ignoring irrelevant information.
This is, for example, required when a child tries to listen to
a teacher but hears other children speak. The last type,
resistance to proactive interference, refers to processes
where previously learned information becomes irrelevant
and interferes with new information. This happens, for
example, when a teacher tries to learn new names of students, but old names interfere. Improvements of inhibition
are evident from childhood to adulthood in typical development (Davidson et al. 2006; Huizinga et al. 2006; Luna
et al. 2004). In ASD, deficits have been observed in each of
the three inhibition domains in some (e.g., Adams and
Jarrold 2011; Christ et al. 2011; Mosconi et al. 2009;
Sinzig et al. 2008; Verté et al. 2005), but not all studies
123
J Autism Dev Disord (2014) 44:1959–1971
(e.g., Christ et al. 2007, 2011; Geurts et al. 2009a). However, this inconsistency might be partly due to differences
in the inhibition type and age group studied, which is
illustrated by two cross-sectional studies. Prepotent
response inhibition seems to improve with age in ASD
(Christ et al. 2011: 8 to 18 years; Luna et al. 2007: 8 to
33 years) and difficulties with resistance to interference
might even fade away after the age of twelve (Christ et al.
2011) as older individuals with ASD showed better skills
than younger individuals (Christ et al. 2011; Luna et al.
2007). The development of proactive interference in people
with ASD seems to parallel the development of TD individuals in the age range of 8–18 years (Christ et al. 2011).
Hence, in people with ASD not all aspects of inhibitory
control might be equally impaired and inhibition problems
might even disappear with age.
In ASD the developmental trajectory of working memory, the second EF component, seems to differ from the
developmental trajectory of inhibition (Luna et al. 2007).
Working memory is the ability to maintain and manipulate
on-line information (Baddeley 1992). A distinction is made
between visual and verbal working memory processes
(Smith et al. 1996). In typical development, both components show a linear development between the age of 4 and
15 years (Gathercole et al. 2004), although improvements of
visual (spatial) working memory processes are seen even
into young adulthood (Luna et al. 2004). Working memory
deficits are commonly observed in ASD (e.g., Minshew and
Goldstein 2001; Steele et al. 2007, but see Edgin and Pennington 2005), especially in spatial working memory (Williams et al. 2005). Compared to TD individuals, working
memory developments seems to be intact in children, but not
in adolescents and adults with ASD (Luna et al. 2007). Thus,
despite parallel development during childhood, working
memory deficits are evident across the lifespan in ASD.
Clear improvements with increasing age are noted in
typically developing children and adolescents for the two
other EF components, cognitive flexibility (i.e., the ability
to intentionally shift thoughts and actions in response to
contextual changes: Monsell 2003) and planning (i.e.,
thinking ahead: Anderson et al. 2001; Davidson et al. 2006;
De Luca et al. 2003; Luciana et al. 2009). Problems in both
cognitive flexibility and planning have been observed in
children with ASD (Hill 2004), but especially for cognitive
flexibility the findings are rather inconsistent (Geurts et al.
2009b). However, cognitive flexibility does seem to
improve in ASD during childhood (Happé et al. 2006) and
development in planning is evident in young children
(Pellicano 2010) as well as in young adolescents (Happé
et al. 2006) with ASD. In sum, little is known about the
developmental patterns of cognitive flexibility and planning in people with ASD, though improvements with age
have been observed for these EF domains.
J Autism Dev Disord (2014) 44:1959–1971
In general, there are three different hypotheses regarding
the development of EF in ASD: (1) the development of EF
in children and adolescents with ASD might be delayed,
but parallel to the typical EF development (Christ et al.
2011; (2) there might be a deviant EF development in ASD
(Ozonoff and McEvoy 1994; or (3) a delayed, but parallel
EF development in childhood might be followed by a
deviant EF development in adulthood (Luna et al. 2007).
All of the above might be true given that different EF
domains follow other developmental trajectories. This
underlines the importance to focus on specific domains
instead of the broad construct of EF.
In everyday lives of individuals with ASD behavioral
problems are observed that seem related to EF. The
Behavioral Rating Inventory Executive Functions (BRIEF:
Gioia 2000), a parent questionnaire that is widely used in
the clinical practise, addresses those everyday behaviors.
Whereas associations are evident with other attention and
behavioral problems (BRIEF: Gioia et al. 2000, 2002;
McAuley et al. 2010), the BRIEF is only minimally related
to laboratory tasks (McAuley et al. 2010). This poses a
problem for the association of BRIEF reports to pure,
actual EF functioning. This has resulted in some authors
arguing that the BRIEF does not actually measure EF
(McAuley et al. 2010). Others (Kenworthy et al. 2008)
question the ecological validity of laboratory tasks as
measurements of EF, arguing that EF related problems in
everyday life are observed in people with ASD, even when
laboratory task performance is intact. Caution should be
made stating that the BRIEF measures actual EF. However,
problems described by the BRIEF take place in a social
context and are relevant in the everyday lives of people
with ASD. Therefore, complementary to what we already
know from laboratory task studies, it is important to study
these behaviors from a developmental perspective and with
respect to individual differences.
Several studies have shown that parents of children and
adolescents with ASD consistently report EF deficits on the
BRIEF (Boyd et al. 2009; Chan et al. 2009; Endedijk et al.
2011; Kalbfleisch and Loughan 2012; Winsler et al. 2007;
Yerys et al. 2009; Zandt et al. 2007). Clinically significant
BRIEF scores on different domains are reported in
35–70 % of children with ASD (Gioia et al. 2002; Kenworthy et al. 2005). In a study with 54 children with ASD
with a mean age of 11 years inhibition problems were
observed in 46 % of the children, working memory problems in 57 %, cognitive flexibility problems in 69 %, and
planning problems in 70 % of these children (Kenworthy
et al. 2005). The high amount of flexibility problems
reported in the BRIEF even discriminates children with
ASD from other clinical groups (Gioia et al. 2002). Furthermore, a higher score on the BRIEF Behavioral Regulation Index (containing the subscales inhibit, shift, and
1961
emotional control) was associated with more symptoms of
communication deficits and restricted repetitive behavior
(Kenworthy et al. 2009) as assessed by the Autism Diagnostic Interview-Revised (ADI-R: Lord et al. 1994), the
Autism Diagnostic Observation Scale- Generic (ADOS-G:
Lord et al. 2000), as well as the Repetitive Behavior ScaleRevised (Boyd et al. 2009). This is in line with the
observed relations between cognitive flexibility and repetitive behavior in ASD in studies using laboratory tasks (de
Vries and Geurts 2012; Yerys et al. 2009). Thus, despite
the weak relations between laboratory task performance
and observed behaviors, autism symptomatology is related
to EF measured with neuropsychological paradigms as well
as to reported EF problems in everyday life. Most studies in
which the BRIEF is used, are focussed on index scores.
However, since the specific EF domains provide more
detailed information than a general EF measure, we focus
on the four well-described and studied domains in the
current study.
The BRIEF has already been used to study developmental trajectories in TD children (Huizinga and Smidts
2011), and recently also in children and adolescents with
ASD (Rosenthal et al. 2013). In a large TD sample (431
boys and 416 girls), BRIEF raw subscale scores of four age
groups (5- to 8-; 9- to 11-; 12- to 15-; and 16- to 18-yearolds) were compared (Huizinga and Smidts 2011). Working memory and flexibility mainly seemed to develop
before the age of 11 years, inhibition appeared to develop
until young adulthood, whereas no development seemed
evident with regard to the planning subscale. The findings
with regard to inhibition fit the conclusions based on laboratory tasks performance. The working memory and
cognitive flexibility findings differ from laboratory tasks
studies, as development was only evident in childhood, and
not during adolescence. The developmental pattern of
planning is in contrast with the observed development of
planning skills during childhood and adolescence based on
EF laboratory tasks performance. In a comparable ASD
study (158 boys and 27 girls, Rosenthal et al. 2013), BRIEF
standardized (T) scores of four age groups (5- to 7-; 8- to
10-; 11- to 13-; and 14- to 18-year-olds) were compared.
Working memory problems were more severe in
14–18 year olds as compared to 6–7 year olds. So, despite
the improved performances on working memory tasks
during childhood in ASD, parent reports indicate an
increase of working memory problems. This is explained
by Rosenthal et al. (2013) by the fact that higher real world
demands during adolescence create a larger discrepancy
with typical development. However, higher demands from
the environment might also produce increasing EF problems in typically developing adolescents, which is not
found in the Huizinga and Smidts study (2011). One
explanation could be that children and adolescents with
123
1962
ASD are more vulnerable to changes in environmental
demands than typically developing peers. Since no agerelated differences were found in other specific EF domains
(e.g. inhibition, cognitive flexibility, and planning), it does
not seem to be the case that reported everyday EF problems
in general increase during adolescence in ASD. As Rosenthal et al. (2013) compared age-related T-scores,
thereby providing information about the relative impairment of participants with ASD compared to a typical age
norm; it is hard to relate these findings to the findings of
Huizinga and Smidts (2011). Trajectories of reported
everyday EF problems in ASD, using raw scores, would
provide the information necessary for these comparisons.
Furthermore, from the sample of Rosenthal et al. (2013) it
is not clear how many children and adolescents actually
showed clinically significant problems. Given the heterogeneity of EF problems in ASD (Hill and Bird 2006; Pellicano et al. 2006; Towgood et al. 2009), this information
seems highly relevant. Although Rosenthal et al. (2013)
controlled for symptomatology, they did not investigate the
unique contribution of symptom severity to everyday EF.
This would be of interest, given the wide range of ASD
severity, and following the theory that some ASD symptoms might stem from EF problems (Pennington and
Ozonoff 1996).
The first and primary aim of the current study is to study
developmental profiles of specific everyday EF domains in
children and adolescents with ASD (6–18 years). For this
purpose we use the same cross-sectional approach as was
taken in TD children and adolescents (Huizinga and Smidts
2011). Hence, we focus on raw scores instead of T-scores.
In line with studies addressing EF in ASD (e.g., Christ
et al. 2011; Happé et al. 2006), we expect to observe agerelated improvements of the subscales inhibition and shift.
We do not expect to find age-related improvements of
working memory and planning, based on the larger BRIEF
working memory discrepancy with TD during adolescence
in ASD (Rosenthal et al. 2013), and the absence of BRIEF
planning improvement in TD (Huizinga and Smidts 2011).
Our secondary aims are: to explore the degree of clinically
relevant EF problems, and to investigate the impact of
ASD severity. Given the heterogeneity of ASD (Hill and
Bird 2006; Pellicano et al. 2006; Towgood et al. 2009), we
expect some, but not all participants to encounter clinically
significant everyday EF deficits. The highest proportions of
reported deficits are expected on the shift and planning
scale (Gioia et al. 2002; Kenworthy et al. 2009). In an
attempt to examine individual differences from a developmental perspective as well, the relative proportions of
clinical scores are examined for different age groups. In
line with Rosenthal et al. (2013), it is expected that the
amount of working memory problems will be significantly
higher in children than in adolescents. Finally, we expect to
123
J Autism Dev Disord (2014) 44:1959–1971
find a positive association between ASD symptom severity
and EF deficits, especially with regard to inhibition and
cognitive flexibility problems (Boyd et al. 2009; Kenworthy et al. 2009).
