Appears in von Hofsten & Rosander (Eds.

)
Progress in Brain Research, Vol. 164

Infancy and autism: progress, prospects, and

challenges
Mayada Elsabbagh* and Mark H. Johnson

Centre for Brain and Cognitive Development, Birkbeck College, University of London, Henry Wellcome
Building, WC1E 7HX, London

Abstract: We integrate converging evidence from a variety of research areas in typical and atypical
development to motivate a developmental framework for understanding the emergence of autism in infancy and
to propose future directions for a recent area of research focusing on infant siblings of children with autism.
Explaining the cognitive profile in autism is best achieved through tracing the process through which associated
symptoms emerge over development. Understanding this process would shed light on the underlying causes of
this multifaceted condition through clarifying how and why a variety of risk factors, single or in combination,
exert an impact on the resulting phenotype. We emphasize the importance of integrating theoretical models of
typical development in understanding atypical development and argue for the need to develop continuous and
individually valid measures for at-risk infants both for predictors and outcomes of autism symptoms.
Keywords: autism; autism spectrum disorder; social orienting; executive function; perception; development;
infancy; infant siblings; core deficits

Autism and infancy: an overview developmental accounts have proved difficult to

empirically verify. Because a confirmed diagnosis
Much research into autism over the past decades of autism can only be made from around three years
has presented a neat developmental story: a form of of age, our knowledge of the early neural,
neurodevelopmental deficit, commonly attributed to behavioural, and cognitive profile in autism is very
brain networks subserving social cognition, leads to poor, and virtually nothing is known about the
decreased attention to, or interest in, the social underlying causes, or the process through which
world. This early lack of attention to or interest in these deficits emerge. As a result, and despite
social stimuli interferes with the emergence of substantial parallel advances in research on typical
critical developmental milestones relevant for development in infancy and research on children
social cognition such as shared attention. These and adults with autism, until very recently the two
cascading influences eventually lead to deficits in disciplines have rarely been combined.
theory of mind, and preclude the typical The purpose of the current chapter is two-fold:
development of socio-communicative skills. First, to review the recent empirical evidence of the
Compelling as they are, the elements of such neural, cognitive, and behavioural profile of autism
in the first few years of life, and second, to situate
this fragmentary knowledge within a developmental
*Corresponding author. Tel.: +44 207 631 6231; Fax: framework relevant for the study of autism in
+44 207 631 6587 infancy, specifically through the study of at-risk
E-mail addresse: m.elsabbagh@bbk.ac.uk infant siblings of children diagnosed with autism.

355
356

We propose that autism in infancy is best Family aggregation studies indicate that the risk of
understood through a combination of genetic, autism for siblings is increased to 5- 10%.
neural, and cognitive risk factors, the presence and One of the most intriguing manifestations of
severity of each as well as their interactions over the genetic risk for autism is that it appears to result
development, give rise to the typical yet in an extended phenotype. This refers to
considerably variable phenotype observed in behavioural and brain characteristics associated
autism. This approach is valuable, not only for with the autism phenotype found not only in
understanding how symptoms unfold over affected individuals, but also in their genetic
development, but also for understanding the relatives, and to a lesser degree in the general
mechanisms which drive and constrain typical population (Bolton et al., 1994; Dawson et al.,
development. In the remainder of this section we 2002b; Pickles et al., 2000). This implies that
will present an overview of how autism is defined several characteristics of autism are not themselves
and identified. In the next section, we will describe uncommon or atypical, but that the number and
some of the theoretical approaches to the severity of these symptoms within an individual
neurocognitive study of this disorder, and review determine whether they manifest as normative
the current evidence relating to them. Finally, in the differences in processing styles or in personalities,
last section, we will discuss a recent approach as opposed to diagnostic symptoms.
focusing on the study of at-risk infants, focusing on Aside from genetic risk, several comorbid
the prospects and challenges of this approach. conditions that constitute diagnosable medical
Autism (or Autism Spectrum Disorders, ASD) conditions are associated with autism, including
is usually defined using a triad of impairment in epilepsy, Moebius, Fragile X, and tuberous
social interaction, communication, and the presence sclerosis. The study of some of these comorbid
of restricted and repetitive patterns of behaviour conditions has provided some evidence that autism
(American Psychiatric Association, 1994). Some may also be associated with prenatal ‘accidents’.
form of autism currently affects 1 in 100 children, For instance, exposure to teratogens in highly
with a higher prevalence in males relative to specific time windows may result in an increased
females (Baird et al., 2006). A confirmed diagnosis risk for autism (Arndt et al., 2005; Miller et al.,
of autism is rarely given before three years of age; 2005). However, the precise contribution of such
only after this age are the core symptoms environmental factors remains unclear.
sufficiently clear and can developmental delay or
other comorbid conditions be clearly ruled out. Approaches to the neurocognitive study of
Deficits or delays in the expected language and autism in infancy
social milestones found in typical development,
e.g., no babbling by 12 months, no pointing, lack of A wide range of theoretical approaches have
social smile, or the loss of language or social skills attempted to specify the “core deficits” in autism in
are considered “red flags” for autism (Filipek et al., adults and older children. This refers to aspects of
1999; Sigman et al., 2004; Zwaigenbaum, 2001), cognition and their neural underpinnings that
but no tool as yet has provided a reliable and stable present a unifying and parsimonious
diagnosis in the first years of life. characterization and/or causal explanation of
Although the precise causes of autism remain autism. Several core deficits have been proposed.
unknown, there is general consensus that it can These include deficits in social orienting, attention,
emerge as a result of interactions among several and motivation; deficits in executive function; and
genetic, prenatal, postnatal, and external differences in low-level perceptual and attentional
environmental factors. Genetic factors appear to processes. Of course, these are not mutually
play a critical role in autism. The genetic basis of exclusive possibilities, at least as far as
autism have been confirmed through converging characterizing the disorder. However, as we will
lines of evidence from familial aggregation, twin discuss in more detail in the next section, these
studies, and overlap with other conditions (Autism characterizations are frequently used to infer
Genome Project, 2007; Bailey et al., 1995; Muhle underlying causal mechanisms, where other
et al., 2004; Newschaffer et al., 2002). abnormalities are viewed as “secondary” or
Nevertheless, there is still no clear pattern of “peripheral” deficits, resulting from the core deficit.
inheritance nor a set of specific genes identified. Our focus in this section is to evaluate, against the
available evidence, the extent to which these
357

