Downloaded from rstb.royalsocietypublishing.

org on May 17, 2014

Why is joint attention a pivotal skill in autism?

Tony Charman

Phil. Trans. R. Soc. Lond. B 2003 358, doi: 10.1098/rstb.2002.1199, published 28

February 2003

References Article cited in:

http://rstb.royalsocietypublishing.org/content/358/1430/315#related-urls

Email alerting service Receive free email alerts when new articles cite this article - sign up in the box at the
top right-hand corner of the article or click here

To subscribe to Phil. Trans. R. Soc. Lond. B go to: http://rstb.royalsocietypublishing.org/subscriptions

 Downloaded from rstb.royalsocietypublishing.org on May 17, 2014

Published online 16 January 2003

Why is joint attention a pivotal skill in autism?

Tony Charman
Behavioural and Brain Sciences Unit, Institute of Child Health, 30 Guilford Street, London WC1N 1EH, UK
([email protected])
Joint attention abilities play a crucial role in the development of autism. Impairments in joint attention
are among the earliest signs of the disorder and joint attention skills relate to outcome, both in the ‘natural
course’ of autism and through being targeted in early intervention programmes. In the current study,
concurrent and longitudinal associations between joint attention and other social communication abilities
measured in a sample of infants with autism and related pervasive developmental disorders at age 20
months, and language and symptom severity at age 42 months, were examined. Extending the findings
from previous studies, joint attention ability was positively associated with language gains and (lower)
social and communication symptoms, and imitation ability was also positively associated with later langu-
age. Some specificity in the association between different aspects of joint attention behaviours and outcome
was found: declarative, triadic gaze switching predicted language and symptom severity but imperative,
dyadic eye contact behaviours did not. Further, although joint attention was associated with later social
and language symptoms it was unrelated to repetitive and stereotyped symptoms, suggesting the latter
may have a separate developmental trajectory. Possible deficits in psychological and neurological processes
that might underlie the impaired development of joint attention in autism are discussed.
Keywords: autism; joint attention; play; imitation; language; symptom severity

1. INTRODUCTION tion with, the environment (‘experience expectant neural

development’; Greenough et al. 1987).
(a) The role of psychological theory in
This paper summarizes the evidence base and presents
understanding autism
new data that highlight the pivotal role that joint attention
Psychological theory helps us understand autism at two
plays in the development of autism. Two outstanding
levels.1 First, it describes and delineates, in psychological
questions are discussed:
terms, the behaviours that characterize individuals with
autism. Second, and more powerfully, it attempts to
(i) In what way might impairments in the development
explain, at a psychological level, the underlying processes
of joint attention have secondary effects on later
that contribute to the abnormal development seen in indi-
development in autism?
viduals with autism. Abnormal psychological processing is
(ii) What primary pathogenic processes at the psycho-
not the primary pathogenesis that ‘causes’ autism. It is
logical and neurological level might lead to impaired
well established that autism is a neurodevelopmental con-
development of joint attention in autism?
dition whose ultimate aetiology is due to the influence of
genetic and other organic disruptions to brain develop-
‘Pivotal’ can refer both to ‘acting as a fulcrum’ and to
ment and organization (Lord & Bailey 2002). It is also
‘being of crucial importance’ or ‘the thing on which pro-
likely that the behavioural phenotype encompassed by the
gress depends’ (Anonymous 1994). Both meanings are
label ‘autism’ and the broader autism spectrum disorders
relevant to discussion of the pivotal role that joint atten-
includes individuals with different and complex aeti-
tion plays in the psychopathology of autism. Evidence
ologies. However, a ‘dynamic systems approach’ to neuro-
comes from several sources, including parental reports of
developmental disorders (Bishop 1997; Karmiloff-Smith
the earliest recognized signs of abnormality, early video-
1997) highlights ways in which abnormal psychological
tapes of infants who later go on to receive a diagnosis of
development, consequent on abnormal brain develop-
autism, attempts to prospectively screen for autism, longi-
ment, can have secondary effects on later brain and
tudinal studies of early predictors of language and social
psychological development and organization. Primary
outcome and intervention studies.
neurobiological deficits may impact on optimal behav-
In typical development, joint attention behaviours
ioural responses and lead to secondary neurological and
emerge between 6 and 12 months and involve the triadic
psychological disturbance, via the interaction of the
coordination or sharing (‘jointness’; Leekam & Moore
developing brain system with the organization of input
2001) of attention between the infant, another person, and
available to children from their processing of, and interac-
an object or event (Bakeman & Adamson 1984). The term
encompasses a complex of behavioural forms including
gaze and point following, showing and pointing. A distinc-
One contribution of 14 to a Theme Issue ‘Autism: mind and brain’. tion has been made between two different functions that

Phil. Trans. R. Soc. Lond. B (2003) 358, 315–324 315 Ó 2003 The Royal Society
DOI 10.1098/rstb.2002.1199
 Downloaded from rstb.royalsocietypublishing.org on May 17, 2014

