Dissociation in performance of children with ADHD and high-functioning autism on a task of sustained attention.

Abstract

Attention deficit hyperactivity disorder (ADHD) and autism are two neurodevelopmental disorders associated with prominent executive dysfunction, which may be underpinned by disruption within fronto-striatal and fronto-parietal circuits. We probed executive function in these disorders using a sustained attention task with a validated brain-behaviour basis. Twenty-three children with ADHD, 21 children with high-functioning autism (HFA) and 18 control children were tested on the Sustained Attention to Response Task (SART). In a fixed sequence version of the task, children were required to withhold their response to a predictably occurring no-go target (3) in a 1-9 digit sequence; in the random version the sequence was unpredictable. The ADHD group showed clear deficits in response inhibition and sustained attention, through higher errors of commission and omission on both SART versions. The HFA group showed no sustained attention deficits, through a normal number of omission errors on both SART versions. The HFA group showed dissociation in response inhibition performance, as indexed by commission errors. On the Fixed SART, a normal number of errors was made, however when the stimuli were randomised, the HFA group made as many commission errors as the ADHD group. Greater slow-frequency variability in response time and a slowing in mean response time by the ADHD group suggested impaired arousal processes. The ADHD group showed greater fast-frequency variability in response time, indicative of impaired top-down control, relative to the HFA and control groups. These data imply involvement of fronto-parietal attentional networks and sub-cortical arousal systems in the pathology of ADHD and prefrontal cortex dysfunction in children with HFA.

Full text

Page 1

Dissociation in performance of children with ADHD and high-
functioning autism on a task of sustained attention
Katherine A. Johnsona,b⁎, Ian H. Robertsona, Simon P. Kellya,c, Timothy J. Silkd,e,f, Edwina
Barryb, Aoife Dáibhisa, Amy Watchorna, Michelle Keaveya, Michael Fitzgeraldb, Louise
Gallagherb, Michael Gillb, and Mark A. Bellgrovea,g
aSchool of Psychology and Trinity College Institute of Neuroscience, Trinity College Dublin, Dublin 2, Ireland
bSchool of Medicine and Health Sciences and Trinity College Institute of Neuroscience, Trinity College
Dublin, Dublin 2, Ireland cCognitive Neurophysiology Laboratory, Nathan S. Kline Institute, Orangeburg,
NY 10962, United States dHoward Florey Institute and Centre for Neuroscience, University of Melbourne,
Australia eAcademic Child Psychiatry Unit, Department of Pediatrics, University of Melbourne, Australia
fDepartment of Psychology, Monash University, Australia gCognitive Neuroscience Laboratory, School of
Psychology and Queensland Brain Institute, University of Queensland, Brisbane, Australia
Abstract
Attention deficit hyperactivity disorder (ADHD) and autism are two neurodevelopmental disorders
associated with prominent executive dysfunction, which may be underpinned by disruption within
fronto-striatal and fronto-parietal circuits. We probed executive function in these disorders using a
sustained attention task with a validated brain-behaviour basis. Twenty-three children with ADHD,
21 children with high-functioning autism (HFA) and 18 control children were tested on the Sustained
Attention to Response Task (SART). In a fixed sequence version of the task, children were required
to withhold their response to a predictably occurring no-go target (3) in a 1–9 digit sequence; in the
random version the sequence was unpredictable. The ADHD group showed clear deficits in response
inhibition and sustained attention, through higher errors of commission and omission on both SART
versions. The HFA group showed no sustained attention deficits, through a normal number of
omission errors on both SART versions. The HFA group showed dissociation in response inhibition
performance, as indexed by commission errors. On the Fixed SART, a normal number of errors was
made, however when the stimuli were randomised, the HFA group made as many commission errors
as the ADHD group. Greater slow-frequency variability in response time and a slowing in mean
response time by the ADHD group suggested impaired arousal processes. The ADHD group showed
greater fast-frequency variability in response time, indicative of impaired top-down control, relative
to the HFA and control groups. These data imply involvement of fronto-parietal attentional networks
and sub-cortical arousal systems in the pathology of ADHD and prefrontal cortex dysfunction in
children with HFA.
© 2007 Elsevier Ltd.
⁎Corresponding author at: School of Psychology and Trinity College Institute of Neuroscience, Trinity College Dublin, Dublin 2, Ireland.
Tel.: +353 1 896 8403; fax: +353 1 671 2006. johnsoka@tcd.ie.
This document was posted here by permission of the publisher. At the time of deposit, it included all changes made during peer review,
copyediting, and publishing. The U.S. National Library of Medicine is responsible for all links within the document and for incorporating
any publisher-supplied amendments or retractions issued subsequently. The published journal article, guaranteed to be such by Elsevier,
is available for free, on ScienceDirect.
1One child was taking 0.5 ml Risperidol per day. One child was taking 30 mg Cipramil per day.
2Hamming-windowing and zero-padding are standard preliminary steps in calculating FFTs and are explained in text books on time-
series analyses, e.g. (Oppenheim, Schafer, & Buck, 1999).
Sponsored document from
Neuropsychologia
Published as: Neuropsychologia. 2007 ; 45(10): 2234–2245.
Sponsored Document
Sponsored Document
Sponsored Document