approaches can provide plausible developmental the model explaining its abnormal functioning in
models for the emergence of autism over the first autism. It will become apparent that these
years of life. approaches often lack a distinction between
Before we begin, it is important to address childhood and adulthood profiles, with the
some relevant methodological issues: How is a expectation that a core deficit will be present early
disorder that is only diagnosed after three years in life, but manifest slightly differently at different
studied in infancy? The fact that a valid and stable ages. Because our focus is to draw the implications
diagnosis of autism cannot be obtained prior to these models have to the development of autism in
three years of age, has not been a deterrent in infancy, we will evaluate each approach against a
recognizing the importance of this area of research. representative review of the literature focusing on
Consequently, several complementary approaches studies of young children (mostly below 6 years)
for investigating autism in the early years have been already diagnosed with autism, as well as selective
pursued, yielding valuable results. Some studies findings from the retrospective approaches
have screened for autism in the general population mentioned above. Several methodological issues,
or in high-risk populations, sampling large groups relevant for the study of special populations, e.g.,
of infants in order to develop appropriate clinical age of participants, their developmental level, the
tools (Baron-Cohen et al., 1992; Lord, 1995; Stone choice of control groups, etc., are frequently
et al., 1999; Wetherby et al., 2004). These children invoked when results across studies do not
are then followed up to the age when some receive converge. To allow comparisons across these
a diagnosis and others do not. Studies of this sort studies, we will include several of these
are effective in providing some general methodological details such as sample
characteristics of autism early on, but are limited as characteristics, information regarding how the
far as in-depth or experimental assessment, as they groups were matched, and brief descriptions of the
require very large samples. Other studies are stimuli and measures used in each study.
retrospective, relying on parental descriptions of
their children prior to diagnosis (Werner et al., Social vs. non-social orienting, attention, and
2005; Wimpory et al., 2000) or on analysis of home cognition
videos of children later diagnosed with autism
(reviewed in Palomo et al., 2006). In general, Autism is primarily defined and diagnosed based on
parents appear to be able to provide some valid and impairment in social interactions and
specific descriptions of developmental delays and communication. Consequently, it is no surprise that
deficits in their children (Filipek et al., 2000), but several theoretical approaches have invoked deficits
these studies are inevitably subject to recollection in the neural systems mediating social interactions
bias. Data from home videos help overcome this as characterizing the disorder. In general, these
problem but are limited by variations in contexts approaches draw on contrasts in the encoding,
and the lack of experimental control. An alternative processing, and interpretation of the social, relative
approach is to study young children as soon as they to the physical, world in autism.
are diagnosed, investigating various aspects of In typical adults social stimuli and social
perception and cognition (several of these will be contexts are processed by specialized neural
presented shortly). Recently, another approach, systems including the fusiform area, the superior
focusing on the study of infants at-risk for autism temporal sulcus, and the orbitofrontal cortex
prior to the onset of overt symptoms, has overcome (Adolphs, 2003). Further specialization within this
some of the major challenges faced by these other “social brain” network has been described. For
approaches (Zwaigenbaum et al., 2007). This instance, the fusiform area appears to be selective
approach will be discussed in detail in the next for faces relative to objects. Detecting the
section. movement of eye gaze in another’s face activates
In what follows we will present some dominant the superior temporal sulcus (STS) and the
theoretical approaches to autism (social orienting, neighbouring MT/V5 region (Allison et al., 2000).
executive function, and low-level perceptual and The basis of these patterns of cortical specialization
attentional processes). For each putative core is currently the subject of debate. Some view
deficit, we will briefly describe the neural specialized areas as “pre-wired” for specific
underpinnings of the target ability in typical adults, stimuli, e.g., the FFA for faces, whereas others
its functional development in typical infants, and suggest that the specialization is driven by
358