316 T. Charman Joint attention in autism

joint attention behaviours serve. Imperative triadic from studies that have attempted to prospectively identify
exchanges serve an instrumental or requesting function, cases of autism using screening instruments (Baird et al.
whereas declarative triadic exchanges serve to share aware- 2001). These have been applied both to general popu-
ness, or the experience, of an object or event (Gómez et lations (Baron-Cohen et al. 1996; Baird et al. 2000; Dietz
al. 1993; Mundy et al. 1993). Individuals with autism are et al. 2001; Robins et al. 2001) and to referred and high-
impaired in the development of both imperative and risk populations (Baron-Cohen et al. 1992; Scambler et al.
declarative acts, although impairments in the latter are 2001). Different aspects of giving, showing, following eye
more severe (Baron-Cohen 1989, 1993; Mundy et al. gaze, and producing and following points, form a key part
1986, 1993; Ricks & Wing 1975; Sigman et al. 1986). of all the screens for autism developed thus far. In the
More recently, it has been shown that the critical distinc- CHAT screening study (Baird et al. 2000), two aspects of
tion may not be the imperative versus declarative level. joint attention behaviour—a lack of gaze monitoring and
Rather, the degree to which a child is monitoring and reg- a lack of pointing for interest—in combination with an
ulating the attention (or attitude) of the other person in absence of simple pretend play at 18 months of age, was
relation to objects and events determines the severity of highly predictive of autism. A proportion of the children
the deficit seen in autism (Mundy et al. 1994; Phillips et prospectively identified had only failed (by parental report
al. 1995; Charman 1998). and health practitioner observation) the items asking
about pointing for interest. An important caution is that
(b) Evidence for the pivotal role of joint attention although the CHAT screen had a high positive predictive
in the early development of autism value its sensitivity was moderate at best, identifying only
The first line of evidence for the central role that joint 38% of cases. It may be that the majority of infants with
attention plays in the development of autism comes from autism did not show impairments in joint attention and
studies that have systematically elicited retrospective par- play behaviours at this age (but might have shown other
ental reports of early symptoms between 12 and 18 developmental impairments and abnormal behaviours not
months (Ohta et al. 1987; Gillberg et al. 1990; Stone et measured in the study). Alternatively, the threshold of
al. 1994). There is some evidence of early abnormalities impairment in these skills may have been set too high (the
in sensory, motor and RSBs, and when such behaviours CHAT asked if children had ever produced such
are present they are highly characteristic of autism (Rogers behaviours).
2001; Charman & Baird 2002). However, most studies Several studies have examined the longitudinal associ-
concur that the best discriminators at this age are likely to ations between joint attention in the pre-school years and
be the social and communicative impairments, in parti- later language and social development. Mundy et al.
cular, joint attention behaviours such as eye contact, gaze (1990) found that joint attention behaviours (alternating
monitoring and response to name (Stone et al. 1997; gaze, pointing, showing and gaze following) measured at
Charman 2000). 45 months were associated with language ability 12
The second source of evidence is the retrospective months later. Social interaction, requesting behaviour,
analysis of home videos taken before children are diag- and initial age, IQ and language ability were not associated
nosed with autism. Adrien et al. (1993) found that within with language at follow-up. Sigman & Ruskin (1999)
the first year children with autism showed impairments in found that responding to joint attention bids measured at
social interaction, lack of social smile and facial the initial time-point was associated with gain in EL at
expression, hypotonia and poor attention. In the second age 12 years. Further, joint attention behaviours measured
year of life, additional impairments included ignoring at 4 years of age were also associated with social and peer
people, preference for aloneness, lack of eye contact and group behaviour 8 years later (Sigman & Ruskin 1999).
lack of appropriate gestures. In a study examining home Stone & Yoder (2001) reported a similar association
videos taken at first birthday parties, Osterling & Dawson between early joint attention ability and later EL ability
(1994) found that children with autism were less likely to from 2 to 4 years of age.
look at others, to show an object or point to objects, and to Another aspect of the pivotal role played by joint atten-
orient to their name, compared with typically developing tion in the development of autism is demonstrated by evi-
controls. In an extension of this study, Werner et al. dence that intervention approaches that have placed an
(2000) found that in videotapes taken between eight and emphasis on the development of non-verbal social–com-
10 months of age children with autism were differentiated municative skills promote enhanced language and social
from typically developing children on the basis of less fre- development (Rogers & Lewis 1989; Koegel 2000; Lord
quent orienting to name. Baranek (1999) found that 2000). Although few, if any, well-controlled randomized
abnormalities in orientation to visual stimuli, aversion to control trials exist, numerous small series case studies have
touch and delayed response to name, all characterized suggested that promoting the NVC competence of chil-
autism (but not developmental delay or typical dren with autism enhances the communicative use of the
development) as early as at nine months of life. In sum- language (Rollins et al. 1998; Kasari et al. 2001).
mary, these studies suggest that alongside a lack of effect, The convergence of these sources of evidence suggests
and in a few cases sensory abnormalities, pre-verbal social that joint attention plays a critical role in the early devel-
communication and social orientating behaviours, includ- opment of autism. Impairments in joint attention behav-
ing joint attention acts, are the most reliably identified iours are among the earliest abnormalities noticed in
early abnormalities (retrospectively) seen towards the end autism, becoming apparent around the end of the first year
of the first year of life in children with autism. of life. Screening instruments that assess (among other
Another demonstration of the importance of joint atten- things) joint attention behaviours can prospectively ident-
tion behaviours in the early development of autism comes ify some cases of autism. Individual differences in joint

Phil. Trans. R. Soc. Lond. B (2003)

 Downloaded from rstb.royalsocietypublishing.org on May 17, 2014

Joint attention in autism T. Charman 317

attention ability relate to later language and social out- Table 1. Age, non-verbal mental age, language scores and
comes over time-periods as long as 8 years, and joint ADI-R scores of participants at both time points.
attention behaviours are emerging as a key target for psy-
cho-educational approaches to early intervention. This time 1 time 2
does not mean that joint attention impairments ‘cause’
autism. However, it does suggest that joint attention is a n = 18 mean n = 18 mean
critical ‘downstream’ effect of earlier brain psychopath- (s.d.) (s.d.)
ology. Understanding why the development of joint atten-
age in months 20.6 (1.3) 42.5 (3.6)
tion skills is impaired in individuals with autism, and the
non-verbal IQ 80.6 (10.1) 83.6 (25.8)
mechanisms by which joint attention behaviours are
ELa raw score 7.3 (3.5) 24.3 (10.5)
related to later outcomes, are important future enterprises RLb raw score 4.8 (2.5) 24.9 (9.5)
for psychological research. RSIc 12.4 (6.3) 12.2 (7.0)
NVCd 9.4 (4.0) 8.8 (4.3)
(c) The present study RSBe 1.8 (1.4) 2.6 (1.8)
The present study took advantage of a small group of
infants (n = 18) with autism and related pervasive develop- a
Reynell EL score.
mental disorders prospectively identified in the CHAT b
Reynell RL score.
screening study (Baron-Cohen et al. 1996, 2000; Baird et c
Reciprocal social behaviour domain of the ADI-R.
al. 2000). We have previously reported findings from a ser-
d
NVC domain of the ADI-R.
ies of experimental tasks of joint attention, attention switch-
e
RSB domain of the ADI-R.
ing, imitation, play and empathy conducted at 20 months
of age (Charman et al. 1997, 1998; Swettenham et al. details of diagnostic assessments). Given the restricted sample
1998). In brief, the group of infants with autism and per- size, we adopted an autism spectrum approach (Lord & Risi
vasive developmental disorder showed very low production 1998) and results were analysed for the group as a whole.
of some behaviours, including empathic responding, pre-
tend play, gaze switching and imitation, in contrast to (b) Experimental measures conducted at age
infants with language delay. The present study reports on 20 months
the longitudinal associations between performance on these Full details of the experimental measures taken at age 20
experimental measures conducted in infancy, and language months are given in Charman et al. (1997, 1998). For the
and behavioural outcomes (symptom severity) from a fol- present analyses, only the key variables entered into the cross-
low-up conducted when the children were aged 42 months. sectional and longitudinal analyses are described.
Although the sample was relatively small, the study pro-
vides a unique contribution because the cohort is signifi- (i) Spontaneous play task
cantly younger than those previously studied. Previous When the child entered the room the following sets of toys were
studies with older samples of children have found positive available (all at once), spread out on the floor: a toy tea-set; a toy
longitudinal associations between early joint attention kitchen stove with miniature pots and pans, spoon, pieces of
behaviour and later language. Consistent with the thesis green sponge; and junk accessories (e.g. brick, straw, rawlplug,
that joint attention is a pivotal skill in the development of cotton-wool, cube, box) and conventional toy accessories (toy
autism, we expected to replicate this finding with our animals, cars, etc.). This combination of objects was based on
younger sample but, in addition, made the prediction that studies by Baron-Cohen (1987) and Lewis & Boucher (1988).
joint attention ability would associate more strongly with The child’s parents and the experimenters remained seated and
language than imitation and play ability. Few studies have offered only minimal and non-specific responses to child-initiated
examined the association between early joint attention approaches. Each child was filmed for 5 min. The presence of
behaviours and later symptom severity but again consist- any functional and pretend play acts on a two-point
ent with our ‘pivotal skill’ thesis we predicted that early scale (0 = no functional or pretend play; 1 = functional play; 2 =
joint attention ability would be (negatively) associated pretend play) was entered into the current analysis.
with later symptom severity.
(ii) Joint attention tasks
Activated toy task
2. METHODS
A series of three active toy tasks based on those described by
(a) Participants Butterworth & Adamson-Macedo (1987) was conducted. The
The participant characteristics are shown in table 1. Non-ver- child stood or sat between their mother and the experimenter.
bal ability was measured using the D and E scales of the Griffiths A series of mechanical toys, designed to provoke an ambiguous
Scale of Infant Development (Griffiths 1986) at age 20 months, response, that is, to provoke a mixture of attraction and uncer-
and either the Griffiths or the Leiter International Performance tainty in the child, were placed one at a time onto the floor of
Scale (Leiter 1952) at age 42 months. A non-verbal IQ was cal- the room 1–2 m from the child. The toys were a robot, which
culated by dividing the age equivalent score by the child’s flashed, beeped and moved around in circular sweeps; a car that
chronological age (MA/CA). RL and EL abilities were assessed followed a circular path around the room; and a pig that made
at both time-points using the Reynell Developmental Language ‘oinking’ noises and shunted backwards and forwards. The toys
Scales (Reynell 1985). At age 42 months, nine subjects met were controlled by the experimenter. They were active for a per-
ICD-10 (World Health Organization 1993) criteria for autism iod of 1 min, during which time they stopped and restarted
and nine subjects met criteria for atypical autism or pervasive twice. The proportion of trials on which the infant produced the
developmental disorder—unspecified (see Cox et al. (1999) for key joint attention behaviour—a gaze switch between the toy and