Page 2

Keywords
Response time; Fast Fourier transform; Variability; Arousal; Response inhibition; Executive function
1  Introduction
Autism and attention deficit hyperactivity disorder (ADHD) are two common, largely genetic,
childhood-onset psychiatric disorders affecting key fronto-striatal and fronto-parietal circuits
that are important for executive function (Courchesne & Pierce, 2005; Filipek et al., 1997;
Mostofsky, Cooper, Kates, Denckla, & Kaufmann, 2002; Schmitz et al., 2006). These two
disorders differ substantially in symptom presentation, but they also share a number of
important features (Sturm, Fernell, & Gillberg, 2004). Despite the exclusion of one disorder
in the formal diagnosis of the other, there appears to be a degree of comorbidity (or a sharing
of symptoms) between the two disorders (Goldstein & Schwebach, 2004; Holtmann, Bölte, &
Poustka, 2005; Stahlberg, Soderstrom, Rastam, & Gillberg, 2004). Both disorders have a strong
genetic component to their aetiology, with heritability estimates of 0.9 for autism and 0.7 for
ADHD (Baron-Cohen & Belmonte, 2005; Faraone et al., 2005); indeed there is preliminary
evidence of genetic linkage in autism and ADHD at chromosomal locations 2q24 and 16p13
(Fisher et al., 2002; International Molecular Genetic Study of Autism Consortium, 2001).
Executive dysfunction is associated with both autism (Geurts, Verte, Oosterlaan, Roeyers, &
Sergeant, 2004) and ADHD (Willcutt, Doyle, Nigg, Faraone, & Pennington, 2005). Some very
specific aspects of executive dysfunction have recently been proposed as endophenotypes in
genetic association studies in ADHD: one in particular is variability in response time (RT) on
tasks that measure sustained attention capabilities (Castellanos & Tannock, 2002; Kuntsi &
Stevenson, 2001). Increasingly, response time variability is being seen as a legitimate marker
of brain pathology, particularly in the frontal areas (Bellgrove, Hester, & Garavan, 2004;
MacDonald, Nyberg, & Bäckman, 2006; Stuss, Murphy, Binns, & Alexander, 2003). It would
be extremely useful to know if a candidate endophenotype is able to distinguish between two
neurodevelopmental disorders. Dissociating autism and ADHD on a task of executive function
may help to define disorder-specific markers for use in genetic association studies. In this
context we sought to determine whether children with ADHD and autism differed on archetypal
executive functions, that of sustained attention and response inhibition, and particularly if
variability in RT specifically differentiated these two groups.
Sustained attention is the endogenous ability to mindfully and consciously process stimuli,
whose non-arousing qualities would otherwise lead to habituation and distraction (Robertson,
Manly, Andrade, Baddeley, & Yiend, 1997). The ability to sustain attention to a task and
produce an appropriate response entails the functioning of the fronto-parietal circuit. The right
dorsolateral prefrontal cortex and the right inferior parietal cortex are activated during sustained
attention tasks (Coull, Frackowiak, & Frith, 1998; Coull, Frith, Frackowiak, & Grasby,
1996; Fassbender et al., 2004; O’Connor, Manly, Robertson, Hevenor, & Levine, 2004; Pardo,
Fox, & Raichle, 1991). In addition, the anterior cingulate, basal ganglia and thalamus are likely
to be involved in regulating and co-ordinating appropriate responses during attentional tasks
(Bradshaw, 2001; Bush et al., 1999; Coull et al., 1998; Konrad, Neufang, Hanisch, Fink, &
Herpertz-Dahlmann, 2006). Finally, areas of the midbrain involved in arousal, including the
reticular formation and the locus coeruleus, may sub-serve the ability to maintain attention to
a task over time (Coull, 1998; Sturm et al., 1999).
There is evidence of anatomical and physiological dysfunction in the fronto-parietal and fronto-
striatal networks in ADHD. Researchers have found bilateral reductions in prefrontal volume
(Castellanos et al., 2002; Filipek et al., 1997; Mostofsky et al., 2002; Sowell et al., 2003),
reduced white matter in the parietal–occipital regions (Filipek et al., 1997) and increased grey
Johnson et al.Page 2
Published as: Neuropsychologia. 2007 ; 45(10): 2234–2245.
Sponsored Document
Sponsored Document
Sponsored Document