perceptual expertise in a given category of stimuli Beyond this consistency in the defining
(Gauthier and Nelson, 2001; Tarr and Gauthier, features of autism, the functioning of other aspects
2000). An intermediate view is that the typical of the social brain network and the developmental
patterns of cortical specialization for social stimuli origins of these difficulties is subject of continued
emerge during development as a result of debate. This is clearly illustrated in research on face
interactions and competition between regions that processing in autism. Behavioral studies on the
begin with only broadly tuned biases (Johnson, scanning and processing of faces (reviewed in
2001). According to the latter view, some skills (Jemel et al., 2006; Sasson, 2006) are equivocal
may be present at birth, while others become regarding the nature of the deficits found in autism.
increasingly specialized over time, showing gradual On the other hand, there is some consensus that
developmental changes leading to adult-like performance of individuals with autism differs from
processing patterns. controls insofar as the tasks employed tap into
Johnson and Morton (1991) proposed that a differences in default encoding and processing
subcortical orienting system (which they termed strategies. For instance, individuals with autism
“conspec”) initially biases the newborn to attend scan faces less than typically-developing controls,
towards faces. While this putative system is based they focus on different regions of the face, and they
on simple low-spatial frequency patterns display greater reliance on featural relative to
characteristic of faces, it is sufficient to bias the configural analysis, whereas the opposite strategy is
input to developing cortical visual areas (Johnson, used more by typical individuals. Depending on the
2005). In response, some of these cortical areas task demands, these atypical strategies may or may
will increase their specialization for faces and not produce differences in overall performance
related social stimuli. A consequence of this will levels relative to those found in controls. This
be increasingly focal patterns of cortical activation implies that individuals with autism may achieve an
with increasing age. Several lines of evidence equivalent level of overall competence on face
indicate that infants as young as two months of age processing tasks, albeit relying on different
show activation in similar areas as those found in underlying strategies. Neuroanatomical studies
adults in response to faces. However, the activated (reviewed in Schultz, 2005) support this
network in infants is broader encompassing areas conclusion. Atypical activation patterns in the
such as the prefrontal cortex (de Haan et al., 2003; fusiform and the amygdala are usually found in
Johnson et al., 2005; Tzourio-Mazoyer et al., 2002). individuals with autism in response to faces.
These findings from the adult and infant Furthermore, these neural patterns appear to
literature motivated the view of autism as a case of correlate with atypical behavioural scanning
damage to, or lack of specialization of, this social patterns in these individuals (Dalton et al., 2007;
brain network (Dalton et al., 2005; Dawson et al., Dalton et al., 2005).
2005; Johnson et al., 2005; Klin et al., 2003; Based on these adult approaches,
Schultz, 2005). Behaviourally, this is manifest in a developmental models of autism have focused on
wide range of impairments or atypical processing of understanding the precursors of their social
social stimuli and contexts. Not only are social- difficulties including orienting to social stimuli and
communicative difficulties diagnostic of the events, joint attention, imitation, and social
disorder, but they have also been clearly interactions. At the neural level, some
demonstrated through a variety of experimental developmental models attribute these social
paradigms and procedures. For instance, one of the difficulties to impairment of mechanisms such as
most popular views of the core deficit in autism is the conspec, which bias infants to orient towards
the marked impairment in inferring other people’s and assign value to socially-relevant stimuli.
beliefs and desires, or theory of mind (Baron-Cohen Different models propose alternative explanations
et al., 1985; Leslie and Thaiss, 1992), even for of the origins of these difficulties. Some propose
individuals with average or above average cognitive that there are impairments in the cortical
ability. Similarly, studies examining the neural mechanisms subserving the social brain network,
underpinning of these behavioural differences, have such as the FFA. Other models propose that
documented atypical activation patterns associated subcortical impairment in the amygdala leads to
with such tasks (Baron-Cohen et al., 1999; Happé et reduced attention to (Dawson et al., 2005; Schultz,
al., 1996). 2005) or even aversion of (Dalton et al., 2005)
socially relevant stimuli such as eye gaze.
359