Phil. Trans. R. Soc. Lond. B (2003)

 Downloaded from rstb.royalsocietypublishing.org on May 17, 2014

318 T. Charman Joint attention in autism

adult (experimenter or parent)—was entered into the current Table 2. Scores for all experimental variables.
analysis.
(a) number of children showing

Goal-detection tasks n %
A series of tasks described by Phillips et al. (1992) were con-
ducted at different times throughout the testing session: (i) The no function or pretend play 4 22.2
blocking task: when the child was manually and visually engaged functional play 12 66.7
pretend play 2 11.1
with a toy, the experimenter covered the child’s hands with his
own, preventing the child from further activity, and held the
(b) percentage of trials key behaviours observed
block for 5 s. This was repeated four times during the session.
(ii) The teasing task: the experimenter offered the child a toy.
mean (%) s.d. (%)
When the child looked at the toy and began to reach out for it,
the experimenter withdrew the toy and held it out of reach for
gaze switch task 45.4 44.2
5 s. The experimenter then gave the toy to the child. This was
goal-detection task 37.7 42.2
repeated four times during the session. The key behaviour imitation task 43.1 35.2
recorded on each trial was whether the child looked up towards
the experimenter’s eyes during the 5 s period immediately after
the block or the tease. The teasing and blocking scores were
3. RESULTS
highly intercorrelated (r = 0.83, p , 0.001). To reduce the num-
ber of variables entered into the analysis, a composite goal- The raw scores of the experimental variables are
detection task score of the proportion of trials in which the infant presented in table 2. The strategy for analysis was to
looked up towards the experimenter on the teasing and blocking present zero-order correlations, followed by partial corre-
trials combined, was entered into the analysis. lations with IQ at 20 months which were partialled out.
Given the sample size, regression analysis was not conduc-
ted.
(iii) Imitation Concurrent associations between the experimental mea-
The materials and method for the procedural imitation task sures and the EL and RL ability at 20 months, and longi-
followed those employed by Meltzoff (1988). The child sat tudinal associations with EL and RL at 42 months, are
opposite the experimenter. Four actions were modelled, all on shown in table 3. Concurrently, only the correlation
objects designed to be unfamiliar to the child. Each act was per- between gaze switching and RL was significant (r = 0.54,
formed three times. At the end of the modelling period (ca. p , 0.05) and remained so when the effects of IQ were
2 min in total), the objects were placed, in turn, in front of the partialled out (r = 0.52, p , 0.05). Longitudinally, the
child. One non-specific prompt (‘What can you do with this?’) presence of functional and pretend play at the initial
was given if the child failed to pick up or manipulate the object assessment was not associated with language ability at 42
at once. The response period was 20 s for each object. The pro- months. The proportion of trials in which a child’s gaze
portion of trials on which the infant imitated the modelled action switched in the joint attention task was significantly corre-
on the objects was entered into the current analysis. lated with both EL (r = 0.55, p , 0.05) and RL (r = 0.74,
p , 0.001). In contrast, the proportion of trials in which
a child made eye contact in the goal-detection tasks was
(c) Symptom severity measured at 20 months and not associated with later language ability. Imitation scores
42 months were significantly correlated with RL (r = 0.63, p , 0.01).
The ADI (ADI-R; Lord et al. 1994) is a semi-structured, stan- When initial IQ was partialled out, the correlations
dardized diagnostic interview that asks parents about the current between gaze switches in the joint attention task and
(and past) functioning. The ADI-R algorithm has three domains EL (r = 0.54, p , 0.05) and RL (r = 0.74, p , 0.001)
or clusters of items that map onto the three symptom areas by remained significant. The partial correlation between imi-
which autism is defined in ICD-10 (World Health Organization tation and RL was also significant (r = 0.65, p , 0.01).
1993): Qualitative impairments in reciprocal social interaction (RSI Concurrent associations between the experimental mea-
or Dimension B), Impairments in verbal and NVC (VNVC or sures and ADI-R symptom domain scores at 20 months
Dimension C), and Repetitive behaviours and stereotyped patterns and longitudinal associations with ADI-R scores at 42
(RSB or Dimension D) (see Lord et al. (1994) for details). ADI- months are shown in table 4. Concurrent performance on
R interviews were conducted with parents of all children at both the play, gaze switch and imitation tasks was significantly
the initial (age 20 months) and follow-up (age 42 months) (negatively) associated with 20-month symptom severity
assessments. For the purposes of the present study the summary measured by the ADI-R algorithm domain scores. Several
algorithm scores (that is, the items that correspond most closely of these associations remained significant for play (NVC:
to characteristic autism symptoms) for each of the three domains r = 20.61, p , 0.01; RSB: r = 20.53, p , 0.05) and for
of behaviour will be entered into the analysis. None of the chil- gaze switch (RSI: r = 20.80, p , 0.001; NVC: r = 20.59,
dren had sufficient language (phrase speech) for the higher-level p , 0.05; RSB: r = 20.69, p , 0.01) when the effect of IQ
verbal items (e.g. stereotyped and idiosyncratic language) to be was partialled out. Fewer associations were found between
scored at 20 months, and only half had sufficient language at performance on the experimental measures at 20 months
42 months. Therefore, the NVC algorithm score was entered and symptom severity at 42 months. In the full corre-
into the analysis for all participants. The ADI-R algorithm lations, performance on the gaze switch task was associa-
domain scores are shown in table 1. ted with scores on the RSI (r = 20.51, p , 0.05) and NVC

Phil. Trans. R. Soc. Lond. B (2003)

 Downloaded from rstb.royalsocietypublishing.org on May 17, 2014

Joint attention in autism T. Charman 319

Table 3. Full and IQ-partialled correlations between the experimental measures and language at 20 and 42 months.