Page 3

matter in the inferior parietal cortices (Sowell et al., 2003). Subcortically, there is reduced
anatomical volume of the caudate nucleus, putamen and cerebellum (Castellanos et al., 1994,
1996, 2002). Functionally, the dorsal anterior cingulate has been found to be hypoactive (Bush
et al., 1999; Durston et al., 2003; Rubia et al., 1999) and functioning of the fronto-striatal circuit
(Schweitzer et al., 2003) and prefrontal cortices are abnormal (Bush et al., 1999; Durston et
al., 2003; Rubia et al., 1999). Dysfunction within the parietal lobe, particularly of the right-
hemisphere, has been noted in a number of recent reports (Booth et al., 2005; Silk et al.,
2005). Please refer to (Bush, Valera, & Seidman, 2005) for a recent review.
There is also anatomical and physiological evidence of fronto-striatal (Abell et al., 1999; Carper
& Courchesne, 2005; Courchesne et al., 2001; McAlonan et al., 2005; Silk et al., 2006; Voelbel,
Bates, Buckman, Pandina, & Hendren, 2006) and fronto-parietal dysfunction in autism (Just,
Cherkassky, Keller, Kana, & Minshew, in press; Schmitz et al., 2006). Anatomical research
suggests that children with autism have abnormally large frontal lobe volumes (Carper &
Courchesne, 2000, 2005; Courchesne et al., 2001; Schmitz et al., 2006) that might reflect a
lack of synaptogenesis early in life (Belmonte et al., 2004). There appears to be diminished
grey matter in fronto-striatal and parietal networks (Courchesne et al., 2001) although there is
some debate as to whether anatomical abnormalities exist in the parietal lobes (Abell et al.,
1999; Courchesne, Press, & Yeung-Courchesne, 1993; Koshino et al., 2005; McAlonan et al.,
2002; Schmitz et al., 2006). Functionally, people with autism show a decrease in regional
cerebral blood flow (rCBF) in left prefrontal cortices (Ohnishi et al., 2000). During response
inhibition tasks, adults with autism show increased activation of the frontal and parietal
cortices, compared with controls, despite showing normal behavioural performance on these
tasks (Schmitz et al., 2006). Greater left hemisphere activity in the inferior and orbitofrontal
cortices may indicate an alternative, compensatory mechanism in adult autism (Schmitz et al.,
2006). The recruitment of additional areas of the brain to aid in task performance and to off-
set the negative effects of a deficient network has been illustrated in recent studies, including
those investigating aged individuals (Nielson, Langenecker, & Garavan, 2002) and first-degree
relatives of patients with schizophrenia (Yeap et al., 2006). Executive dysfunction in ADHD
and autism may be related to the compromised workings of the fronto-striatal and fronto-
parietal circuits.
Equivocal evidence exists for sustained attention deficits in both ADHD and autism, although
a greater amount of research has been directed at elucidating sustained attention deficits in
ADHD, when compared with autism (e.g. Bellgrove, Hawi, Gill, & Robertson, 2006; Manly
et al., 2001). Some researchers have argued that in order to show a deficit in sustained attention,
a time-on-task effect for the number of errors must be shown (van den Bergh et al., 2006). In
ADHD research, a number of studies have failed to show time-on-task effects with children
with ADHD (Stins et al., 2005; van der Meere, Wekking, & Sergeant, 1991), although other
studies have demonstrated significant deficits in performance over the duration of the task
(Epstein et al., 2003; Heinrich et al., 2001; Johnson et al., 2007). Only a few studies have
investigated sustained attention in children with autism. Most studies have used the continuous
performance task (CPT) and reported intact sustained attention in autism. Unfortunately, RT
was not recorded in three studies (Buchsbaum et al., 1992; Garretson, Fein, & Waterhouse,
1990; Siegel, Nuechterlein, Abel, Wu, & Buchsbaum, 1995), only median RT was analysed
in one study (Noterdaeme, Amorosa, Mildenberger, Sitter, & Minow, 2001) and two studies
used food or money as obvious (and thus attention-attracting) rewards every time a correct hit
was made by the child (Garretson et al., 1990; Pascualvaca, Fantie, Papageorgiou, & Mirsky,
1998). One recent study has directly compared the performance of children with ADHD and
HFA on the Integrated Visual and Auditory (IVA) Continuous Performance Test (IVA), a test
that combines inattention and impulsivity in the visual and auditory domains (Corbett &
Constantine, 2006). Unfortunately, the dependent variables for this task are a combination of
re-weighted dependent variables (Corbett & Constantine, 2006). Nevertheless, this study found
Johnson et al.Page 3
Published as: Neuropsychologia. 2007 ; 45(10): 2234–2245.
Sponsored Document
Sponsored Document
Sponsored Document