Cascading effects would then follow, precluding Face and Gaze processing
the development of social-communicative skills in
the typical fashion. With these theoretical proposals Available studies of face processing in young
in mind, we will now present the available evidence children with autism point to a pattern of
from the early years including studies on social impairment in comparison to both children of the
orienting and scanning (Table 1); face, gaze, and same age as well as developmentally matched
emotion processing (Table 2); imitation and children (see Table 2 for summary). Difficulties in
understanding of contingency (Table 3); joint face processing include discrimination as well as
attention and communication (Table 4). understanding of emotion and eye gaze (but see
Chawarska et al., 2003). Interestingly, there is
Social orienting and scanning evidence for improvement in behavioural
performance in slightly older children by around 4
Retrospective studies looking back at the first two years of age (Chawarska and Volkmar, 2007).
years of life consistently show less orienting Consistent with these results, a recent review
towards social stimuli and a reduced response to concluded that impairments in face processing
name calling from as early as 9 months (Maestro et appear to be clearer in children relative to adults
al., 2005; Mars et al., 1998; Osterling et al., 2002; with autism (Sasson, 2006). However,
Werner and Dawson, 2005) or younger (Maestro et electrophysiological studies suggest that atypical
al., 2002) in children later diagnosed with autism as underlying neural processes mediate performance
compared to those later diagnosed with in both children (Table 2) and adults with autism
developmental delay. Consistent with this, (Dawson et al., 2004b; Dawson et al., 2005).
experimental studies with already diagnosed
children show that they orient less to social stimuli Imitation and understanding of contingency
such as people, voices, etc. as compared to non-
social stimuli such objects, sounds, etc., relative to As evident from the studies summarised in Table 3,
both children of the same age and developmental imitation is an area of serious difficulty for young
level (see Table 1 for summary). It is worth noting children with autism. However, it appears that
that these latter studies also report an overall certain forms of imitation may pose more difficulty
decreased level of orienting to both social and non- than others. For instance, manual imitation is less
social stimuli, but the impairment is greater for the problematic for young children with autism than
social stimuli. Similarly, in retrospective studies, imitation of actions on objects. Interestingly, these
excessive looking at objects is a reliable risk marker studies suggest that imitative responses, and
but only when combined with reduced social interactions with others in general, may be
orienting (Osterling et al., 2002). One possibility is facilitated in children with autism, if they are based
that the increased interest in objects is a on perfect contingency, rather than on high, but
consequence of decreased interest in people. imperfect, contingency patterns preferred by control
Compared with numerous studies documenting children. This is consistent with studies of older
atypical face scanning in older children and adults, children (Gergely and Watson, 1999). High, but
only one study examined this in young children imperfect, contingency characterises most social
with autism. Anderson and colleagues (2006) found interaction and this may therefore underlie a greater
no differences in visual scanning of social stimuli difficulty in dealing with people rather than objects
but they did for landscape images. On the other (Dawson and Adams, 1984; Escalona et al., 2002).
hand, pupillary responses which index attentional
engagement and information processing Joint attention and Communication
differentiated the groups when children’s faces
were viewed, but not when animal faces or Relative to developmentally matched controls,
landscapes were viewed. Interestingly, scanning of young children with autism demonstrate serious
landscapes correlated with diagnostic symptoms in impairments in several tasks of joint attention
the autism group; toddlers who fixated more briefly (summarized in Table 4). Additional data (not
showed less repetitive behaviours (Anderson et al., reported in Table 4) assessing these abilities using
2006). standardized or diagnostic instruments provide
similar findings (e.g., Dawson et al., 1998; Dawson
et al., 2004a; Dawson et al., 2004b; Rogers et al.,
360