Methods
Participants
The participants, 155 children and adolescents with ASD
from 6 to 18 years, took part in a larger study (Scheeren
et al. 2010, 2012, 2013) for which they were recruited from
a specialized school for normally intelligent pupils with
ASD. Inclusion criteria for the present study were threefold. First, ASD was diagnosed by a team of clinicians
according to the criteria of the DSM-IV-TR (American
Psychiatric Association APA 1994, 2000). The clinicians
worked independently from the authors and were blind to
the outcomes of the current study. The diagnostic process
included an examination of psychiatric, neuropsychological, and speech functioning. Second, clinical diagnoses
were verified by a raw score above the Dutch threshold for
ASD (which is C60 for boys and C51 for girls) on the
Social Responsiveness Scale (SRS: Constantino and Gruber 2007; Dutch version: Roeyers et al. 2011). The SRS is a
parent questionnaire with 65 items, ranging from 0 (never
true) to 3 (almost always true). It provides an index score
for autism symptomatology; a higher total score indicates
more autistic traits (Constantino and Gruber 2007). Third,
to be included participants had to have a receptive verbal
IQ score C70 on the Dutch version of the Peabody Picture
Vocabulary Test-III (PPVT-III: Dunn and Dunn 2005).
After the exclusion of participants with a score below the
Dutch SRS threshold score (n = 27), or an IQ \ 70
(n = 6), the final sample consisted of 118 participants (16
girls, 102 boys), diagnosed with autism (n = 24), Asperger
(n = 14) and PDD-NOS (n = 80). The final sample was
divided into four different age groups (i.e. 6- to 8 yearolds, 9- to 11 year-olds, 12- to- 14 year-olds and 15- to
18 year olds: see for similar procedure Huizinga and
Smidts 2011). Descriptives of the included participants are
presented in Table 1. Supplementary Table 1 provides
information about the normed performances on the BRIEF
scales, other than those of interest in the current study.
Although the ADOS score was used as a predictor in this
study, it might concern some that over half of the participants did not score above the ASD threshold on the ADOS.
However, all participants are clinically diagnosed within
the spectrum, and the SRS was used to confirm this.
Therefore, it might be the case that the low ADOS scores
indicate a sensitivity problem of the ADOS (Bastiaansen
et al. 2011; Gotham et al. 2008). Nonetheless, we repeated
Range
6.3 (4.1)
63.1 (11.5)
58.5 (7.8)
64.2 (10.6)
57.7 (8.7)
ADOSa
Inhibition T
WM T
Shift T
Plan Tb
6–18
36–82
29–88
39–94
40–74
0–19
61–133
72–132
8.1 (.8)
55.3 (6.8)
59.6 (10)
65 (7)
61.4 (8.4)
7.4 (4.5)
91.2 (15.5)
105.6 (14.1)
47–64
45–81
54–79
43–74
2–19
67–126
84–126
6.4–8.9
Range
55 (8.5)
65.5 (11.5)
63 (9.4)
58.9 (7)
6 (4.4)
91.3 (17.5)
98.3 (11.3)
10.8 (.9)
M (SD)
4/1/9
13/1
60.5 (8.9)
65.2 (9.9)
63.1 (11.5)
58.2 (7.3)
6.3 (4.1)
87.2 (18.1)
109 (11.7)
13.6 (.8)
M (SD)
8/1/20
24/5
40–82
46–83
39–94
41–73
0–19
61–133
80–130
12.1–14.8
Range
9- to 11- (29)
57.6 (8.7)
63.6 (11)
65 (7)
57.1 (8.9)
7.4 (4.5)
91.2 (15.5)
105.9 (13.7)
16.5 (1)
M (SD)
6/7/29
38/4
36–71
42–84
54–79
40–71
2–19
67–126
76–132
15–18.4
Range
12- to 14- (42)
–
–
–
–
.67
2.61
3.40
386.50
F*
–
–
–
–
.57
.06
–
–
–
–
.02
.06
.08
.91
\.001
.02
g2p
p
6/5/22
27/6
15- to 18- (33)
a
df = 3,112/ n = 116;
* df = 3,114
b
n = 117
AUT autism, AS asperger, BRI Behavior Regulation Index, BRIEF Behavioral Rating Inventory Executive Functions, IQ intelligence quotient, M mean, PDD-NOS pervasive developmental
disorder not otherwise specified, Plan plan/organize, PPVT-III Peabody Picture Vocabulary Test-III, SD standard deviation, SRS Social Responsiveness Scale, T T score
38–72
29–88
43–80
44–71
0–18
66–128
72–118
9–11.8
Range
6-to 8- (14)
Autism Diagnostic Observation Scale Generic communication and social reciprocity
87.2 (18.1)
SRS total
13.1 (2. 9)
105.3 (12.9)
PPVT-III
Age
M (SD)
24/14/80
M (SD)
102/16
Diagnosis (AUT/AS/PDD-NOS)
Total (118)
Groups, age in years (n)
Gender (m/v)
Measure
Table 1 Descriptive statistics of the participants
J Autism Dev Disord (2014) 44:1959–1971
1963
123
1964
J Autism Dev Disord (2014) 44:1959–1971
Table 2 Age-related differences on the BRIEF raw subscale scores
Subscale
Groups, age in years (n)
Total
(n = 118)a
M (SD)
6–8
(n = 14)b
M (SD)
9–11
(n = 29)
M (SD)
12–14
(n = 42)
M (SD)
15–18
(n = 33)
M (SD)
F
p
g2p
Posthoc* (p)
Inhibit
20.5 (4.7)
23.8 (2.4)
21.1 (3.9)
20.5 (4.6)
18.6 (5.3)
4.59
.01
.11
6–8 [ 9–11 (.04), 12–14 (.01), 15–18 (.00)
WM
22.1 (3.9)
24 (4.2)
22.3 (3.5)
22.1 (3.3)
21.2 (4.7)
1.67
.18
.04
–
Shift
17.5 (3.4)
17.3 (2.7)
18.2 (2.9)
17.9 (3.5)
16.3 (3.7)
2.11
.10
.05
–
Plan
25.5 (4.5)
24 (3.7)
23.9 (4.6)
26.9 (4.1)
25.7 (4.8)
3.13
.03
.08
9–11 \ 12–14 (.04)
BRIEF behavioral rating inventory executive functions, M mean, Plan plan/organize, SD standard deviation, WM working memory
* Corrected for multiple comparisons
a
n for analyses is 117
b
n for analyses is 13
all analyses with the subsample and describe this sample in
supplementary Table 2. When results with regard tot the
subsample show different patterns compared to findings
concerning the total group this will be mentioned.
With the exception of verbal receptive IQ and age,
group descriptives did not significantly differ between age
groups. In the ADOS subsample receptive IQ didn’t differ
significantly between age groups.
Measurements
Behavioral Rating Inventory Executive Functions
The Behavior Rating Inventory of Executive functions
(BRIEF: Baron 2000; Gioia et al. 2000; Dutch translation
Huizinga and Smidts 2012) is a parent questionnaire that
includes 86 statements regarding behavior. Items can be
rated on a 3-point frequency scale (1 = never; 2 = sometimes; 3 = often). Raw subscale scores can be calculated for
eight subscales (inhibit, working memory, shift, emotional
control, initiate, plan/organize, organization of materials,
and monitoring). Aggregated subscales provide three index
scores: (1) the Behavior Regulation Index: consisting of the
scales inhibit, shift, and emotional control; (2) the Metacognition Index: consisting of the working memory, initiate,
plan/organize, organization of materials, and monitoring
scale; and (3) a total score: a composite of all subscales. For
each composed score a higher score means more everyday
executive problems. T-scores (M = 50, SD = 10) can be
calculated to determine whether scores are potentially clinically significant (T [ 65), according to gender- and agespecific norms (Huizinga and Smidts 2012). Please note that,
in line with the experimental literature concerning EF, we
focused on four relevant BRIEF subscales: inhibit (10 items,
e.g., ‘‘Interrupts others’’); working memory (10 items, e.g.,
‘‘When given three things to do, remembers only the first or
last’’); shift (8 items, e.g., ‘‘Acts upset by a change in
123
Fig. 1 Mean BRIEF planning subscale scores per age group. The
error bars represent 95 % CI
plans’’); and plan/organize (12 items, e.g., ‘‘Does not bring
home homework, assignment sheets, materials, and so on’’).
The BRIEF is a reliable and valid questionnaire (Baron
2000; Gioia et al. 2000, 2002; Huizinga and Smidts 2011).
The Autism Diagnostic Observation Scale Generic
The Autism Diagnostic Observation Scale Generic
(ADOS-G: Lord et al. 2000) is a semi-structured instrument used to observe social interaction, communication,
imagination and repetitive interests in people with supposed ASD. The developmental and language level determine which one of four modules of approximately 45 min
is carried out. Behaviors are scored on a 3-point scale
(0 = normal; 1 = slightly abnormal; 2 = clearly abnormal). Scores are calculated for five domains: communication; social reciprocity; fantasy; repetitive interests; and
other behaviors. A threshold (C7) for the aggregated scores
J Autism Dev Disord (2014) 44:1959–1971
1965
et al. 2010, 2012, 2013), in which various other questionnaires and experimental tasks were used.
Outliers and Missing Datapoints
Fig. 2 Mean BRIEF inhibition subscale scores per age group. The
error bars represent 95 % CI
of communication and social reciprocity determines whether the cut-off for ASD is met. There is a standardized
continuous ADOS score (Gotham et al. 2009), which
would be most suitable to measure severity. However,
module 4 is not standardized. Therefore, we used the
ADOS aggregated scores of communication and social
reciprocity as a measure of severity. A higher score reflects
more autism symptoms. Note, that in the current study
module 3 (N = 39) or 4 (N = 77), for children or adolescents with fluent speech, were used. The validity and
reliability of the ADOS are reported as good to excellent
(Lord et al. 2000).
Procedure
For each participant there was parental informed consent.
Participants from 12 year and older also gave informed
consent for themselves. Both the ADOS and BRIEF were
administered as part of a larger research project (Scheeren
There were no outliers ([3 SE) in the current sample on
any of the included BRIEF scales. Following the BRIEF
manual (Huizinga and Smidts 2009), if no more than 2
subscale items were missing, missing items were scored as
1 and subscales were computed. SRS missing items were
scored as the mean score of the completed subscale items,
only when no more than 2 items per scale were missing.