20 months 42 months

EL RL EL RL

(a) full correlations

play 0.16 0.30 0.43 0.34
gaze switch 0.28 0.54¤ 0.55¤ 0.74¤ ¤ ¤

goal-detection composite 20.07 0.06 0.41 0.34

imitation 0.04 0.25 0.46 0.63¤ ¤

(b) IQ-partialled correlations a

play 0.16 0.29 0.42 0.33
gaze switch 0.28 0.52¤ 0.54¤ 0.74¤ ¤ ¤

goal-detection composite 20.09 0.00 0.40 0.32

imitation 0.02 0.18 0.47 0.65¤ ¤

a
IQ at 20 months partialled out.
¤
p , 0.05, ¤ ¤ p , 0.01, ¤ ¤ ¤ p , 0.001.

Table 4. Full and IQ-partialled correlations between the experimental measures and symptom severity at 20 and 42 months.

20 months 42 months

RSIb NVCc RSBd RSIb NVCc RSBd

(a) full correlations

play 20.45 20.59¤ 20.52¤ 20.08 20.27 20.15
gaze switch 20.81 20.62¤ ¤
20.65¤ ¤
20.51¤ 20.66¤ ¤
20.30
goal-detection composite 20.14 20.42 20.38 20.32 20.12 20.27
imitation 20.23 20.48¤ 20.29 20.37 20.52¤ 20.14

(b) IQ-partialled correlations a

play 0.44 20.61¤ ¤
20.53¤ 20.03 20.25 20.14
gaze switch 20.80¤ ¤ ¤
20.59¤ 20.69¤ ¤
20.46 20.65¤ ¤
20.27
goal-detection composite 20.10 20.35 20.41 20.22 0.02 20.23
imitation 20.16 20.31 20.38 20.14 20.34 20.05

a
IQ at 20 months partialled out.
b
Reciprocal social behaviour domain of the ADI-R.
c
NVC domain of the ADI-R.
d
RSB domain of the ADI-R.
¤
p , 0.05, ¤ ¤ p , 0.01, ¤ ¤ ¤ p , 0.001.

(r = 20.66, p , 0.01) domains of the ADI-R and imitation the former, but only longitudinally for the latter. By con-
was associated with NVC (r = 20.52, p , 0.05). When the trast, the other joint attention measure (proportion of
effect of IQ was partialled out, only one correlation trials in which a child looked to the adult in the goal-
remained significant: performance on the gaze switch task detection tasks) and the measure of functional and pre-
was associated with NVC score at 42 months (r = 20.65, tend play were not associated with language ability either
p , 0.01), although the correlation between gaze switch concurrently or longitudinally. The play, gaze switch and
and RSI fell only just short of significance (r = 20.46, imitation measures were all associated with measures of
p = 0.06). Performance on the goal-detection tasks was not symptom severity, across all three domains of symptoms,
associated with symptom severity scores cross-sectionally concurrently at 20 months. However, only gaze switches
or longitudinally. on the active toy task and imitation were associated with
symptom severity at 42 months. The gaze switch measure
was more robustly associated with later symptom severity
4. DISCUSSION
than imitation, in that it was associated with both the RSI
A clear pattern of findings emerged in terms of the con- and the NVC ADI-R domains, and these associations held
current and longitudinal associations between the experi- up when the effect of initial IQ was controlled. By
mental measures at 20 months and language ability and contrast, none of the experimental measures taken at 20
symptom severity at 20 and 42 months. One measure of months was associated with severity of RSB measured by
joint attention (frequency of gaze switches in the active toy the ADI-R at 42 months.
task) and the measure of imitation were associated with These results extend downwards in age the findings of
language ability, both concurrently and longitudinally for previous studies that have shown longitudinal associations

Phil. Trans. R. Soc. Lond. B (2003)

 Downloaded from rstb.royalsocietypublishing.org on May 17, 2014

320 T. Charman Joint attention in autism

between early social communication behaviours and later gaze switch task goal detection task
language ability in samples seen first in the third and eye contact
fourth years of life (Mundy et al. 1990; Stone et al. 1997; shift in attention
Sigman & Ruskin 1999; Stone & Yoder 2001). This dem- distal
onstrates that within the present sample of infants with request/demand
autism individual differences in early social communi- share attention
cation skills relate to one critical outcome measure: langu- triadic
age ability. They also extend previous findings by
examining longitudinal associations with symptom sever- TOY
ity as well as language ability. That is, the greater the
facility a child demonstrated in gaze switching during the
active toy task at 20 months of age, the less severe were
that child’s social and communication symptoms at 42
months. This is consistent with a recent finding that an C E C TOY E
early joint attention behaviour (looking at an object held
out by other), rated retrospectively from home videos of Figure 1. Differences in the form and function of joint
first birthday parties, was associated with symptom sever- attention behaviours in the gaze-switching and goal-detection
ity (rated on the Childhood Autism Rating Scale; Schopler tasks.
et al. (1980)) at age 5 years (Osterling et al. 2002). The
findings are also consistent with many studies that have
demonstrated longitudinal associations between joint terize the gaze switch, but not the goal-detection task, are
attention abilities, including proto-declarative pointing, that the former but not the latter involves shared attention
following eye gaze and pointing, and language learning and a directly referential goal (‘Look at that!’). It has been
and later language ability in typically developing infants suggested that these aspects—shared attention and com-
(Bates et al. 1979; Tomasello & Farrar 1986; Mundy & municative reference—of joint attention behaviour are
Gomes 1996; Carpenter et al. 1998). early evidence for the infant’s emerging understanding of
Taken together, this pattern of findings provides further others as intentional agents, and that understanding the
support for the thesis that joint attention is a pivotal skill mental state of attention in episodes of shared attention
in autism. Only joint attention abilities were related to may be a precursor to understanding mental states or
both later language ability and symptom levels. Further, ‘theory of mind’ ability (Baron-Cohen 1993; Tomasello
the present study demonstrates that the pivotal role of 1995). Some empirical evidence from typically developing
joint attention can be demonstrated in infants with autism, children supports this claim (see Charman et al. 2000).
representing the youngest cohort of children with autism The present study demonstrates that individual differ-
studied to date. However, although one measure of joint ences in these specific aspects of joint attention in infants
attention (gaze switches in the active toy task) was associa- with autism are related to later language ability and social
ted with later language and symptom severity, another and communication symptoms more than other joint
measure of joint attention (looks to the experimenter in attention skills and imitation and play.
the blocking and teasing tasks) was not. This suggests that One other notable finding emerged. Although early joint
the underlying competencies tapped by the two tasks may attention and imitation abilities were related to both later
differ, despite the fact that both have been described language and symptom severity (above and beyond initial
under the umbrella term ‘joint attention tasks’. Previously, IQ), this only held for social and communication symp-
it has been suggested that looking to the experimenter in toms and not for repetitive behaviours and stereotyped
the blocking and teasing tasks might be a questioning patterns. This suggests that the developmental trajector-
(‘What are you doing?’; Phillips et al. (1992)) or an ies, and perhaps at a psychological level the underlying
imperative communicative act (‘Give me that back!’; psychopathology, of these symptom domains may be sep-
Charman 1998). This differs in both nature and form arable. There is other evidence to suggest that this might
from the more clear declarative act involved in switching be the case (see Charman & Swettenham 2001, for a
gaze between the active toy and an adult in the active toy review). Tanguay et al. (1998), for example, found that
task. One suggestion is that the triangulation and shifting three factors derived from factor analysis of the social and
of attention in this task may have a more direct social goal communication items on the ADI-R (‘affective recip-
(‘Look at that!’) and this may involve sharing one’s mental rocity’, ‘joint attention’ and ‘theory of mind’) did not cor-
state of perception with others. relate with scores on the repetitive behaviours and
An analysis of the differences in both the form and the stereotyped interests ADI-R domain score. At least two
function of the two joint attention tasks is shown in figure studies have found that RSBs were identified less consist-
1. Both forms of joint attention behaviour involve eye con- ently in the second and third years of life compared with
tact (or at least looking to an adult’s face) and a shift in older samples of 4- and 5-year-old children with autism
attention, and both might have an imperative function (see Cox et al. 1999; Stone et al. 1999). Consistent with
(‘Start up that toy!’ in the gaze-switch task; ‘Give me that this, two recent studies have found social communication
back!’ in the goal-detection task). The aspects of communi- impairments (including in joint attention) but not execu-
cative form that characterize the gaze switch but not the tive function deficits in 3-year-olds with autism relative to
goal-detection response include the distal position of the controls, in contrast to studies with school-age children
object and the triadic nature of attention focus (child–toy– with autism (see Griffith et al. 1999; Dawson et al. 2002).
adult). The aspects of communicative function that charac- It is possible that in at least a subgroup of children with