2003; Wetherby et al., 2004). Further

manifestations of social-communicative impairment Executive function
include difficulties in social referencing, requesting,
and noticing or attending to others’ distress. Executive function is an umbrella term which refers
Furthermore, retrospective studies in the second to mechanisms that guide actions in an endogenous
year of life also report that children later diagnosed goal-driven way, despite conflicting demands from
with autism exhibit decreased initiation of social the immediate environment. These include
interactions and differences in verbal planning, monitoring, working memory, and
communication (reviewed in (Palomo et al., 2006). inhibition. In typical adults, multiple distributed
In fact, because the social-communicative neuroanatomical systems mediate these functions
difficulties encompass several areas, some studies but they tend to be frequently associated with the
have attempted to specify which aspects best prefrontal cortex. So-called “executive
differentiate children with autism from those with dysfunction” in autism is commonly manifest in
developmental delay. These studies suggest that terms of rigidity and preference for sameness,
problems in joint attention are specifically repetitive and restricted interests and behaviors, the
characteristic of young children with autism, latter being among the diagnostic symptoms. There
relative to children with developmental delay or is considerable debate as to whether executive
other comorbid conditions. Moreover, joint function can be regarded as a core deficit in autism.
attention skills have been frequently found to These impairments tend to not be as universal as
correlate with difficulties in other areas reported those found in social-communication. Executive
above, including social orienting (Dawson et al., function deficits tend to be present in various
1998, 2004a), emotion processing (Dawson et al., developmental conditions including Attention
2004b), and imitation (Rogers et al., 2003). Deficit Disorder and Tourette Syndrome, but some
argue that an atypical profile of executive
In sum, the evidence suggests that autism in early dysfunction is specific to autism (Ozonoff and
childhood is associated with serious impairments in Jensen, 1999). Difficulty arises, in part, due to the
several social-communicative skills. On the other wide range of tasks and different levels of
hand, certain difficulties, namely social orienting complexity assessed by these executive function
and joint attention appear to specifically tasks. Performance of individuals with autism on
characterize the disorder, relative to other tasks of planning and mental flexibility is more
conditions. Taken together, despite considerable consistently impaired relative to performance on
research in to social-communicative abilities in tasks of inhibition and self-monitoring (see Hill,
young children with autism and retrospective 2004 for a review).
findings on these abilities, the evidence is limited as Our understanding of the functional
far as distinguishing fine-grained hypotheses development of the prefrontal cortex, the area
regarding the origins of these difficulties. Research frequently associated with executive function, is
with older children and adults has suggested that still limited. However, recent evidence suggests that
the origin of differences in orienting to, and task-dependent developmental changes in this area
scanning of, social stimuli is a precursor to the begin very early in life and some continue well into
more complex social deficits. However, this claim adulthood. There is also evidence that this area is
is difficult to verify given that these problems can important for the development of expertise in other
also be the result of the social impairment rather domains including face processing (Johnson et al.,
than its precursor. By the age of diagnosis and in 2005) and speech perception (Dehaene-Lambertz et
the few years that follow, several aspects of social- al., 2002).
communicative functions are seriously disrupted, Developmental models of autism have taken
making it difficult to differentiate the causes from different views on the emergence of executive
the consequences. Another challenge for this view dysfunction. Some view impairment in the
is the suggestion that these social difficulties are the prefrontal cortex as secondary to those found in
product of a qualitatively different form of social-communicative systems in the medial
impairment in non-social abilities. We now turn to temporal lobe (Dawson et al., 2002a; Griffith et al.,
review some of these findings, presenting the 1999). Others suggest the opposite pattern, where
second general theoretical approach to autism, certain social deficits such as those in theory of
focusing on difficulties in non-social domains. mind are a consequence of executive dysfunction
361