The SRS total score was calculated as the sum of the
subscales. More than 2 items were missing for one participant on the BRIEF plan/organize scale. Consequently,
this score is missing for this participant. The ADOS was
only assessed with 116 participants. Therefore, for two
participants the ADOS score is missing.
Results
Are there Differences Between the Age Groups
on the BRIEF Scores?
A multivariate analysis of variance (MANOVA) was
conducted with the four BRIEF scale raw scores as
dependent variables, and age group as between subject
factor. Alpha level for the MANOVA and the subsequent
ANOVAs were set on .05. Given the unequal sample sizes
and, for inhibition and working memory, the failure to meet
the heteroscedasticity criterion (Field 2009), the Games
Howell procedure was used in posthoc procedures. Effect
sizes were expressed with partial g2, representing small,
medium, and large effects by values of respectively C.01,
.06, and .14 (Cohen 1992). In Table 2 results of the MANOVA, and subsequent ANOVAs, were reported.
A significant main effect of age was found, Wilks
K = .654 F(12,291.32) = 4.23, p = \ .001, partial
Table 3 Percentages of clinical scores on the BRIEF subscales
Subscale
Groups, age in years (n)
Total
(n = 118)
6–8
(n = 14)
9–11
(n = 29)
12–14
(n = 42)
15–18
(n = 33)
Fisher’s exact test
p
Cramer’s V
Inhibit
50.0
48.3
38.1
36.4
41.5
1.58
.68
.11
WM
50.0
13.8
19.0
24.2
22.9
6.77
.07
.25
Shift
28.6
51.7
57.1
51.5
50.8
3.41
.34
.17
Plana
0.0
10.3
33.3
18.2
19.7
9.21
.02
.29
ADOS autism diagnostic observation scale generic communication and social reciprocity, BRIEF behavioral rating inventory executive functions,
M mean, Plan plan/organize, SD standard deviation, WM working memory
a
n = 117
123
1966
J Autism Dev Disord (2014) 44:1959–1971
Table 4 Regression analyses with BRIEF subscales as dependent variables
BRIEF scales
Predictors
Inhibit
DR2
Step 1
Total R2
p
-.35
\.001
.05
.61
.999
.001
.00
ADOS
.00
.12
DR2
b
.12
Age
PPVT-III
Step 2
WM
b
p
DR2
.07
.02
-.22
.02
.08
.41
.25
.05
.01
.999
.002
Shift
-.11
.06
b
P
DR2
.28
.05
-.17
.08
.01
.95
.11
.02
.25
.09
Plan
-.15
.09
p
.16
.09
.14
.14
.59
.06
.00
.11
.16
b
.05
.16
.59
.12
Beta’s are standardized beta’s from the full model
ADOS autism diagnostic observation scale generic communication and social reciprocity, BRIEF behavioral rating inventory executive functions,
PPVT-III Peabody Picture Vocabulary Test-III
g2 = .13, indicating differences between the age groups on
both the inhibit scale, and the plan/organize scale, but not
on the working memory and the shift scale. The follow up
analyses revealed that compared to 6- to 8- year-olds,
inhibition problems were less severe in the older groups
(see Fig. 1).
However, for planning there was another pattern of
results (see Fig. 2) as compared to 9- to 11-year-olds, 12to 14- year-olds showed more planning problems, p = .04.
None of the other post hoc comparisons for planning were
significant.
As some aspects of IQ cover EF, controlling for IQ
might provide overcorrected or unlikely results (Dennis
et al. 2009). However, we repeated the analyses with
receptive verbal IQ as a covariate: the same pattern was
found, except that age-related differences in planning were
only moderately significant, F(3,112) = 2.49, p = .06.
In the ADOS cut-off subsample there was a main age
effect as well, Wilks K = .487 F(12,108.77) = 2.83,
p = .002, partial g2 = .21. However, in this sample this
was caused by age differences in inhibition,
F(4,43) = 2.92, p = .04, partial g2 = .17 (less problems
in 15-18- year-olds, compared to 9- to 11- year olds,
p = .04) and shift, F(4,43) = 5.37, p = .003, partial
g2 = .27 (less problems in 15-18- year-olds, compared to
9- to 11-, p = .04; and 12- to 14- year olds, p = (\ .01).
When repeating the analyses of the subsample with
receptive verbal IQ as covariate the same pattern was
found.
What is the Proportion of Children and Adolescents
with Clinical BRIEF Scores, and are Proportions
Different Between Age Groups?
The BRIEF scores were first categorized as clinical
(T [ 65) or within normal limits (T B 65). Based on this,
the proportions of actual clinical scores per age group were
123
calculated. Next, Fisher’s exact tests with adjusted alpha
levels of .01 (.05/4) were used to test differences in relative
proportions across age groups. Effect sizes were expressed
with Cramer’s V. Effects from ±.1 represent a small effect,
this is medium for ±.3, and large for ±.5 (Field 2009).
Clinical scores were observed for inhibition in 42 % of the
participants, for working memory in 23 %, for shift in
51 % and for plan/organize in 20 % of participants. For the
plan/organize scale clinical scores were absent in the
youngest group, but present in 33 % of the children from to
9- to 11-years-old. Nonetheless, no significant age group
differences were found with regard to clinical percentages
of all BRIEF scale scores (see Table 3). In the ADOS cutoff subsample this was 43 % for inhibition, 20 % for
working memory, 47 %, for shift, and 21 % for plan/
organize and, again, there were no age differences.
What is the Unique Contribution of ASD Severity
to BRIEF Scores?
Four multiple, hierarchical regression analyses were performed with BRIEF subscale scores as dependent variables. As receptive verbal IQ differed between age groups,
we included IQ together with age in the first step, and
ADOS aggregated scores in the second step. Alpha level
was set at .01 (.05/4). Prior to the regression analyses,
correlation analyses were performed. Alpha level was set at
.01 (05/4). Correlation analyses revealed no relations
between the BRIEF subscales and the ADOS: Inhibit,
r (114) = .04, p = .70; working memory, r (114) = .1,
p = .33; shift, r (114) = .13, p = .16; plan/organize,
r (114) = .02, p = .87. In the ADOS cut-off subsample
relations were absent as well: Inhibit, r (49) = .04,
p = .81; working memory, r (49) = .03, p = .83; shift,
r (49) = .02, p = .92; plan/organize, r (48) = .032,
p = .84. Regression analysis showed that 12 % of the
variance in the inhibit scale was explained. This was 16 %
J Autism Dev Disord (2014) 44:1959–1971
in the ADOS cut-off subsample. However, in both cases
this was fully attributable to the effect of age. Autism
severity did not add uniquely to the variance of the inhibition scale, nor did IQ (see Table 4). With respect to the
other BRIEF subscales (working memory, shift and plan/
organize), none of the regression models were significant.
See the results and statistics in Table 4.
Discussion
In the current cross-sectional study we focused on agerelated differences in specific domains of everyday EF in
children and adolescents with ASD as reported by their
parents. Moreover, we explored age differences in proportions of clinically significant EF problems, and the
relationship between ASD symptom severity and everyday
EF. Age-related differences were found with regard to
inhibition and planning. Compared to the 6- to 8-year-olds,
inhibition problems were reported less for the older children and adolescents. For planning, an opposite pattern was
found. Compared to the 9- to 11-year-olds, in 12- to 14year-olds more planning problems were observed. Hence,
planning problems may be especially apparent in young
adolescents with ASD during the transition period from
primary to secondary education. In a subsample of children
and adolescent that scored above the ASD threshold on the
ADOS, differences were found between age groups for
inhibition and cognitive flexibility. Cognitive flexibility
problems were less apparent in the oldest group (15- to
18-year olds), compared to 9- to 11-year-olds and 12- to
14- year-olds. Consistent with former studies, everyday EF
deficits were found in children and adolescents with ASD,
although to a smaller extent than expected. Only 20 %
(planning) to 51 % (cognitive flexibility) of the participants
encountered clinical EF problems. In the ADOS cut-off
subsample this was approximately the same. This highlights the heterogeneity of ASD, and underlines the
importance of focusing on individual differences when
studying EF in ASD. Counter to our expectation, no relations between ASD symptom severity and everyday EF
were found.
Developmental Profiles of Everyday EF
As expected, in general, older children and adolescents
with ASD showed fewer inhibition problems. However,
this was only the case when comparing the older age
groups with the youngest group. This might suggest a
delayed or protracted development of inhibition, rather
than a deviated development, which is in line with Rosenthal et al. (2013). Longitudinal research with both ASD
and TD participants is necessary to test this hypothesis.
1967
Age-related differences were not found in working
memory and cognitive flexibility. With respect to working
memory, this contrasts the observed decrease of everyday
working memory problems during childhood in TD (Huizinga and Smidts 2011), which consequently fits the
increasing T-scores found by Rosenthal et al. (2013).
Hence, it might be the case that working memory problems
in ASD are specifically revealed during adolescence.
Increased real world expectations (Rosenthal et al. 2013)
do not seem to explain this though, because age-related
differences of raw scores could not be found. Considering
that laboratory task performance revealed a deviant
development of working memory during adolescence in
ASD (Luna et al. 2007) an explanation at the cognitive
level might be more valid. Hence, it seems that working
memory development in children and adolescents with
ASD differs from that of typical development, which, with
age, might lead to an increase of impairments compared to
TD. Nonetheless, differences between real life observations
and observed laboratory performance make it hard to draw
firm conclusions. Further research could focus on the
unique contribution of specific cognitive developmental
processes on the one hand, and increasing environmental
demands on the other hand. With respect to cognitive
flexibility, the lack of age-related differences in the current
study is against our expectations and inconsistent with
previously observed age-related improvements in cognitive
flexibility performance in ASD (Happé et al. 2006).
However, in a subsample of children and adolescents that
scored above the ASD threshold on the ADOS we did note
age related differences, as the 15- to 18- year olds had
fewer problems than the 9- to 11- and 12- to 14-year-olds.
One explanation might be that children and adolescents
that score above the ASD threshold on the ADOS (and
have more severe ASD than those who score below the
threshold) receive more professional help or help from
their families with switching. However, this is only speculative. In line with Rosenthal et al. (2013), in both of the
current samples an increase of flexibility problems was
absent. This might suggest that, if cognitive flexibility
problems are characteristic for ASD, there is at least no
increase of problems during childhood and adolescence.
Longitudinal research with ASD and TD participants could
test this hypothesis.
Contrary to findings based on laboratory tasks (Happé
et al. 2006; Pellicano 2010), in the total group of children
and adolescents with ASD, more planning problems were
observed in 12- to 14-year-olds compared to 9- to 11-yearolds. This might well be explained by two interdependent
factors. First, this could stem from changing demands of
the environment. Naturally, children from 12 years and
older are expected to behave more independently with
regard to school tasks, homework, and everyday activities
123
1968
than younger children. In the Netherlands it is uncommon
for children younger than twelve to receive homework.