Page 4

little difference in performance of the ADHD and HFA groups except on a measure containing
elements of impulsivity, consistency of RT and sustained attention (VRCQ) (Corbett &
Constantine, 2006). It is difficult to determine which element of the VRCQ was driving this
difference in performance between the two groups. Thus, the nature of sustained attention
deficits in autism remains to be fully determined. Error rates and RT performance both have
the capability of furthering our understanding of sustained attention deficits in children with
ADHD, with autism and in control children.
Recently we described a new procedure to analyse RT data to dissociate variability in RT into
temporal components of fast (moment-to-moment) and slow variability using a Fast Fourier
Transform (FFT) (Johnson et al., 2007), based on the work of Castellanos et al. (2005). The
task employed was the Sustained Attention to Response Task (SART), which requires
participants to withhold a response to an infrequent target and respond to all other stimuli
(Robertson et al., 1997). This task differs substantially from the traditional CPTs, in which
participants monitor a stream of stimuli for the occurrence of an infrequent target, which by
its very nature has an attention-arousing quality. Instead, the SART tests the ability of the
participant to inhibit the automatised act of button pressing when the target appears.
Withholding to a rare target, as opposed to responding to a rare target in CPT tasks, shifts the
automatic response set to the non-targets. Successful withholding of the primed response places
greater demand on the sustained attention system in order to interrupt the ongoing action
(Bellgrove, Hawi, Kirley, Gill, & Robertson, 2005; Robertson et al., 1997). In addition, the
SART provides an ample amount of time-series RT data for analysis using the FFT. Different
forms of variance can be measured from the FFT spectrum, distinguishing distinct components
of RT variability: (1) gradual variability, which has a slow temporal characteristic and (2) trial-
to-trial variability, which has a fast temporal characteristic. In contrast, variability measured
simply as standard deviation over a task run represents the combined influence of these
components but provides no indication of relative contributions. Slow variability is thought to
reflect declining arousal over the course of the task, whereas fast variability may reflect
fluctuations in top-down attentional control (Johnson et al., 2007).
The SART activates the same right fronto-parietal attentional network (Fassbender et al., 2004;
Manly et al., 2003) that appears dysfunctional in ADHD and autism (Giedd, Blumenthal,
Molloy, & Castellanos, 2001; McAlonan et al., 2005; Silk et al., 2005). In this study we
employed fixed- and random-sequence versions of the SART. The Random SART has a greater
response inhibition loading than the Fixed SART, due to the random stimuli presentation. The
Fixed SART places a larger endogenous demand upon the sustained attention system, due to
the predictability of the stimuli presentation. Errors of commission (responding to the no-go
stimuli) on the Fixed SART primarily reflect lapses of sustained attention and to a lesser degree
deficits in response inhibition. Commission errors on the Random SART reflect to a greater
degree response inhibition deficits, in addition to lapses of sustained attention. Errors of
omission (failure to respond to the go-stimuli) in either SART reflect a break from task
engagement and thus are reflective of lapsing attention. The current study therefore employed
both the fixed and random version of the SART to examine the dissociation between sustained
attention and response inhibition in children with ADHD, autism and controls.
The central aim of this study was to assess the ability of children with ADHD, with autism and
normal healthy children on these tasks, in order to examine if these groups differed on aspects
of sustained attention, response inhibition and response time variability (fast and slow). Based
on the previous experimental literature, we hypothesised that the ADHD group would make
more errors of commission and omission and show greater fast and slow variability in RT than
the control group, on both versions of the SART. Based on anatomical and physiological
evidence, we hypothesised that the HFA group, on both versions of the SART, would make
more errors of commission and omission and show greater fast and slow variability in RT than
Johnson et al.Page 4
Published as: Neuropsychologia. 2007 ; 45(10): 2234–2245.
Sponsored Document
Sponsored Document
Sponsored Document