(Russell et al., 1999). Yet another view, is that both more serious impairments in this area become
social orienting deficits as well as executive apparent later in development.
dysfunction originate from early impairments in Interestingly, however, some of these studies
disengagement of visual attention (Bryson et al., found that performance on these tasks correlates
2004). According to the latter view, the infant’s with measures of social-communicative abilities
inability to flexibly switch the locus of attention such as joint attention. More specifically, one study
leads to problems in self-regulation as well as a that differentiated several tasks based on their
decrease in social orienting. We now evaluate these corresponding neuroanatomical systems, found that
suggestions against the evidence available from the tasks tapping into the ventromedial cortex were
early years. more strongly correlated with measures of joint
attention relative to those tapping into the
Studies on executive function dorsolateral system (Dawson et al., 2002a). This
highlights the possibility that some aspects of
Little is known regarding restricted and repetitive executive dysfunction relate to social impairment in
behaviours in the early years which become autism.
characteristic symptoms of older children and In sum, unlike the adult phenotype, executive
adults. Findings from retrospective studies indicate function in autism appears to be not as seriously
that these may not be prevalent prior to diagnosis, affected as social skills. Over development,
especially when children who are later diagnosed difficulties in the area become more apparent. In
with autism are compared to children later relation to evaluating theoretical models of the
diagnosed with mental retardation (reviewed in emergence of autism, although the evidence is
Palomo et al., 2006). Complicating this area of limited, the conclusion that children with autism are
research is the fact that repetitive behaviours and not specifically impaired on these tasks seems to
preference for sameness manifest in qualitatively suggest that executive function may not be the
different ways is young children relative to adults. underlying causal mechanism for social-
Moreover, some of these behaviours are abundant communicative deficits. However, even though
in typical infants and toddlers, e.g., reading the tasks of executive functions do not distinguish
same book, holding onto the same blanket. young children with autism from those with
However, by the age of two, studies using developmental delay, these tasks still correlate with
diagnostic instruments report a higher frequency of social skills, suggesting developmental interactions
repetitive sensorimotor behaviours, some of which between those systems. Because executive
are atypical, in young children with autism (Lord et functions and abilities like theory of mind are to
al., 2006; Richler et al., 2007). some extent similar problem domains, we might
As far as experimental work is concerned, a expect that these social and non-social skills, which
handful of studies have evaluated executive require common computations, are mediated by
functions in young children with autism focusing overlapping neural systems. As a result, the
on a range of skills, including rule-learning, direction of causality is still difficult to establish.
working memory, and response inhibition (see
Table 5 for summary). With the exception of one Sensation, perception, and attention
study (McEvoy et al., 1993), the remaining ones
found that young children with autism do not The two approaches described above focusing on
perform at the level expected based on their social-communicative skills and executive function
chronological age but their performance is capture areas of serious difficulties for individuals
equivalent to children matched on the basis of with autism. However, there are several other
mental age. This implies that unlike some social- characteristics associated with the disorder which
communicative difficulties, performance on these are not clearly related to either of these categories.
tasks is not specific to the autism phenotype but is Some of these differences have been broadly
better predicted by mental age. Although the described in relation to “central coherence”. This
literature evaluating these abilities in older children refers to the ability to perceptually process the
and adults in far more comprehensive, the studies coherent whole of sensory input in order to
with younger children with autism indicate that ascertain meaning (Frith and Happé, 1994).
they are not particularly impaired, suggesting that Individuals with autism are often described to have
a narrow focus of attention and interest as well as
362

acute perception for details. Interestingly, these global processing style, and this has some
differences have been described at multiple levels implications for autism. As typically developing
including sensation, perception, attention, and even children begin to flexibly scan their environment
style of thought and interests. It is unclear, and switch their attention among different objects,
however, whether these areas are related at all, and global forms are processed quickly and efficiently.
whether they share common underlying Infants who exhibit a pattern of prolonged look
mechanisms. Unlike the debate over whether duration, rely more on local elements when
executive dysfunction is a primary or secondary processing visual stimuli (Freeseman et al.,1993;
deficit in autism, central coherence is usually Frick and Colombo, 1996). Furthermore, in adults,
viewed as a complementary explanation, local directed attention and cue-driven attentional
accounting for aspects of performance which are shifting appear to share similar neural substrates in
positive in autism. the left superior parietal cortex (Fink et al., 1997).
Some of the differences in perceptual Hence, it is possible that the narrow focus of
sensitivity in autism relate to processing of low attention in autism is a developmental consequence
spatial frequency relative to high spatial frequency of early difficulties in visual disengagement
visual inputs, associated with the magnocellular and (Landry and Bryson, 2004). Alternatively, it can
the parvocelluar pathways respectively (Behrmann also result from atypical modulation of early visual
et al., 2006; Deruelle et al., 2004). Research in this processing areas by top-down feedback. More
area is limited even in adults, except for a few research is required to test whether these
studies focusing on motion processing where the differences in information processing found in
results appear to be inconsistent (Gepner and autism are related to differences in attentional
Mestre, 2002; Milne et al., 2006; Spencer et al., modulation.
2000). More attention has been paid to aspects of
central coherence related to differences in local vs. Summary
global processing of visuospatial stimuli. Unlike
typically developing individuals, the cognitive Converging lines of evidence indicate that general
processing style in autism appears to be biased deficits as well as specific precursors to some
towards local rather than global or configural symptoms are present early on in autism. By the
information (Happé and Frith, 2006). This time of diagnosis, several co-occurring impairments
sometimes results in superior performance on tasks are clear, and encompass social and non-social
which benefit from these abilities (Behrmann et al., domains. After the onset of these symptoms in the
2006; Jarrold et al., 2005; Jolliffe and Baron- early years, different sets of abilities show varying
Cohen, 1997; Plaisted et al., 1998). In brain trajectories of development. Some abilities, such as
imaging studies of adults with autism, increased face processing, begin as seriously impaired, but
activation of sensory ventral-occipitotemporal areas over time, compensatory strategies and atypical
and decreased activation in prefrontal areas has neural systems may restore behavioural
been described as reflective of these processing performance to within the typical range. Other
differences in local versus global information (Ring deficits, such as executive dysfunction, may not be
et al., 1999). Some studies have suggested an evident at younger ages but become clearer over
association between measures of central coherence development. In all cases, apart from aspects of
and theory of mind (Baron-Cohen and Hammer, behaviour that define the disorder itself, substantial
1997; Jarrold et al., 2000), where individuals with a variability is seen in the resulting phenotype.
more focal processing style are more likely to have In this section we have briefly reviewed some
difficulties with theory of mind tasks. of the most prominent hypotheses concerning core
Because central coherence is a theoretical deficits in autism. Since all of these studies have
model specific to autism and because it describes been conducted with children and adults with
several levels of information processing, it is autism, it is very difficult to unravel the causal
difficult to trace its specific developmental pathways involved and identify the critical primary
trajectory in infants. Furthermore, we could not find deficits. Therefore, in the next section we turn our
any studies examining these abilities in young focus to the much smaller literature on infants and
children with autism. However, in typical infancy toddlers who are at-risk for autism.
research, a link has been made between early
variation in attentional skills in relation to local vs.
363