Although an increase of planning problems in ASD was not
found in their study, this is in line with the arguments of
Rosenthal et al. (2013) about increased expectations from
the environment. Second, parents of younger children
might consequently rate behaviors as ‘‘never a problem’’,
when these behaviors are not yet asked for in real life.
Thus, this does not mean that planning problems are
absent, but that the BRIEF might be insufficient to detect
these problems. Indeed, the observed age-related differences disappeared in an exploratory analysis with a smaller
planning scale where we excluded the items referring to
homework or tasks (5 items).1 To determine whether or not
daily life planning abilities change across the life span in
children and adolescents with ASD, it is of importance that
a planning scale is developed that is more valid with
respect to a broader age range. In the ADOS cut-off subsample an increase of planning problems was absent.
Again, although speculative, this might be explained by the
fact that the family of children and adolescents with more
severe ASD or professionals provide more help compared
to those supporting participants with milder symptoms,
thereby preventing an increase of planning problems when
homework is involved.
Summing up, inhibition problems were most apparent in
the 6- to 8- year-olds, whereas in the total group studied
more planning problems were observed in 12- to 14- yearolds compared to 9- to 11- year-olds. With regard to
working memory and cognitive flexibility no age-related
differences were found, except that in the ADOS cut-off
subsample cognitive flexibility problems were less apparent in the oldest group (15- to 18- year olds), compared to
9- to 11-, and 12- to 14- year-olds. This clearly underlines
the importance of studying EF domains separately.
The Proportion of Children and Adolescents
with Clinically Significant EF Problems
Compared to previous studies (Gioia et al. 2002; Kenworthy et al. 2005), we found relatively small proportions
of children and adolescents with ASD with clinically significant EF problems (inhibition: 42 %; working memory:
23 %; cognitive flexibility: 51 %; planning: 20 %). This
could be influenced by the fact that the samples from the
previous studies came from a hospital based neuropsychology service, to which children might have been
referred because of behavioral problems. The current
sample might be more representative for the prevalence of
1
For this restricted planning scale we eliminated all questions with
the words ‘‘homework’’, ‘‘task’’, or ‘‘assignment’’ in the in Dutch
translated descriptions, F (3,112) = 1.19, p = .32.
123
J Autism Dev Disord (2014) 44:1959–1971
EF deficits in a more general, ASD population with normal
to high intelligence levels.
Consistent with other studies, we found that flexibility
belongs to the most affected domains of EF (Gioia et al.
2002; Kenworthy et al. 2005). In fact, everyday problems
in inhibition, working memory, and planning were only
visible in a minority of the children and adolescents with
ASD. No age-related differences in proportions of clinically significant scores were observed. This clearly shows
that, although group differences exist between groups with
ASD and TD (Boyd et al. 2009; Chan et al. 2009; Endedijk
et al. 2011; Kalbfleisch and Loughan 2012; Winsler et al.
2007; Yerys et al. 2009; Zandt et al. 2007), not all individuals with ASD encounter EF problems in one or more
EF domains.
The Contribution of ASD Severity
Autism severity, as observed by an objective informant, did
not contribute to any of the four EF domains (inhibition,
working memory, cognitive flexibility, and planning). To
analyse whether the lack of association was due to the
relatively mild severity of autistic problems in the sample,
we have repeated the analyses with a subsample of participants that scored above the ASD threshold on the
ADOS. In this subsample no relation with severity was
found either. This is surprising, since the BRIEF is widely
used in the clinical practice with regard to ASD, and the
ADOS score is based on a broad spectrum of ASD symptoms. Exploratory analyses2 revealed moderate to strong
relations between the SRS and the aforementioned BRIEF
scores, while no relation between the ADOS score and the
SRS index score was found. Relations between the parent
report (SRS) and the BRIEF can be explained by informant- and content-overlap. A lack of relations between the
BRIEF and the ADOS might be explained by construct
validity problems of the BRIEF. However, it could also be
the case that the BRIEF is a more general measurement of
behavioral disruption (McAuley et al. 2010) instead of EF.
Items like ‘‘interrupts others’’, ‘‘gets out of seat at the
wrong times, and ‘‘gets out of control more than friends’’
might refer to more generally disruptive and/or inattentive
behaviors than to EF. Rather than ASD symptoms, the
BRIEF could thus pick up on other symptoms, like
comorbid ADHD. This is a question that should be
addressed in future research.
Next to the already mentioned shortcomings, one
shortcoming of this study is that the youngest age group
was too small to detect differences with a medium or small
2
Relations ADOS severity score: SRS index score, r = .07;
inhibition, r = .49; working memory, r = .39; shift, r = .86; and
planning, r = .56. Data can be obtained from the first author.
J Autism Dev Disord (2014) 44:1959–1971
effect. It could, therefore, be argued that we had insufficient power to determine age effects. However, we solved
this by using more powerful regression analyses with age
as a predictor to confirm findings.
To conclude, the BRIEF was developed to measure
everyday behaviors that are often found disturbed in children and adolescents with ASD and that seem related to
EF. Based on the current study, we wonder whether the
BRIEF is an appropriate instrument to measure everyday
EF from a developmental perspective, however. The perceptions of parents on behaviors take place in a context of
environmental demands that vary with age. It is, therefore,
complex to determine whether observed differences
between age groups point to true developmental changes of
an underlying deficit or simply reflect responses to
changing environmental demands. The planning scale in
particular seems more appropriate to measure problems in
adolescents than in children. Therefore, adjustments might
be needed in order to increase ecological validity for a
broad age range. In a clinical assessment it would be best to
study EF with laboratory tasks as well as the BRIEF
questionnaire. When both EF skills and environmental
demands are taken into account this might help determining whether interventions are needed to aim at actual EF,
the social environment, or both. Furthermore, profiles of
deficits might predict developmental (Berger et al. 2003) or
intervention outcomes (van der Oord, et al. 2008), which
could benefit the efficiency and effectiveness in the clinical
practice. Nonetheless, likewise to EF studies that take place
in the laboratory, the current study highlights that some,
but not all domains of everyday EF are impaired in some,
but not all children and adolescents with ASD. Since the
majority of the children and adolescents does not seem to
have clinically significant problems, even within a subsample of children and adolescent who scored above the
ASD threshold on the ADOS, and because autism severity
is not related to everyday EF problems it is important to
focus on individual differences in ASD.
Acknowledgments This work was partly funded as part of the
research program ‘‘Autism and Aging: A Double Jeopardy’’, which is
financed (personal VIDI grant HM Geurts, No. 452-10-003) by NWO,
and was partly financially supported by Stichting Nuts Ohra [SNO-T0701-116]. The authors would like to thank all children, adolescents,
parents and teachers of the Berg en Boschschool who took part in this
study.
References
Adams, N., & Jarrold, C. (2011). Inhibition in autism: Children with
autism have difficulty inhibiting irrelevant distractors but not
prepotent responses. Journal of Autism and Developmental
Disorders, 42(6), 1052–1063.
1969
American Psychiatric Assosiation. (1994). Diagnostical and statistical manual of mental disorders (4th ed.). Washington, DC:
American Psychiatric Association.
American Psychiatric Assosiation (2000). Diagnostical and statistical
manual of mental disorders (4th ed., text revision). Washington,
DC: American Psychiatric Association.
Anderson, V. A., Anderson, P., Northam, E., Jacobs, R., & Catroppa,
C. (2001). Development of executive functions through late
childhood and adolescence in an Australian sample. Developmental Neuropsychology, 20(1), 385–406.
Baddeley, A. (1992). Working memory. Science, 255(5044),
556–559.
Baron, I. S. (2000). Behavior rating inventory of executive function.
Child Neuropsychology, 6(3), 235–238.
Bastiaansen, J. A., Meffert, H., Hein, S., Huizinga, P., Ketelaars, C.,
Pijnenborg, M., et al. (2011). Diagnosing autism spectrum
disorders in adults: The use of Autism diagnostic observation
schedule (ADOS) module 4. Journal of Autism and Developmental Disorders, 41, 1256–1266.
Berger, H. J. C., Aerts, F. H. T. M., Spaendonck, K. P. M., Cools, A.
R., & Teunisse, J. P. (2003). Central coherence and cognitive
shifting in relation to social improvement in high-functioning
young adults with autism. Journal of Clinical and Experimental
Neuropsychology, 25(4), 502–511.
Best, J. R., Miller, P. H., & Jones, L. L. (2009). Executive functions
after age 5: Changes and correlates. Developmental Review,
29(3), 180–200.
Boyd, B. A., McBee, M., Holtzclaw, T., Baranek, G. T., & Bodfish, J.
W. (2009). Relationships among repetitive behaviors, sensory
features, and executive functions in high functioning autism.
Research in Autism Spectrum Disorders, 3(4), 959–966.
Bramham, J., Ambery, F., Young, S., Morris, R., Russell, A.,
Xenitidis, K., et al. (2009). Executive functioning differences
between adults with attention deficit hyperactivity disorder and
autistic spectrum disorder in initiation, planning and strategy
formation. Autism, 13(3), 245–264.
Chan, A. S., Cheung, M., Han, Y. M. Y., Sze, S. L., Leung, W. W.,
Man, H. S., et al. (2009). Executive function deficits and neural
discordance in children with autism spectrum disorders. Clinical
Neurophysiology, 120(6), 1107–1115.
Christ, S. E., Holt, D. D., White, D. A., & Green, L. (2007). Inhibitory
control in children with autism spectrum disorder. Journal of
Autism and Developmental Disorders, 37(6), 1155–1165.
Christ, S. E., Kester, L. E., Bodner, K. E., & Miles, J. H. (2011).
Evidence for selective inhibitory impairment in individuals with
autism spectrum disorder. Neuropsychology, 25(6), 690–701.
Constantino, J. N., & Gruber, C. P. (2007). Social responsiveness
scale. Los Angeles: Western Psychological Services.
Corbett, B. A., Constantine, L. J., Hendren, R., Rocke, D., & Ozonoff,
S. (2009). Examining executive functioning in children with
autism spectrum disorder, attention deficit hyperactivity disorder
and typical development. Psychiatry Research, 166(2–3),
210–222.
Damasio, A. R., & Maurer, R. G. (1978). A neurological model for
childhood autism. Archives of Neurology, 35(12), 777–786.
Davidson, M. C., Amso, D., Anderson, L. C., & Diamond, A. (2006).
Development of cognitive control and executive functions from 4
to 13 years: Evidence from manipulations of memory, inhibition,
and task switching. Neuropsychologia, 44(11), 2037–2078.
De Luca, C. R., Wood, S. J., Anderson, V., Buchanan, J., Proffitt, T.