Phil. Trans. R. Soc. Lond. B (2003)

 Downloaded from rstb.royalsocietypublishing.org on May 17, 2014

Joint attention in autism T. Charman 321

autism, repetitive, restricted and stereotyped abnormali- frontally mediated neuroaffective motivation systems, that
ties only begin to emerge in children with autism after serve to prioritize social information processing, are appar-
infancy, later than the social and communication deficits ent in development in advance of cognition as the primary
are apparent. regulator of behaviour (Mundy 1995; Mundy & Neal
Thus, although joint attention may be a pivotal skill in 2001). Dawson and colleagues have developed a similar
the development of individuals with autism, it may not be account and provided experimental evidence of a deficit
related to RSBs and restricted interests, which may be due in social orienting in pre-school children with autism
to different underlying pathology at both the psychological (Dawson & Lewy 1989; Dawson et al. 1998; see also Hob-
and neurological level. It is not well understood how the son 1993). Consistent with this account, Leekam et al.
social and communication symptoms in autism ‘hang (2000) found that children with autism were unimpaired
together’ with the repetitive and stereotyped symptoms. in low-level exogenous orienting to objects but they were
This has implications both for our understanding of impaired in exogenous orienting to a social cue (a head
autism and for interventions. Whereas there is evidence turn), and the latter was strongly related to joint attention
that intervention approaches that place an emphasis on behaviour (gaze following). This suggests that impair-
the development of non-verbal social–communicative ments in dyadic social engagement may be present in
skills promote enhanced language and social development autism and may relate to the triadic social engagement
(Rogers & Lewis 1989; Koegel 2000; Lord 2000; Kasari impairments, most notably in joint attention behaviours
et al. 2001), we do not know if such approaches have a (Leekam & Moore 2001). Under such accounts, primary
direct effect on RSBs. Although some of the latter may be neurobiological deficits that underlie impaired social ori-
secondary to communication difficulties they might also enting impact on optimal behavioural responses from as
be expected to ameliorate as communication improves. early as the first few months of life. This may lead to sec-
However, direct interventions that target RSBs may be ondary neurological (and later psychological) disturbance
required (National Research Council 2001). via the interaction of the developing brain system with the
organization of social input available to the children from
(a) What are the origins of the joint attention their processing of, and interaction with, the environment
deficit in autism? (‘experience expectant neural development’; Greenough et
Much theoretical interest has focused on the role of al. 1987) (see Mundy & Crowson 1997; Mundy & Neal
joint attention behaviours as ‘precursors’ to later language 2001).
(Tomasello 1995) and theory of mind development Another clue as to what impairments might underlie
(Charman et al. 2000) in both typically developing chil- disturbances in joint attention behaviours in autism is pro-
dren and children with autism. Acting as a ‘precursor’ vided by a secondary finding from the study of Dawson et
involves either joint attention growing or transforming into al. (1998). Although the impairment in orienting found in
language or theory of mind ability (i.e. it is an earlier form children with autism was greatest for social stimuli (e.g.
of these behaviours) or via experiences gained through the name called), they also showed impairments in orienting
precursor behaviour (e.g. jointly attending to events in the to non-social stimuli (e.g. jack-in-the-box). Results from
world) the child acquires the later abilities. However, rec- studies examining attention orienting to non-social stimuli
ognition that joint attention is not a starting point but are mixed, with some studies finding slow spatial orienting
merely a staging post in early social communicative devel- to cues implicating cerebella dysfunction (Townsend et al.
opment, and hence a ‘postcursor’ of earlier psychological 1996) and others finding no deficit in automatic shifts of
and developmental processes (Tomasello 1995), focuses attention, but impairments in suppression of context-inap-
attention on what earlier impairments underlie the propriate responses implicating executive brain systems in
impaired development of joint attention skills in autism. the prefrontal and parietal cortex (Minshew et al. 1999).
Several candidate precursors have been suggested. Other, more recent, lines of research suggest that the fun-
Using a paradigm measuring spontaneous attention damental cognitive impairments that underlie these abnor-
switching during free play, Swettenham et al. (1998) malities might be at a more basic, low-level perceptual
found that, compared with controls, infants with autism processing level (Happé 1999; Plaisted et al. 1999; Milne
looked less and for shorter duration at people, and more et al. 2002) or at the level of processing and understanding
and for longer duration at objects. They also switched emotions (Baron-Cohen et al. 2000; Klin et al. 2002a).
attention less frequently between social and non-social The association between executive deficits and joint
stimuli. This mirrors recent findings using a sophisticated attention impairments has been directly explored in sev-
eye-tracking methodology to examine where adults with eral recent studies of 3- and 4-year-old children with
autism look when watching film of social interactions autism. As noted above, two recent studies have found no
(Klin et al. 2002b). Thus, individuals with autism (from executive function deficits in 3 year olds with autism rela-
a very early age) may have less exposure to people and the tive to controls, in contrast to studies with school-age chil-
facial, gestural and eye gaze information that, in the typical dren (see Griffith et al. 1999; Dawson et al. 2002). Both
case, draw them into social interaction and an understand- studies also examined the longitudinal associations
ing of the social world. In one sense, this reduced exposure between executive measures and joint attention. Griffith
to social information means that they are less ‘expert’ in et al. (1999) found that performance on a spatial reversal
social interactions than typically developing children. task at age 3 years was associated with joint attention
What might underlie this preference for directing attention ability at age 4 years (but not vice versa) for children with
to objects rather than to people? autism but not for controls. Dawson et al. (2002) found
Mundy and colleagues have proposed a ‘social orien- that tasks tapping ventromedial but not dorsolateral pre-
ting’ model of autistic pathology, whereby disturbances to frontal function were correlated with joint attention