Table 1. Social vs. non-social orienting and scanning

Study Group n (M:F) Age (sd) Matching Stimuli Measures Key Group differences

Swettenham et ASD 10 20.7 (1.3) CA 5-minute free play session Looking time, duration, attention Relative to both groups, overall ASD looked less and
al. (1998) DD 17 19.9 (1.4) shifting to objects or people switched attention less
TD 16 20.1 (1.3) ASD looked less and for shorter duration at people
and looked more and for longer at objects. They
also shifted attention less between objects and
people
Dawson et al. ASD 20 (19:1) 64.6 MA Social: clapping, name calling; Head turns ASD oriented less to all stimuli but the impairment
(1998) DS 19 (16:3) 65.3 Non-social: musical toy, rattle was worse for social stimuli
TD 20 (19:1) 30.9

Dawson et al. ASD 72 (60:12) 43.5 (4.3) MA Social: Humming, name Head turns ASD oriented less to all stimuli but the impairment
(2004a) DD 34 (18:16) 44.8 (5.3) calling, snapping fingers, was worse for social stimuli
TD 39 (30:9) 27.1 (8.9) tapping thighs;
Non-social: timer bleep, phone
ring, whistle, car horn

Anderson et al. ASD 9 (8:1) 49.6 CA Static unimodal images of Scanning and attentional Relative to both groups, ASD scanned landscapes
(2006) DD 6 (6:0) 46.3 animal faces, children’s faces, engagement (papillary responses) less but did not differ in scanning faces
TD 9 (8:1) 49.8 landscapes ASD showed pupil restriction to internal features of
children’s faces relative to both groups who
showed dilation

ASD, children with autism; DD, children with developmental delay; DS, children with Down Syndrome; TD, typically-developing children; n (M:F), number of participants and ratio of males to females; Age (sd),
mean age in months and standard deviation; CA, group matching based on chronological age; MA, group matching based on mental age
364

Table 2. Face, gaze, and emotion processing

Study Group n (M:F) Age (sd) Matching Stimuli Measures Key Group differences

Chawarska and ASD 14 (10:4) 27.9 (4.9) MA Human and monkey faces Novelty preference (face TD showed enhanced recognition of human but not
Volkmar, (2007) DD 16 (12:4) 23.7 (6.7) recognition) monkey faces, but neither ASD nor DD showed
Experiment 1 TD 10 (8:2) 19.7 (4.7) evidence of face recognition regardless of the
species.
Chawarska and ASD 15 (14:1) 45.9 (6.7) MA Human and monkey faces Novelty preference (face No group differences. Both human and monkey
Volkmar, (2007) DD 10 (8:2) 39.5 (6.7) recognition) faces are recognized by all groups.
Experiment 2 TD 13 (7:6) 38.0 (8.2)

Chawarska et al. ASD 15 (12:3) 28 (6.6) CA Images of faces where eye Reaction time differences in No interaction between group (ASD vs. TD) and
(2003) TD 12 (7:5) 26 (6.5) gaze is congruent or congruent vs. incongruent trials condition (congruent vs. incongruent)
incongruent with the
appearance of a lateral object

Johnson et al. ASD 9 33.4 (5.8) VMA Images of faces where eye Reaction time differences in ASD show no difference between congruent and
(2005) LD 13 34.7 (7.4) gaze is congruent or congruent vs. incongruent trials incongruent conditions whereas TD and LD are
TD 15 36.2 (7.4) incongruent with the faster in the congruent condition
appearance of a lateral object

Dawson et al. ASD 33 (29:4) 43.5 (4.3) CA Familiar and unfamiliar faces Amplitude and latency of ERP Relative to both groups, ASD did not show a
(2002a) DD 17 (11:6) 45.1 (4.9) and objects differential response to faces relative to objects
TD 21 (19:2) 45.5 (5.8) but do show P400 and Nc amplitude differences
to objects.