Page 5

the control group and in a similar manner as the ADHD group, but that there may be some
evidence of compensatory mechanisms in the measures.
2 
2.1 
 Method
 Participants
Twenty three children with ADHD (3 females), 21 children with HFA (1 female) and 18 control
children (3 females) participated in the study (see Table 1). There was no significant difference
between the mean ages or in IQ, as measured using four subtests (picture completion,
vocabulary, information, block design) of the Weschler Intelligence Scale for Children
(Weschler, 1992), between the three groups. The data of 15 of the children with ADHD and 5
controls had previously been published and these children were randomly chosen to match the
children with HFA according to age and IQ (Johnson et al., 2007).
Exclusion criteria for participation in the study included known neurological conditions or
pervasive developmental disorders (apart from the presently studied disorders for each group),
serious head injuries and below average intelligence (below 70 on the WISC-III) (Weschler,
1992). Control children were also excluded if they had first degree relatives with ADHD or
HFA. Handedness was measured using the Edinburgh Handedness Inventory (Oldfield,
1971).
The participants with ADHD and HFA were recruited as part of ongoing genetic studies
(Gallagher, Hawi, Kearney, Fitzgerald, & Gill, 2004; Kirley et al., 2002). These participants
were either referred by consultant psychiatrists or recruited through support groups.
Diagnosis for the participants with ADHD was confirmed by psychiatrists using the parent
form of the Child and Adolescent Psychiatric Assessment (CAPA) (Angold et al., 1995).
Twenty-two children with ADHD met DSM-IV diagnosis for Combined-type ADHD and one
for the Inattentive subtype (American Psychiatric Association, 1995). At the time of testing,
all parents of the children with ADHD completed the Conners’ Parent Rating Scale—Revised:
Short Version (CPRS-R:S) (Conners, 1997) and all had ADHD Index T-scores greater than 65
(mean 79, S.D. 6). In addition, parents completed the Asperger Syndrome (and high-
functioning autism) Diagnostic Interview (ASDI) (Gillberg, Gillberg, Råstram, & Wentz,
2001), either at the time of testing (for 16 children) or retrospectively (for 7 children)
approximately 24 months post-testing. The mean of the whole ADHD group on the ASDI was
0.9 (S.D. 1.0). Fifty-six percent of the children with ADHD met diagnostic criteria for
oppositional defiant disorder and 13% met diagnostic criteria for conduct disorder. Any
stimulant medication was withdrawn for at least 24 h prior to testing. Seventy-four percent of
the children with ADHD were stimulant naïve.
The control children were recruited from Dublin schools. Parents of control children completed
the CPRS-R:S (Conners, 1997) at the time of testing and all had ADHD Index T-scores less
than 60 (mean 45, S.D. 5). Seventeen parents also completed the ASDI, retrospectively,
approximately 6 months post-testing (mean 0.1, S.D. 0.2). Consent was obtained from parents
of all children and the experimental work was conducted under the approval of local ethical
committees in accordance with the Declaration of Helsinki.
2.2  Apparatus and procedure
All participants performed the Fixed and Random versions of the Sustained Attention to
Response Task, presented on a laptop computer (Robertson et al., 1997) (please see Fig. 1). In
the Fixed version, a repeating fixed sequence of digits (1–9) was presented. In the Random
version, the digits appeared in a pseudorandom order. For both versions, a single digit appeared
on the screen for 313 ms; a mask was then presented for 125 ms, after which a response cue
Johnson et al. Page 5
Published as: Neuropsychologia. 2007 ; 45(10): 2234–2245.
Sponsored Document
Sponsored Document
Sponsored Document