Phil. Trans. R. Soc. Lond. B (2003)

 Downloaded from rstb.royalsocietypublishing.org on May 17, 2014

322 T. Charman Joint attention in autism

ability. They suggested that impairments in rule learning Bakeman, R. & Adamson, L. B. 1984 Coordinating attention
regarding the relations between stimuli and reward that to people and objects in mother–infant and peer–infant inter-
is mediated by the ventromedial system may underlie the action. Child Dev. 55, 1278–1289.
deficits in the development of joint attention (and later Baranek, G. T. 1999 Autism during infancy: a retrospective
video analysis of sensory-motor and social behaviours at 9–
theory of mind) abilities in autism.
12 months of age. J. Autism Devl Disord. 29, 213–224.
Even when these developmental processes are better
Baron-Cohen, S. 1987 Autism and symbolic play. Br. J. Devl
understood, the need to study and understand joint atten- Psychol. 5, 139–148.
tion and other early social communication impairments in Baron-Cohen, S. 1989 Perceptual role-taking and protodeclar-
autism will not disappear. The pivotal role that joint atten- ative pointing in autism. Br. J. Devl Psychol. 7, 113–127.
tion appears to play in the course and outcome of develop- Baron-Cohen, S. 1993 From attention–goal psychology to
ment for individuals with autism, and its potential as a belief–desire psychology: the development of a theory of
target for intervention, will remain, whatever its neurologi- mind and its dysfunction. In Understanding other minds: per-
cal and psychological antecedents. One example is a spectives from autism (ed. S. Baron-Cohen, H. Tager-Flus-
recent study by Siller & Sigman (2002) who found that berg & D. Cohen), pp. 59–82. Oxford University Press.
individual differences in the degree to which mothers Baron-Cohen, S., Allen, J. & Gillberg, C. 1992 Can autism be
synchronized their focus of attention with that of their detected at 18 months? The needle, the haystack, and the
CHAT. Br. J. Psychiat. 161, 839–843.
child were associated with child language gains up to 16
Baron-Cohen, S., Cox, A., Baird, G., Swettenham, J.,
years later. The authors suggest several mechanisms that
Nightingale, N., Morgan, K., Drew, A. & Charman, T. 1996
may underlie this association. These include providing Psychological markers of autism at 18 months of age in a
attentional, social and language experiences that partly large population. Br. J. Psychiat. 168, 158–163.
compensate for the child’s attentional impairments, pro- Baron-Cohen, S., Ring, H. A., Bullmore, E. T., Wheelwright,
viding a more consistent model of being an agent directing S., Ashwin, C. & Williams, S. C. 2000 The amygdala theory
attention (and having intentions) in relation to objects and of autism. Neurosci. Biobehav. Rev. 24, 355–364.
events in the world, or simply being a more fun and motiv- Bates, E., Benigni, L., Bretherton, I., Camaioni, L. & Volterra,
ational partner in social exchanges (Siller & Sigman V. 1979 The emergence of symbols: cognition and communication
2002). Understanding how early psychopathological pro- in infancy. New York: Academic Press.
cesses affect joint attention ability, and what the mech- Bishop, D. V. M. 1997 Cognitive neuropsychology and devel-
anism of transmission is of the associations identified opmental disorders: uncomfortable bedfellow. Q. J. Exp.
between joint attention and later social and language Psychol. A 50, 899–923.
Butterworth, G. E. & Adamson-Macedo, E. 1987 The origins
development, remain important goals for psychological
of pointing: a pilot study. In Annual Conference of the Develop-
research into autism. mental Psychology Section of the British Psychological Society
York, September, UK. [Abstract.]
The author is grateful to his research collaborators on the Carpenter, M., Nagell, K. & Tomasello, M. 1998 Social cog-
CHAT project for many discussions over the years that have nition, joint attention and communicative competence from
helped develop and inform his views on this topic: Simon 9 to 15 months of age. Monog. Soc. Res. Child Dev. 63, 1–
Baron-Cohen, Gillian Baird, Antony Cox, John Swettenham, 143.
Auriol Drew and Sally Wheelwright.
Charman, T. 1998 Specifying the nature and course of the
joint attention impairment in autism in the preschool years:
implications for diagnosis and intervention. Autism 2, 61–79.
ENDNOTE Charman, T. 2000 Theory of mind and the early diagnosis of
1
For reasons of parsimony the term ‘autism’ will be used throughout to autism. In Understanding other minds: perspectives from autism
describe individuals with autism and the related ‘pervasive developmental and developmental cognitive neuroscience, 2nd edn (ed. S.
disorders’ described in DSM-IV (American Psychiatric Association 1994) Baron-Cohen, H. Tager-Flusberg & D. Cohen), pp. 422–
and ICD-10 (World Health Organization 1993), commonly referred to 441. Oxford University Press.
as ‘autism spectrum disorders’.
Charman, T. & Baird, G. 2002 Practitioner review: diagnosis
of autism spectrum disorder in 2- and 3-year-old children.
J. Child Psychol. Psychiat. 43, 289–305.
REFERENCES Charman, T. & Swettenham, J. 2001 Repetitive behaviors and
Adrien, J. L., Lenoir, P., Martineau, J., Perrot, A., Hameury, social-communicative impairments in autism: implications
L., Larmande, C. & Sauvage, D. 1993 Blind ratings of early for developmental theory and diagnosis. In The development
symptoms of autism based upon family home movies. J. Am. of autism: perspectives from theory and research (ed. J. A. Bur-
Acad. Child Adol. Psychiat. 32, 617–626. ack, T. Charman, N. Yirmiya & P. R. Zelazo), pp. 325–345.
American Psychiatric Association 1994 Diagnostic and statistical Hillsdale, NJ: Lawrence Erlbaum Associates.
manual of mental disorders, (DSM-IV), 4th edn. Washington, Charman, T., Swettenham, J., Baron-Cohen, S., Cox, A.,
DC: American Psychiatric Association. Baird, G. & Drew, A. 1997 Infants with autism: an investi-
Anonymous 1994 Collins English dictionary, 3rd edn. Glasgow: gation of empathy, pretend play, joint attention and imi-
Harper Collins. tation. Devl Psychol. 33, 781–789.
Baird, G., Charman, T., Baron-Cohen, S., Cox, A., Swetten- Charman, T., Baron-Cohen, S., Swettenham, J., Cox, A.,
ham, J., Wheelwright, S. & Drew, A. 2000 A screening Baird, G. & Drew, A. 1998 An experimental investigation
instrument for autism at 18 month of age: a six-year follow- of social-cognitive abilities in infants with autism: clinical
up study. J. Am. Acad. Child Adol. Psychiat. 39, 694–702. implications. Infant Mental Hlth J. 19, 260–275.
Baird, G., Charman, T., Cox, A., Baron-Cohen, S., Swetten- Charman, T., Baron-Cohen, S., Swettenham, J., Baird, G.,
ham, J., Wheelwright, S. & Drew, A. 2001 Screening and Cox, A. & Drew, A. 2000 Testing joint attention, imitation
surveillance for autism and pervasive developmental dis- and play as infancy precursors to language and theory of
orders. Arch. Dis. Child 84, 468–475. mind. Cogn. Dev. 15, 481–498.