M., Mahony, K., et al. (2003). Normative data from the cantab. I:
Development of executive function over the lifespan. Journal of
Clinical and Experimental Neuropsychology, 25(2), 242–254.
de Vries, M., & Geurts, H. M. (2012). Cognitive flexibility in ASD;
task switching with emotional faces. Journal of Autism and
Developmental Disorders, 42(12), 2558–2568.
123
1970
Dennis, M., Francis, D. J., Cirino, P. T., Schachar, R., Barnes, M. A.,
& Fletcher, J. M. (2009). Why IQ is not a covariate in cognitive
studies of neurodevelopmental disorders. Journal of the International Neuropsychological Society, 15(3), 331–343.
Dunn, L. M., & Dunn, L. M. (2005). Peabody Picture Vocabulary
Test (PPVT)-III-NL. Amsterdam: Hartcourt Test Publishers.
Edgin, J. O., & Pennington, P. F. (2005). Spatial cognition in autism
spectrum disorders: Superior, impaired, or just intact? Journal of
Autism and Developmental Disorders, 35(6), 729–745.
Endedijk, H., Denessen, E., & Hendriks, A. W. (2011). Relationships
between executive functioning and homework difficulties in
students with and without autism spectrum disorder: An analysis
of student- and parent-reports. Educational Neuroscience,
Learning and Individual Differences, 21(6), 765–770.
Field, A. (2009). Discovering statistics using SPSS (3rd ed.). London:
Sage.
Friedman, N. P., & Miyake, A. (2001). The relations among inhibition
and interference control functions: A latent-variable analysis.
Journal of Experimental Psychology: General, 133(1), 101–135.
Gathercole, S. E., Pickering, S. J., Ambridge, B., & Wearing, H.
(2004). The structure of working memory from 4 to 15 years of
age. Developmental Psychology, 40(2), 177–190.
Geurts, H. M., Begeer, S., & Stockmann, L. (2009a). Brief report:
Inhibitory control of socially relevant stimuli in children with
high functioning autism. Journal of Autism and Developmental
Disorders, 39(11), 1603–1607.
Geurts, H. M., Corbett, B., & Solomon, M. (2009b). The paradox of
cognitive flexibility in autism. Trends in Cognitive Sciences,
13(2), 74–82.
Gioia, G. A., Isquith, P. K., Guy, S. C., & Kenworthy, L. (2000).
Behavior rating inventory of executive function. Odesse, FL:
Psychological Assessment Resources.
Gioia, G. A., Isquith, P. K., Kenworthy, L., & Barton, R. M. (2002).
Profiles of everyday executive function in acquired and development disorders. Child Neuropsychology, 8(2), 121–137.
Gotham, K., Pickles, A., & Lord, C. (2009). Standardizing ADOS
scores for a measure of severity in autism spectrum disorders.
Journal of Autism and Developmental Disorders, 39(5),
693–705.
Gotham, K., Risi, S., Dawson, G., Tager-Flusberg, H., Joseph, R.,
Carter, A., et al. (2008). A replication of the Autism Diagnostic
Observation Schedule (ADOS) revised algorithms. Journal of
the American Academy of Child and Adolescent Psychiatry, 47,
642–651.
Happé, F., Booth, R., Charlton, R., & Hughes, C. (2006). Executive
function deficits in autism spectrum disorders and attentiondeficit/hyperactivity disorder: Examining profiles across
domains and ages. Brain and Cognition, 61(1), 25–39.
Hill, E. L. (2004). Executive dysfunction in autism. Trends in
Cognitive Sciences, 8(1), 26–32.
Hill, E. L., & Bird, C. M. (2006). Executive processes in asperger
syndrome: Patterns of performance in a multiple case series.
Neuropsychologia, 44(14), 2822–2835.
Hughes, C. (2011). Changes and challenges in 20 years of research
into the development of executive functions. Infant and Child
Development, 20(3), 251–271.
Hughes, C., Ensor, R., Wilson, A., & Graham, A. (2009). Tracking
executive function across the transition to school: A latent variable
approach. Developmental Neuropsychology, 35(1), 20–36.
Huizinga, M., Dolan, C. V., & van der Molen, M. W. (2006). Agerelated change in executive function: Developmental trends and
a latent variable analysis. Neuropsychologia, 44(11), 2017–2036.
Huizinga, M., & Smidts, D. P. (2011). Age-related changes in
executive function: A normative study with the Dutch version of
the behavior rating inventory of executive function (BRIEF).
Child Neuropsychology, 17(1), 51–66.
123
J Autism Dev Disord (2014) 44:1959–1971
Huizinga, M., & Smidts, D. P. (2012). BRIEF Vragenlijst executieve functies voor 5- tot 18-jarigen. Amsterdam: Hogrefe
Uitgevers.
Kalbfleisch, M. L., & Loughan, A. R. (2012). Impact of IQ
discrepancy on executive function in high-functioning autism:
Insight into twice exceptionality. Journal of Autism and
Developmental Disorders, 42(3), 390–400.
Kenworthy, L., Black, D. O., Harrison, B., della Rosa, A., & Wallace,
G. L. (2009). Are executive control functions related to autism
symptoms in high-functioning children? Child Neuropsychology,
15(5), 425–440.
Kenworthy, L. E., Black, D. O., Wallace, G. L., Ahluvalia, T.,
Wagner, A. E., & Sirian, L. M. (2005). Disorganization: The
forgotten executive dysfunction in high-functioning autism
(HFA) spectrum disorders. Developmental Neuropsychology,
28(3), 809–827.
Kenworthy, L., Yerys, B. E., Anthony, L. G., & Wallace, G. L.
(2008). Understanding executive control in autism spectrum
disorders in the lab and in the real world. Neuropsychology
Review, 18(4), 320–338.
Lezak, M. D., Howieson, D. B., Bigler, E. D., & Tranel, D. (2012).
Neuropsychological assessment (5th ed.). New York, NY:
Oxford University Press.
Lord, C., Risi, S., Lambrecht, L., Cook, E. H, Jr, Leventhal, B. L.,
DiLavore, P. C., et al. (2000). The autism diagnostic observation
schedule-generic: A standard measure of social and communication deficits associated with the spectrum of autism. Journal of
Autism and Developmental Disorders, 30(3), 205–223.
Lord, C., Rutter, M., & Le Couteur, A. (1994). Autism diagnostic
interview, revised: A revised version of a diagnostic interview
for caregivers of individuals with possible pervasive developmental disorders. Journal of Autism and Developmental Disorders, 24(5), 659–685.
Luciana, M., Collins, P. F., Olson, E. A., & Schissel, A. M. (2009).
Tower of London performance in healthy adolescents: The
development of planning skills and associations with selfreported inattention and impulsivity. Developmental Neuropsychology, 34(4), 461–475.
Luna, B., Doll, S. K., Hegedus, S. J., Minshew, N. J., & Sweeney, J.
A. (2007). Maturation of executive function in autism. Biological Psychiatry, 61(4), 474–481.
Luna, B., Garver, K. E., Urban, T. A., Lazar, N. A., & Sweeney, J. A.
(2004). Maturation of cognitive processes from late childhood to
adulthood. Child Development, 75(5), 1357–1372.
McAuley, T., Chen, S., Goos, L., Schachar, R., & Crosbie, J. (2010).
Is the behavior rating inventory of executive function more
strongly associated with measures of impairment or executive
function? Journal of the International Neuropsychological
Society, 16(3), 495–505.
Minshew, N. J., & Goldstein, G. (2001). The pattern of intact and
impaired memory functions in autism. Journal of Child
Psychology and Psychiatry, 42(8), 1095–1101.
Miyake, A., Friedman, N. P., Emerson, M. J., Witzki, A. H., &
Howerter, A. (2000). The unity and diversity of executive
functions and their contributions to complex ‘‘frontal lobe’’
tasks: A latent variable analysis. Cognitive Psychology, 41(1),
49–100.
Monsell, S. (2003). Task switching. Trends in Cognitive Sciences,
7(3), 134–140.
Mosconi, M. W., Kay, M., D’Cruz, A. M., Seidenfeld, A., Guter, S.,
Stanford, L. D., et al. (2009). Impaired inhibitory control is
associated with higher-order repetitive behaviors in autism
spectrum disorders. Psychological Medicine, 39(9), 559–1566.
Ozonoff, S., & McEvoy, R. E. (1994). A longitudinal study of
executive function and theory of mind development in autism.
Development and Psychopathology, 6(3), 415–431.
J Autism Dev Disord (2014) 44:1959–1971
Pellicano, E. (2010). The development of core cognitive skills in
autism: A 3-year prospective study. Child Development, 81(5),
1400–1416.
Pellicano, E., Maybery, M., Durkin, K., & Maley, A. (2006). Multiple
cognitive capabilities/deficits in children with an autism spectrum disorder: ‘‘Weak’’ central coherence and its relationship to
theory of mind and executive control. Development and
Psychopathology, 18(1), 77–98.
Pennington, B. F., & Ozonoff, S. (1996). Executive functions and
developmental psychopathology. Journal of Child Psychology
and Psychiatry, 37(1), 51–87.
Roeyers, H., Thys, M., Druart, C., De Schryver, M., & Schittekate, M.
(2011). SRS screeningslijst voor autismespectrumstoornissen,
handleiding. Amsterdam: Hogrefe Uitgevers.
Rosenthal, M., Wallace, G. L., Lawson, R., Wills, M. C., Dixon, E.,
Yerys, B. E., et al. (2013). Impairments in real-world executive
function increase from childhood to adolescence in autism
spectrum disorders. Neuropsychology, 27(1), 13–18.
Russo, N., Flanagan, T., Iarocci, G., Berringer, D., Zelazo, P. D., &
Burack, J. A. (2007). Deconstructing executive deficits among
persons with autism: Implications for cognitive neuroscience.
Brain and Cognition, 65(1), 77–86.
Scheeren, A. M., Begeer, S., Banerjee, R., Meerum Terwogt, M., &
Koot, H. M. (2010). Can you tell me something about yourself?
Self-presentation in children and adolescents with high functioning autism spectrum disorder in hypothetical and real life
situations. Autism, 14(5), 457–473.
Scheeren, A. M., de Rosnay, M., Koot, H. M., & Begeer, S. (2013).
Rethinking theory of mind in high-functioning autism spectrum
disorder. Journal of Child Psychology and Psychiatry, 54(6),
628–635.
Scheeren, A. M., Koot, H. M., & Begeer, S. (2012). Social interaction
style of children and adolescents with high-functioning autism
spectrum disorder. Journal of Autism and Developmental
Disorders, 42(10), 2046–2055.
Sinzig, J., Morsch, D., Bruning, N., Schmidt, M. H., & Lehmkuhl, G.
(2008). Inhibition, flexibility, working memory and planning in
1971
autism spectrum disorders with and without comorbid ADHDsymptoms. Child and Adolescent Psychiatry and Mental Health,
13(2), 117–126.
Smith, E. E., Jonides, J., & Koeppe, R. A. (1996). Dissociating verbal
and spatial working memory using PET. Cerebral Cortex, 6(1),
11–20.