Phil. Trans. R. Soc. Lond. B (2003)

 Downloaded from rstb.royalsocietypublishing.org on May 17, 2014

Joint attention in autism T. Charman 323

Cox, A., Klein, K., Charman, T., Baird, G., Baron-Cohen, S., Lewis, V. & Boucher, J. 1988 Spontaneous, instructed and
Swettenham, J., Wheelwright, S. & Drew, A. 1999 Autism elicited play in relatively able autistic children. Br. J. Devl
spectrum disorders at 20 and 42 months of age: stability of Psychol. 6, 325–339.
clinical and ADI-R diagnosis. J. Child Psychol. Psychiat. 40, Lord, C. 2000 Commentary: achievements and future direc-
719–732. tions for intervention in communication and autism spec-
Dawson, G. & Lewy, A. 1989 Arousal, attention and the social trum disorders. J. Autism Devl Disord. 30, 393–398.
impairments of individuals with autism. In Autism: nature, Lord, C. & Bailey, A. 2002 Autism spectrum disorders. In
diagnosis and treatment (ed. G. Dawson), pp. 49–74. New Child and adolescent psychiatry: modern approaches, 4th edn
York: Guilford Press. (ed. M. Rutter & E. Taylor), pp. 636–663. Oxford:
Dawson, G., Meltzoff, A. N., Osterling, J., Rinaldi, J. & Blackwell Scientific.
Brown, E. 1998 Children with autism fail to orient to nat- Lord, C. & Risi, S. 1998 Frameworks and methods in diagnos-
urally occurring social stimuli. J. Autism Devl Disord. 28, ing autism spectrum disorders. Mental Retard. Devl Disab.
479–485. Res. Rev. 4, 90–96.
Dawson, G., Munson, J., Estes, A., Osterling, J., McPartland, Lord, C., Rutter, M. & Le Couteur, A. 1994 Autism diagnostic
J., Toth, K., Carver, L. & Abbott, R. 2002 Neurocognitive interview-revised. J Autism Devl Disord. 24, 659–686.
function and joint attention ability in young children with Meltzoff, A. N. 1988 Infant imitation and memory: nine-
autism spectrum disorder versus developmental delay. Child month-olds in immediate and deferred tests. Child Dev. 59,
Dev. 73, 345–358. 217–225.
Dietz, C., Willemsen-Swinkels, S. H. N., Buitelaar, J. K., van Milne, E., Swettenham, J., Hansen, P., Campbell, R., Jeffries,
Daalen, E. & van Engeland, H. 2001 Early detection of H. & Plaisted, K. 2002 High motion coherence thresholds in
autism: population screening. Presentation at the Bienn. children with autism. J. Child Psychol. Psychiat. 43, 255–263.
Meeting Soc. Res. Child Development, Minneapolis, April. Minshew, N. J., Luna, B. & Sweeney, J. A. 1999 Oculomotor
Gillberg, C., Ehlers, S., Schaumann, H., Jakobsson, G., evidence for neocortical systems but not cerebellar dysfunc-
Dahlgren, S. O., Lindblom, R., Bagenholm, A., Tjuus, T. & tion in autism. Neurology 52, 917–922.
Blinder, E. 1990 Autism under age 3 years: a clinical study Mundy, P. 1995 Joint attention and social-emotional approach
of 28 cases referred for autistic symptoms in infancy. J. Child behavior in children with autism. Dev. Psychopathol. 7, 63–
Psychol. Psychiat. 31, 921–934. 82.
Gómez, J. C., Sarria, E. & Tamarit, J. 1993 The comparative Mundy, P. & Crowson, M. 1997 Joint attention and early
study of early communication and theories of mind: social communication: implications for research on inter-
ontogeny, phylogeny and pathology. In Understanding other vention with autism. J. Autism Devl Disord. 27, 653–676.
minds: perspectives from autism (ed. S. Baron-Cohen, H. Mundy, P. & Gomes, A. 1996 Individual differences in joint
Tager-Flusberg & D. Cohen), pp. 397–426. Oxford Univer- attention skill development in the second year. Infant Behav.
sity Press. Dev. 21, 469–482.
Greenough, W. T., Black, J. E. & Wallace, C. S. 1987 Experi- Mundy, P. & Neal, R. 2001 Neural plasticity; joint attention
ence and brain development. Child Dev. 58, 539–559. and autistic developmental pathology. Int. Rev. Res. Mental
Griffith, E. M., Pennington, B. F., Wehner, E. A. & Rogers, Retard. 23, 139–168.
S. J. 1999 Executive functions in young children with
Mundy, P., Sigman, M., Ungerer, J. & Sherman, T. 1986
autism. Child Dev. 70, 817–832.
Defining the social deficits of autism: the contribution of
Griffiths, R. 1986 The abilities of babies. University of Lon-
non-verbal communication measures. J. Child Psychol. Psy-
don Press.
chiat. 27, 657–669.
Happé, F. 1999 Autism: cognitive deficit or cognitive style?
Mundy, P., Sigman, M. & Kasari, C. 1990 A longitudinal
Trends Cogn. Sci. 3, 216–222.
study of joint attention and language development in autistic
Hobson, R. P. 1993 Autism and the development of mind. Lon-
children. J. Autism Devl Disord. 20, 115–128.
don: Lawrence Erlbaum.
Mundy, P., Sigman, M. & Kasari, C. 1993 The theory of mind
Karmiloff-Smith, A. 1997 Crucial differences between devel-
and joint attention in autism. In Understanding other minds:
opmental cognitive neuroscience and adult neuropsychol-
ogy. Devl Neuropsychol. 13, 513–524. perspectives from autism (ed. S. Baron-Cohen, H. Tager-
Kasari, C., Freeman, S. F. N. & Paparella, T. 2001 Early inter- Flusberg & D. Cohen), pp. 181–203. Oxford University
vention in autism: joint attention and symbolic play. Int. Press.
Rev. Res. Mental Retard. 23, 207–237. Mundy, P., Sigman, M. & Kasari, C. 1994 Joint attention,
Klin, A., Jones, W., Schultz, R., Volkmar, F. & Cohen, D. developmental level, and symptom presentation in young
2002a Defining and quantifying the social phenotype in children with autism. Devl Psychopathol. 6, 389–401.
autism. Am. J. Psychiat. 159, 895–908. National Research Council 2001 Educating children with autism.
Klin, A., Jones, W., Schultz, R., Volkmar, F. & Cohen, D. Committee on educational interventions for children with autism.
2002b Visual fixation patterns during viewing of naturalistic Division of Behavioral and Social Sciences and Education.
social situations as predictive of social competence in indi- Washington, DC: National Academy Press.
viduals with autism. Arch. Gen. Psychiat. 59, 809–816. Ohta, M., Nagai, Y., Hara, H. & Sasaki, M. 1987 Parental
Koegel, L. K. 2000 Interventions to facilitate communication perception of behavioral symptoms in Japanese autistic chil-
in autism. J. Autism Devl Disord. 30, 383–391. dren. J. Autism Devl Disord. 17, 549–563.
Leekam, S. & Moore, C. 2001 The development of attention Osterling, J. & Dawson, G. 1994 Early recognition of children
and joint attention in children with autism. In The develop- with autism: a study of first birthday home videotapes. J.
ment of autism: perspectives from theory and research (ed. J. A. Autism Devl Disord. 24, 247–257.
Burack, T. Charman, N. Yirmiya & P. R. Zelazo), pp. 105– Osterling, J. A., Dawson, G. & Munson, J. A. 2002 Early rec-
129. Hillsdale, NJ: Lawrence Erlbaum Associates. ognition of 1-year-old infants with autism spectrum disorder
Leekam, S. R., López, B. & Moore, C. 2000 Attention and versus mental retardation. Devl Psychopathol. 14, 239–251.
joint attention in preschool children with autism. Devl Psy- Phillips, W., Baron-Cohen, S. & Rutter, M. 1992 The role of
chol. 36, 261–273. eye-contact in goal-detection: evidence from normal toddlers
Leiter, R. G. 1952 Leiter international performance scale. Wood and children with autism or mental handicap. Devl Psycho-
Dale, IL: Stoetling. pathol. 4, 375–384.