Steele, S. D., Minshew, N. J., Luna, B., & Sweeney, J. A. (2007).
Spatial working memory deficits in autism. Journal of Autism
and Developmental Disorders, 37(4), 605–612.
Towgood, K. J., Meuwese, J. D. I., Gilbert, S. J., Turner, M. S., &
Burgess, P. W. (2009). Advantages of the multiple case series
approach to the study of cognitive deficits in autism spectrum
disorder. Neuropsychologia, 47(13), 2981–2988.
Van der Oord, S., Prins, P. J. M., Oosterlaan, J., & Emmelkamp, P.
M. G. (2008). Treatment of attention deficit hyperactivity
disorder in children: predictors of treatment outcome. European
Child and Adolescent Psychiatry, 17(2), 73–84.
Verté, S., Geurts, H., Roeyers, H., Oosterlaan, J., & Sergeant, J. A.
(2005). Executive functioning in children with autism and
Tourette syndrome. Development and Psychopatholy, 17(2),
415–445.
Williams, D. L., Goldstein, G., Carpenter, P. A., & Minshew, N. J.
(2005). Verbal and spatial working memory in autism. Journal of
Autism and Developmental Disorders, 35(6), 747–756.
Winsler, A., Abar, B., Feder, M. A., Schunn, C. D., & Rubio, D. A.
(2007). Private speech and executive functioning among highfunctioning children with autistic spectrum disorders. Journal of
Autism and Developmental Disorders, 37(9), 1617–1635.
Yerys, B. E., Wallace, G. L., Sokoloff, J. L., Shook, D. A., James, J.
D., & Kenworthy, L. (2009). Attention deficit/hyperactivity
disorder symptoms moderate cognition and behavior in children
with autism spectrum disorders. Autism Research, 2(6),
322–333.
Zandt, F., Prior, M., & Kyrios, M. (2007). Repetitive behavior in
children with high functioning autism and obsessive compulsive
disorder. Journal of Autism and Developmental Disorders, 37(2),
251–259.
123
Copyright of Journal of Autism & Developmental Disorders is the property of Springer
Science & Business Media B.V. and its content may not be copied or emailed to multiple sites
or posted to a listserv without the copyright holder's express written permission. However,
users may print, download, or email articles for individual use.
Journal of Child Psychology and Psychiatry 46:4 (2005), pp 401–408
doi: 10.1111/j.1469-7610.2004.00361.x
Continuity and change from early childhood
to adolescence in autism
Corina W. McGovern and Marian Sigman
University of California at Los Angeles, USA
Background: This longitudinal study of 48 children diagnosed with autism at 2–5 years of age was
designed to test the hypothesis that diagnosis would remain stable for most of the sample but that
there would be improvements in symptom severity, adaptive behavior, and emotional responsiveness
in adolescence. Methods: A sample of children with autism assessed in both early and middle
childhood were observed in late adolescence with the Autism Diagnostic Observation Scale (ADOS)
and their parents were administered the Autism Diagnostic Interview-Revised (ADI-R) and the
Vineland Adaptive Behavior Scale. Results: All but 2 adolescents (46 of 48) met lifetime criteria for
autism according to the ADI-R, and all but 4 adolescents (40 of 44) met criteria for autism spectrum
disorder on the ADOS. In contrast to the continuity in diagnosis, parents described improvements in
social interactions, repetitive/stereotyped behaviors, adaptive behaviors, and emotional responsiveness to others’ distress in adolescence compared to middle childhood. High-functioning adolescents
with autism showed more improvement in these domains than low-functioning adolescents with
autism. The extent to which the adolescents were observed to be socially engaged with their peers in
school in middle childhood predicted adaptive behavior skills even when intelligence level was
statistically constrained. Conclusions: The developmental trajectory of children with autism appears
to show both continuity and change. In this sample, most individuals continued to be diagnosed in
the autism spectrum but parents reported improvements in adolescence. The results suggest that
social involvement with peers improves adaptive behavior skills, and this argues for focusing
intervention programs in this area. In addition, it is clear that high-functioning adolescents improve
more than low-functioning individuals not only in cognitive abilities but also in social interaction
skills. Thus, any early intervention that impacts the cognitive abilities of young children with autism
is likely to have a parallel influence on their social skills as they mature into late adolescence and
early adulthood. Keywords: Autism, developmental continuity, diagnosis, adaptive behavior,
emotional responsiveness.
The developmental course from early childhood into
adolescence and young adulthood of children with
autism has rarely been described. The lack of a clear
understanding of the developmental trajectory in
autism has several explanations. First, the very few
long-term longitudinal studies of children with autism have generally included only one or two measures often administered at only a few time points. In
addition, the measures used have tended to be based
on parental recollection of general categories of behavior over fairly long time periods. Although parental reports can provide reliable information, this is
more likely to be true for reports of current characteristics and behaviors. Finally, it is difficult to generalize findings across studies because of variations
in the methods used and sample characteristics
since agreement on diagnostic criteria and reliable
diagnostic instruments are relatively new.
The current study was formulated to contribute to
the understanding of developmental trajectories in
autism by following a moderate size sample of children with autism with assessments in early childhood, the mid-school years, and late adolescence/
young adulthood. Although the children were diagnosed with a variety of measures at intake, diagnoses were confirmed with a standardized interview
(the ADI-R) by middle childhood. While there was
some reliance on parental recollections, much of the
information about the children came from concurrent parental reports, observations of behavior, and
assessments of the children. The study was designed
to assess continuity and change in a variety of
characteristics important for the adjustment of the
individuals over the life course. The factors investigated consist of diagnosis, symptoms, adaptive
functioning, and emotional responsiveness. The
literature on these factors and our hypotheses
concerning change and continuity will be reviewed in
the following sections.
Diagnosis and symptoms
Stability of diagnosis and symptoms across the lifespan is central to understanding any disorder and
yet remains a relatively unexplored area in the
study of autism. Mesibov and colleagues (Mesibov,
Schopler, Schaffer, & Michal, 1989) reported decreases in autistic symptoms in a group of 89 children prior to 10 years of age and again after 13 years
of age using the Childhood Autism Rating Scale
(CARS) (Schopler, Reichler, & Renner, 1988). The
CARS is a 15-item rating scale of behaviors asso-
Association for Child Psychology and Psychiatry, 2004.
Published by Blackwell Publishing, 9600 Garsington Road, Oxford OX4 2DQ, UK and 350 Main Street, Malden, MA 02148, USA
402
Corina W. McGovern and Marian Sigman
ciated with autism that is generally completed after
an observation period. All participants were in North
Carolina’s TEACCH (Treatment and Education of
Autistic and Related Communication Handicapped
Children) program, a series of educational and skills
training programs designed to meet the needs of
individuals with autism throughout the lifespan.
Analysis revealed a significant decrease in mean
CARS scores in adolescence, indicating a significant
improvement over time in 9 of 15 areas.
Decreases in CARS scores and DSM-III criteria
from early childhood to early adolescence were also
noted among a group of 76 subjects with autism
spectrum disorder who had not received any special
treatment other than standard services such as early
identification, special education, and speech therapy
(Eaves & Ho, 1996). Despite the lower scores
suggesting an amelioration of autistic symptoms
with age, the entire sample continued to receive a
diagnosis of autism spectrum disorder at follow-up.
In line with these studies, Piven and colleagues
(Piven, Harper, Palmer, & Arndt, 1996) compared
parental reports of autistic symptom expression in
their adolescent and young adult children to retrospective reports when their children were 5 years of
age. These comparisons indicated amelioration of
social and communicative symptoms but not repetitive behaviors and stereotyped interest. Mapping
parental report to DSM-IV criteria, 5 of the 38 participants in their sample no longer met criteria for
autism when they reached the adolescent/young
adult period. However, all the participants were
reported to have persistent autistic characteristics
and substantial social impairments in adolescence
and young adulthood. Based on the findings from
these three studies, we predicted that our sample of
adolescents and young adults would show a moderate improvement in autistic symptoms but that the
majority would continue to meet diagnostic criteria
for autism.
ively associated with intelligence. In addition, we
hypothesized that children who were more involved
with their peers in the mid-school period would
make bigger gains in adaptive behavior than children
who were less socially engaged. This hypothesis was
based on the observation that, during high-level social play, typically developing children are able to
learn to interpret and respond to others’ social cues
and to acquire skills to initiate and maintain a social
interaction (Schuler & Wolfberg, 2000).
Emotional responsiveness
Young children with autism show deficits in responsiveness to the emotions of others (Sigman,
Kasari, Kwon, & Yirmiya, 1992). In a study comparing young children with autism to mental age
matched controls, children with autism looked less
at an adult showing fear, pleasure, distress, or discomfort. In addition, children with autism were rated
as less empathic and showed less inhibition of play
during the adult’s distress (Sigman et al., 1992). In
the mid-school follow-up the children with autism
continued to show less orienting and attending to an
adult’s display of distress and appeared less concerned by the adult’s distress than did children with
Down syndrome and children with other developmental disabilities (Sigman & Ruskin, 1999).
Research suggests that during the period of adolescence and adulthood social interest expands and
social skills continue to develop (Ando, Yoshimura, &
Wakabayashi, 1980; Mesibov & Handlan, 1997;
Mesibov et al., 1989). Given that autistic people tend
to become less socially withdrawn as they age, we
predicted that there would be more responsiveness
to other people’s emotions in adolescence and young
adulthood than was seen in the mid-school years.
Methods
Participants
Adaptive behavior
Adaptive behavior is generally defined as the individual’s ability to meet the personal and social demands of the environment expected by the culture
for someone of his or her age (Nihira, Leland, &
Lambert, 1993). Very little information is available
about developmental change in the adaptive behavior of individuals with autism or about factors that
might predict improvements in adaptive behavior. In
one recent study, individuals with autism improved
in adaptive behavior with age (Freeman, Del’Homme,
Guthrie, & Zhang, 1999). Rates of improvement were
related to initial IQ level for two areas of adaptive
behavior, communication and daily living skills, but
not for socialization skills. Based on the findings of
this study, we expected improvements in adaptive
behavior from middle childhood to adolescence and
young adulthood with rates of improvement posit-
The original sample seen in early childhood consisted of
70 children with autism. Because recruitment took
place over many years, the location for recruiting subjects varied, although most of the subjects came from
either the outpatient or inpatient services of the UCLA
Neuropsychiatric Institute or from local elementary or
preschools. Fifty-one of the original sample of 70 children (73%) participated in the mid-school follow-up
(Sigman & Ruskin, 1999). The children who participated in the mid-school follow-up did not differ from the
non-participants in initial age, gender, mental age, DQ/
IQ, or socio-economic status, although the mothers
of participants were more highly educated than the
mothers of non-participants.
At the most recent follow-up, 48 of the original 70
children (68%) were seen. Of the 48 participants from
the current follow-up, 45 were also seen during the
mid-school follow-up. Six participants from the midschool follow-up did not participate in the most recent
follow-up for the following reasons: one participant had
Continuity and change from early childhood to adolescence
died; one participant who was living in a group home
did not have contact with her parents who were needed
to provide consent; one participant had re-entered the
foster-care system and his original foster mother was
unable to locate him; and two families declined to participate. Three families whom we were unable to locate
at the mid-school follow-up were located and participated in the current follow-up.
The current sample of 48 was composed of 6 females
and 42 males. The ethnic composition of the group was
as follows: 31 Caucasian, 7 African-American, 7 Asian,
2 Hispanic, and 1 other. The majority of participants
(42 of 48) continued to live at home with their family; six
lived in residential facilities. Of the 48 participants seen
at the most recent follow-up, the mean age of the participants during early childhood, middle childhood/
early adolescence, and later adolescence/young adulthood, respectively, was as follows: 3 years, 11 months
of age (SD ¼ 1 year), 12 years, 8 months of age (SD ¼
3 years, 9 months), and 19 years of age (SD ¼ 3 years,
10 months). The degree to which the group seen in later
adolescence/young adulthood was representative of the
original sample was analyzed. Repeated measures
ANOVA revealed that the current participant group did
not differ significantly from the original sample in
terms of chronological age, mental age, developmental
or intelligence quotient, nor language age at intake.
Descriptive information about the sample at intake is
provided in Table 1.
Procedures
Parents were interviewed about their child’s past and
current autistic symptoms and adaptive behavior using
standardized assessments at two time points: when
their children were around 12–13 years of age and
approximately seven years later when their children
were about 19–20 years of age. During the adolescent/
young adult follow-up, participants’ autistic symptoms
were assessed directly using a standardized observation and participants’ cognitive ability was evaluated
using standardized assessments. Approximately half
the families came to UCLA for testing and the other half
of the sample was assessed in their residence based on
parental request and/or distance from UCLA. Testing
did not begin until parents signed a consent form and
participants who had sufficient capacity signed an assent form. Families were paid for their participation.
Testing of all participants was conducted by experimenters who were naive as to the scores of any participant on previous testing.
At the time when this study was initiated in the late
1970s, there were few specialized intervention programs available to most children with autism. For this
403
reason, no assessment of intervention experiences was
carried out at recruitment or at the mid-school followup. However, parents were asked in the current
follow-up whether their children had participated in
intervention programs other than their school experiences. Fourteen parents stated that their children had
been involved in behavioral interventions with a mean
duration of 54 months. Seven individuals were involved
in speech therapy and two were involved in occupational
therapy. Given that these descriptive data were based on
parental recollection of their children’s intervention
programs over a long period of time, no evaluation of the
impact of these interventions was carried out.
Diagnostic measures
Initial diagnoses. Initial diagnoses of the sample
during early childhood changed over time as diagnostic
tools and standards evolved. The first half of the 70
children with autism recruited for this study were
diagnosed by a team of clinicians based on the standards of the Diagnostic and Statistical Manual of Mental
Disorders, third edition (APA, 1980) and third editionrevised (APA, 1987). The team of clinicians who made
the diagnoses all had extensive experience with childhood psychopathology. For the second half of the
sample, clinical judgment was supplemented by two
other diagnostic methods: the Childhood Autism Rating
Scale (CARS) (Schopler et al., 1988), a structured
observation, and the Autism Behavior Checklist (ABC)
(Krug, Arick, & Almond, 1980), used as a parent interview. To be classified as autistic, participants were required to meet diagnostic criteria on two of the three
methods. Thus, all participants had a well-established
diagnosis of autism during early childhood.
Follow-up diagnoses. Parents were interviewed
regarding their children’s past and current autistic
symptoms using the Autism Diagnostic InterviewRevised (ADI-R) (Lord, Rutter, & Le Couteur, 1994) during the middle childhood and adolescent/young adult
follow-ups. The examiners who administered the ADI-R
had all attended training workshops and were trained to
administer and score the ADI-R reliably. Diagnosis of
autism on the ADI-R is primarily based on the child’s
history of behaviors associated with autism. To establish
a current diagnosis in the adolescent and young adult
years, parental report was supplemented with the Autism Diagnostic Observation Schedule-Generic (ADOS-G)
(Lord et al., 2000), a diagnostic instrument based on
current expressed autistic symptoms displayed during a
structured observation. The ADOS-G had not been used
at the middle childhood follow-up because it was still
under development at the time.
Table 1 Descriptive information of sample at intake
Early childhood
Chronological age (mos.)
Language age (mos.)
Mental age (mos.)
IQ
Adolescent/Young adult
n
M
SD
n
M
SD
70
69
70
70
47.21
16.49
23.71
49.31
12.14
7.62
9.81
13.27
48
47
48
48
47.27
17.43
25.04
51.13
12.12
8.62
11.00
12.77
404
Corina W. McGovern and Marian Sigman
Adaptive behavior measures
Adaptive behavior during the middle childhood/early
adolescent follow-up and the adolescent/young adult
follow-up was assessed using the Vineland Adaptive
Behavior Scales (VABS) (Sparrow, Balla, & Cicchetti,
1984). The validity and usefulness of the VABS has
been well established with autistic samples (Perry &
Factor, 1989). The VABS raw scores were converted to
age equivalents using standardized norms. Percentile
rankings were computed using the supplementary
norms for individuals with autism (Carter et al., 1998).
Emotional responsiveness measure
Emotional responsiveness was assessed via parent interview as a question on the ADI-R during the middle
school and adolescent/young adult follow-ups. Parents
were asked if their child ever tried to comfort them if
they were sad, hurt or ill. Parental descriptions of their
child’s behavior were coded on a 4-point scale from 0 to
3, with lower scores indicating more comforting behavior. Parents were also asked to report retrospectively
on their child’s offering comfort when he/she was
4–5 years of age.
Direct observation of peer involvement
School assessments were conducted during the middle
childhood/early adolescent period to determine the
nature of peer interaction of children with autism. Direct observations of the participants were conducted
during structured and unstructured periods of two
different school days to observe ongoing social interaction. Measures of peer interactions were based on the
Peer Play Scale (Howes, 1988) and were coded as nonsocial play, low-level social play, or high-level social
play. Peer interactions were coded continuously every
15 seconds, with the highest level of play recorded
during every 15-second block. Each observation period
lasted at least 30 minutes, ranging from approximately
30 minutes to approximately 60 minutes with an average observation time of 41 minutes. Frequency counts
were transformed to proportion scores for the total
observation period to account for differences in observation periods (Sigman & Ruskin, 1999). Proportion
of time engaged in high-level play was used as the
measure of peer involvement.
Results
Diagnosis and symptoms
At the middle childhood/early adolescent follow-up
the ADI-R was administered to the parents of 51 of
the children originally diagnosed with autism in
early childhood. At the adolescent/young adult
follow-up, parents of 48 participants originally
diagnosed with autism were interviewed with the
ADI-R. Of the 48 parents interviewed at the adolescent/young adult follow-up, 45 had also been
interviewed at the middle childhood/early adolescent follow-up.
Diagnosis. Overall, the results support the ADI-R as
a valid and reliable instrument for diagnosing autism in middle childhood, adolescence, and young
adulthood. Ninety-eight percent (50/51) of participants originally diagnosed with autism in early
childhood met ADI-R lifetime criteria for autism in
middle childhood and 96% (46/48) of participants
met lifetime criteria during adolescence and young
adulthood based on parental report.
Of the 48 participants seen at time 3, 44 were
administered the ADOS-G to assess current expressed autistic symptoms. Four participants were
not administered the ADOS-G because of child noncompliance or parental refusal. Most participants
(40 of 44) met ADOS-G criteria for autism spectrum
disorder. Thirty-three children met criteria for autism, and seven met criteria for the broader classification of pervasive developmental disorder-not
otherwise specified (PDD-NOS). There is little qualitative distinction between autism and PDD-NOS;
rather, PDD-NOS is diagnosed as a case of subthreshold autistic symptomatology.
By adolescence/young adulthood, four individuals
showed sufficient diminution of symptoms to no
longer meet criteria for autism or the broader autism
spectrum on the ADOS-G. These four participants
ranged in age from 15 years, 3 months of age to
18 years, 8 months of age (mean age was 16 years,
7.5 months). All four were verbal and during the
ADOS-G used language in a largely correct fashion,
were able to maintain a back and forth conversation
and showed good use of gestures for non-verbal
communication. However, on an independent measure of language, the Clinical Evaluation of Language
Fundamental-Revised (Semel, Wiig, & Secord, 1987),
all four participants continued to show language
delays (2 years, 3 months; 2 years, 10 months;
6 years exactly; and 6 years, 11 months delayed). To
consider the general level of functioning of these four
individuals we also examined IQ scores. Three of the
four participants had IQs within the average range of
intelligence (IQs of 101, 106, and 110) and one had
an IQ in the borderline range of intelligence (IQ ¼ 77)
as measured using the Stanford–Binet Intelligence
Scale: Fourth Edition (Thorndike et al., 1986). In
terms of functioning in the school system, all four
were mainstreamed in regular classes, two did not
receive any supports or services and two had a resource class to monitor their progress.
During the ADOS-G, all four responded effectively
and appropriately to social situations and presses,
made clear and appropriate social overtures, and
appeared comfortable interacting with the examiner.
To obtain a more global sense of social functioning
for these four individuals, parental reports of social
functioning were considered. According to parental
report on the ADI-R and VABS, two of these participants had a group of friends and a best friend and led
rather active social lives, showing typical to above
average social behavior (70th and 95th percentile on
Continuity and change from early childhood to adolescence
national percentile ranks), whereas the other two
were reported to have just one friend each that they
did not socialize with often and showed marked social deficits relative to their age (at or below the 1st
percentile on national percentile rankings). Thus,
two of the four individuals who did not meet ADOS
criteria for autism appear to have relatively good
social competence. The other two individuals appear
to experience general social difficulties despite not
manifesting overt social-communicative impairments during the semi-structured social situations
in the ADOS-G.
Symptoms. Domain algorithm scores on the ADI-R
for socialization, non-verbal communication, verbal
communication, and repetitive interests and stereotyped behaviors for current and past expression of
autistic symptoms were compared from middle
childhood/early adolescence to adolescence/young
adulthood. Verbal communication questions were
only asked for those participants with phrase
speech. For the purpose of comparing current versus
past recollection across both interviews, items that
were not applicable through adolescence and young
adulthood were dropped from the algorithm (a list of
items used is available from the authors).