Phil. Trans. R. Soc. Lond. B (2003)

 Downloaded from rstb.royalsocietypublishing.org on May 17, 2014

324 T. Charman Joint attention in autism

Phillips, W., Gómez, J. C., Baron-Cohen, S., Laá, V. & Rivi- Stone, W. L., Hoffman, E. L., Lewis, S. E. & Ousley, O. Y.
ère, A. 1995 Treating people as objects, agents, or ‘subjects’: 1994 Early recognition of autism: parental reports vs. clinical
how children with autism make requests. J. Child Psychol. observation. Arch. Ped. Adol. Med. 148, 174–179.
Psychiat. 36, 1383–1398. Stone, W. L., Ousley, O. Y. & Littleford, C. D. 1997 Motor
Plaisted, K. C., Swettenham, J. & Rees, L. 1999 Children with imitation in young children with autism: what’s the object?
autism show local precedence in a divided attention task and J. Abn. Child Psychol. 25, 475–485.
global preference in a selective attention task. J. Child Psy- Stone, W. L., Lee, E. B., Ashford, L., Brissie, J., Hepburn,
chol. Psychiat. 40, 733–742. S. L., Coonrod, E. E. & Weiss, B. 1999 Can autism be diag-
Reynell, J. K. 1985 Reynell developmental language scales, 2nd nosed accurately in children under three years? J. Child Psy-
edn. Windsor, UK: NFER Nelson. chol. Psychiat. 40, 219–226.
Ricks, D. N. & Wing, L. 1975 Language, communication, and Swettenham, J., Baron-Cohen, S., Charman, T., Cox, A.,
symbols in normal and autistic children. J. Autism Child Baird, G., Drew, A., Rees, L. & Wheelwright, S. 1998 The
Schizophrenia 5, 191–222. frequency and distribution of spontaneous attention shifts
Robins, D. L., Fein, D., Barton, M. L. & Green, J. A. 2001 between social and non-social stimuli in autistic, typically-
The modified checklist for autism in toddlers: an initial study developing and non-autistic developmentally delayed
investigating the early detection of autism and pervasive infants. J. Child Psychol. Psychiat. 39, 747–754.
developmental disorders. J. Autism Devl Disord. 31, 131– Tanguay, P. E., Robertson, J. & Derrick, A. 1998 A dimen-
144. sional classification of autism spectrum disorder by social
Rogers, S. 2001 Diagnosis of autism before the age of 3. Int. communication domains. J. Am. Acad. Child Adol. Psychiat.
Rev. Mental Retard. 23, 1–31. 37, 271–277.
Rogers, S. & Lewis, H. 1989 An effective day treatment model Tomasello, M. 1995 Joint attention as social cognition. In Joint
for young children with pervasive developmental disorders. attention: its origins and role in development (ed. C. Moore &
J. Am. Acad. Child Adol. Psychiat. 28, 207–214. P. Dunham), pp. 85–101. Hillsdale, NJ: Lawrence
Rollins, P. R., Wambacq, I., Dowell, D., Mathews, L. & Reese, Erlbaum Associates.
P. B. 1998 An intervention technique for children with autis- Tomasello, M. & Farrar, M. J. 1986 Joint attention and early
tic spectrum disorders: joint attentional routines. J. Com- language. Child Dev. 57, 1454–1463.
mun. Disord. 31, 181–193. Townsend, J., Harris, N. S. & Courchesne, E. 1996 Visual
Scambler, D., Rogers, S. J. & Wehner, E. A. 2001 Can the attention abnormalities in autism: delayed orienting to
checklist for autism in toddlers differentiate young children location. J. Int. Neuropsychol. Soc. 2, 541–550.
with autism from those with developmental delays? J. Am. Werner, E., Dawson, G., Osterling, J. & Dinno, N. 2000 Brief
Acad. Child Adol. Psychiat. 40, 1457–1463. report: recognition of autism spectrum disorder before one
Schopler, E., Reichler, R. J., DeVellis, R. & Daly, K. 1980 year of age: a retrospective study based on home videotapes.
Towards objective classification of childhood autism: child- J. Autism Devl Disord. 30, 157–162.
hood autism rating scale (CARS). J. Autism Devl Disord. 10, World Health Organization 1993 Mental disorders: a glossary
91–103. and guide to their classification in accordance with the 10th
Sigman, M., Mundy, P., Sherman, T. & Ungerer, J. 1986 revision of the international classification of diseases: research
Social interactions of autistic, mentally retarded and normal diagnostic criteria (ICD-10). Geneva: World Health Organi-
children and their caregivers. J. Child Psychol. Psychiat. 27, zation.
647–656.
Sigman, M. & Ruskin, E. 1999 Continuity and change in the GLOSSARY
social competence of children with autism, Down syndrome ADI: autism diagnostic interview
and developmental delays. Monogr. Soc. Res. Child Dev. 64,
ADI-R: autism diagnostic interview—revised
1–114.
CHAT: CHecklist for Autism in Toddlers
Siller, M. & Sigman, M. 2002 The behaviors of parents of chil-
dren with autism predict the subsequent development of EL: expressive language
their children’s communication. J. Autism Devl Disord. 32, IQ: intelligence quotient
77–89. NVC: non-verbal communication
Stone, W. L. & Yoder, P. J. 2001 Predicting spoken language RL: receptive language
in children with autistic spectrum disorders. Autism 5, RSB: repetitive and stereotyped behaviour
341–361. RSI: reciprocal social interaction

Phil. Trans. R. Soc. Lond. B (2003)