Parental reports of autistic symptoms were analyzed for each of the sub-domains of the modified
algorithm using single factor repeated measures GLM
with four levels (mid-school/current, mid-school/
past, adolescent/current and adolescent/past). Only
the data from subjects with all four data points were
included in these analyses. All sub-domains showed
significant differences in reported autistic symptoms: social, F (3, 129) ¼ 39.55, p < .001; non-verbal
communication, F (3, 129) ¼ 21.33, p < .001; verbal
communication, F (3, 45) ¼ 32.18, p < .001; and
repetitive behaviors/stereotyped interests, F (3,
132) ¼ 18.46, p < .001. Tests of within-subjects
contrasts revealed that current symptoms in adolescence were reported as less severe than past symptoms: social, F (1, 43) ¼ 48.83, p < .001; non-verbal
communication, F (1, 43) ¼ 36.53, p < .001; verbal communication, F (1, 15) ¼ 59.27, p < .001;
and repetitive behaviors/stereotyped interests,
F (1, 44) ¼ 49.11, p < .001 (see Table 2).
Current symptoms were reported to improve from
mid-school to adolescence for two of the subTable 2 Mean recollected and current domain scores on the
ADI-R
Current
Impairments
Social*
Nonverbal communication*
Verbal communication*
Repetitive/stereotyped
Behaviors*
*p < .001.
Ever
n
M
SD
M
SD
44
44
16
45
8.48
3.82
3.56
4.36
4.92
3.17
2.78
2.06
12.93
6.23
8.06
6.02
3.23
2.40
2.79
1.60
405
Table 3 Current domain scores on the ADI-R
Mid-school
Social*
Repetitive/
stereotyped behaviors*
Adolescent
n
M
SD
M
SD
44
45
11.75
5.42
4.10
1.66
8.48
4.36
4.92
2.06
*p < .001.
domains: social, F (1, 43) ¼ 45.77, p < .001, and
repetitive behaviors/stereotyped interests, F (1,
44) ¼ 14.74, p < .001 (see Table 3). Thus, there was
an improvement in the areas of socialization and
repetitive behaviors/stereotyped interests based on
parents’ current judgments of their children’s
symptoms from mid-school to adolescence/young
adulthood. There were no significant differences in
reporting of past symptoms across the mid-school
and adolescent/young adult follow-ups.
In order to determine whether the change in
symptoms differed for high-functioning and lowfunctioning individuals, participants were grouped
into high (IQ P 70) and low (IQ/DQ < 70). Independent two-tailed t-tests were conducted for high
(n ¼ 12) and low IQ/DQ (n ¼ 35) groups for change in
ADI-R current versus past expression of autistic
symptoms for each of the modified algorithm domain
totals. High IQ participants were reported to show
larger reductions in reported social impairments than
low IQ/DQ participants, t (45) ¼ )3.85, p < .001;
larger reductions in verbal communicative impairments than low IQ/DQ participants, t (19) ¼ )2.76,
p < .05, and larger reductions in repetitive behaviors
and stereotyped interests than low IQ/DQ participants, t (45) ¼ )4.05, p < .001. Only the nonverbal
communication domain did not show a significant
difference in reported change in autistic symptoms for
high and low IQ/DQ participants (see Table 4). In
terms of improvement in current difficulties, there
was no significant difference for the high and low
samples in the social domain. However, the highfunctioning sample was reported to have improved
more in repetitive and stereotyped interests than the
low-functioning sample, t (43) ¼ 2.75, p < .01.
The lower scores on the ADI-R for parental report of
current behavior vs. their recollections of past behavior at both middle childhood and adolescence/
young adulthood may be interpreted in two ways.
First, it may be that parents simply remember their
children as more impaired earlier. Because we do not
have comparable parent reports in the 3–5-year-old
period, this interpretation cannot be ruled out. The
second hypothesis is that the characteristics of autism diminish from an early age, particularly for
higher-functioning individuals. The strongest evidence for this hypothesis is that parents described
fewer current social impairments and less repetitive
current behaviors and stereotyped interests in adolescence/young adulthood than in middle childhood.
406
Corina W. McGovern and Marian Sigman
Table 4 Mean change in ADI-R symptom domain score and VABS composite score
High IQ ‡ 70
Low IQ < 70
Impairments
Social**
Nonverbal communication
Verbal communication*
Repetitive/stereotyped behaviors**
VABS composite score**
n
M
SD
n
M
SD
35
33
10
35
29
)3.85
)2.15
)2.40
)2.23
3.66
3.56
2.83
2.88
1.94
14.69
12
12
11
12
12
)8.58
)3.50
)5.36
)4.83
46.33
4.19
2.35
2.01
2.04
45.31
*p < .05; **p < .001.
Adaptive behavior
Adaptive behavior in adolescence and young adulthood was representative of individuals with autism,
with average VABS percentile rankings ranging from
62 to 65% compared to a normative sample of adolescents and adults with autism. Individual rankings
ranged from the 1st percentile to the 99th percentile
(Carter et al., 1998). Comparison of age-equivalent
domain scores on the VABS revealed that our sample
did not show a particular weakness in socialization
relative to other areas of adaptive functioning as is
often cited in the literature (Carter et al., 1998; Kraijer,
2000; Rodrigue, Morgan, & Geffken, 1991).
Paired samples two-tailed t-tests of VABS adaptive
behavior composite age equivalent scores revealed
that parents reported that their child’s adaptive behavior improved from the mid-school years to the
adolescent and young adult years, t (40) ¼ 3.07,
p < .01. Further analysis revealed improvement for 2
of the 3 VABS domain scores: daily living skills improved from the mid-school years to the adolescent
and young adult years, t (40) ¼ 3.59, p < .01; socialization also improved from the mid-school years
to the adolescent and young adult years, t (40) ¼
2.39, p < .05. However, scores for the communication domain were not significantly different from the
mid-school years to the adolescent/young adult
years (see Table 5).
In order to determine whether the change in
symptoms differed for high-functioning and lowfunctioning individuals, participants were divided
into high (IQ P 70) and low (IQ/DQ < 70) groups.
Independent samples t-tests revealed that high IQ
participants gained significantly more months in
adaptive behavior on average than low IQ/DQ participants, t (39) ¼ 4.59, p < .001 (see Table 4).
The hypothesis that peer social engagement would
contribute to gains in adaptive behaviors was tested
with Pearson correlation coefficients. These analyses
showed that the percentage of time that the children
with autism spent in high-level play with peers in the
mid-school years predicted gains in VABS adaptive
behavior composite scores, r (32) ¼ .69, p < .001,
daily living skills scores, r (32) ¼ .51, p < .01, communication scores, r (32) ¼ .64, p < .001, and socialization scores, r (32) ¼ .70, p < .001 from the
mid-school years to the adolescent and young adult
years. Even after controlling for mid-school IQ/DQ,
percentage of time engaged with peers in high-level
play in the mid-school years continued to predict
gains in VABS adaptive behavior composite scores, r
(30) ¼ .51, p < .01, communication scores, r (30) ¼
.52, p < .01, and socialization scores, r (30) ¼ .56,
p < .01, from the mid-school years to the adolescent
and young adult years. Thus, the hypothesis that
peer social engagement would predict later adaptive
behavior was confirmed in this study.
Emotional responsiveness
The hypothesis that children with autism would
manifest increasing emotional responsiveness to
others’ distress as they matured was supported by
parental report. Paired samples two-tailed t-tests of
parental reports of their child’s ‘current’ comforting
behavior, assessed during the mid-school years and
assessed again in the adolescent and young adult
years, revealed significant differences across the two
assessment periods, t (44) ¼ 4.957, p < .001. Parents
reported that their children offered more comfort to
them if they were sad, hurt or ill in the adolescent and
young adult follow-up than they did in the mid-school
follow-up. Pearson correlations revealed that par-
Table 5 Adaptive behavior composite age equivalent score
Mid-school
Impairments
Total score*
Daily living*
Socialization*
Communication
*p < .05.
Adolescent/Young adult
n
M
SD
n
M
SD
41
41
41
41
57.65
65.95
46.85
60.98
42.76
39.16
37.31
55.96
41
41
41
41
73.59
88.17
62.93
68.83
61.76
62.25
63.99
64.97
Continuity and change from early childhood to adolescence
ental reports of their child’s ‘current’ comforting behavior in the mid-school assessment was predictive
of parental reports of their child’s ‘current’ comforting
behavior in the adolescent and young adult assessment, r (43) ¼ .52, p < .001, even after controlling for
participants’ IQ/DQ scores in the mid-school years, r
(36) ¼ .37, p < .05.
Discussion
The results of this study provide evidence of both
continuity and change in the developmental trajectory of children with autism from early childhood to
late adolescence. Diagnosis of autism spectrum
disorder shows very strong stability over time in that
almost all the children continue to meet diagnostic
criteria as adolescents and young adults even when
current behavioral observations were used to establish diagnosis. On the other hand, improvement was
reported by parents in terms of symptoms, adaptive
behavior, and behavioral responsiveness to the
emotions of others. Parents mentioned fewer symptoms in adolescence/early adulthood than they
remember when the children were 3–5 years of age.
Moreover, they reported fewer symptoms in the areas
of socialization and repetitive behaviors/stereotyped
interests in adolescence than in middle childhood.
They also described stronger adaptive behavior in all
areas except communication skills and more empathic behavior in adolescence than in childhood.
The behavioral observations largely support the
findings from parental reports.
It must be pointed out that the improvements in
symptoms, adaptive behavior and empathy were
mostly confined to the higher-functioning individuals. The parents of the subset of the sample with
lower IQs reported much less progress than that
reported for children with higher IQs. This finding
lends significance to intervention efforts directed
toward improving these cognitive and language skills
of children with autism in early life. Successful
interventions in the cognitive and communicative
domains are likely to generalize to improvements in
social functioning, even if diagnosis is not altered.
Despite the severe limitations of high-functioning
adolescents with autism, they still enjoy relatively
more adaptive and social skills than low-functioning
adolescents with autism.
The findings also point to another possible area
where interventions may have important pay-offs
and that is on the school playground. In the current
study, the social involvement of the children was
observed and recorded by non-participant observers
who did not intervene in the social environment of
the children. Elementary school children who were
more socially engaged gained more adaptive skills
than less engaged elementary school children. Although it is true that social engagement was partly a
function of cognitive skills, social engagement predicted improvements in adaptive behavior even when
407
the effects of intelligence were statistically constrained. Thus, it appears that engagement with
peers improves the social skills of children with
autism as is true for typically developing children. To
the extent that peer engagement can be increased in
schoolchildren with autism, there may be improvements in their adaptive and social skills.
Improvements in social behavior were reported
from middle childhood/early adolescence to later
adolescence/young adulthood, suggesting the
transition from mid-childhood to adolescence may
coincide with emerging social interest for individuals
with autism. Although it is clear that parents see
their children as less symptomatic as the children
mature, the interpretation of this finding is not clear.
It could be that children actually change in their
behavior, as has been suggested by previous research (Eaves & Ho, 1996; Mesi...
Purchase answer to see full
attachment