Gepner et al. ASD 13 (12:1) 69.3 (11.0) MA Still, dynamic, and strobe Accuracy on matching Mode of presentation affects performance in the
(2001) TD 13 (12:1) 40.7 (13.6) emotional or non-emotional expressions emotional condition for TD but not for ASD.
facial expressions

Dawson et al. ASD 29 (26:3) 44.8 (10.0) CA Static photos of an unfamiliar Amplitude and latency of ERP Both groups showed faster response to fear.
(2004b) TD 22 (19:3) 43.7 (7.0) woman posing neutral and fear TD showed larger N300 and NSW amplitudes to
faces fear compared to neutral whereas ASD showed
no difference.
ASD, children with autism or Autism Spectrum Disorder; DD, children with developmental delay; LD, children with language delay; TD, typically-developing children; n (M:F), number of participants and ratio of
males to females; Age (sd), mean age in months and standard deviation; CA, group matching based on chronological age; MA, group matching based on mental age; VMA, group matching based on verbal mental
age; P400, Nc, N300 and NSW, ERP components sensitive to face processing
365

Table 3. Imitation and understanding of contingency

Study Group n (M:F) Age (sd) Matching Stimuli Measures Key Group differences

Rogers et al. (2003) ASD 24 (20:4) 34.2 (3.6) MA Manual, oral-facial, object- Imitative behaviours Relative to both groups, ASD showed overall worse
DD 20 (9:11) 34.1 (6.5) oriented imitations performance
TD 15 (6:9) 21.3 (1.5) ASD were worse on oral-facial and object-oriented
imitation relative to manual imitation.
Imitation problems were unrelated to motor
demands.

Dawson and Adams ASD 15 (12:3) 60.6 (8.2) norm Simultaneous imitation of the Frequency of imitation, social ASD performed well on object permanence, they did
(1984) child’s own actions, a familiar responsiveness poorly on imitation tasks
or a novel action; spontaneous Those who did poorly on imitation were more
imitation; object permanence socially responsive when the examiner imitated
their actions. They were also more likely to
imitate familiar relative to novel actions

Escalona et al. (2002) ASD 20 (12:8) 61.2 Children were randomly Behaviours indicating engagement Both conditions resulted in increased responsiveness
assigned to groups where the with the adult from the child
examiner either imitated their The imitation condition showed additional
action or was contingently improvement such as decrease in motor activity
responsive and a greater increase in touching the adult.

ASD, children with autism; DD, children with developmental delay; TD, typically-developing children; n (M:F), number of participants and ratio of males to females; Age (sd), mean age in months and standard
deviation; MA, group matching based on mental age
366

Table 4. Joint attention and communication

Study Group n (M:F) Age (sd) Matching Stimuli Measures Key Group differences

Charman et al. ASD 8 (7:1) 21.4 (1.8) MA Simulated distress by the Number and quality of target Relative to both groups, ASD did not use gaze
(1998) PDD 13 (11:2) 20.5 (1.7) examiner, spontaneous play, responses declaratively, and showed less target behaviours
DD 8 (3:5) 20.5 (1.0) joint attention, goal detection, including empathetic responses, imitation of
imitation actions on objects, and pretend play.

Stone et al. (1997) ASD 14 32.8 (3.5) CA & MA 16 situations which elicit Number and quality of ASD showed less communicative acts.
DD/LD 14 31.8 (4.4) requesting or commenting communicative acts The quality of communicative acts differed: ASD
behaviour were more likely to directly request
objects/actions than to direct the examiner’s
attention through pointing or eye gaze.

Mundy et al. (1986) ASD 18 (14:4) 53.3 MA ESCS; structured and Initiating and responding to social Relative to both groups, ASD showed less frequency
DD 18 50.2 unstructured play interactions, joint attention, and of higher-level social interaction and requesting,
TD 18 22.2 behaviour regulation but they showed the same frequency of lower-
level ones.
Relative to both groups, ASD showed deficits in
structured but not unstructured play behaviour
Initiating joint attention was the best predictor of
group membership

Dawson et al. ASD 72 (60:12) 43.5 (4.3) MA Simulated distress by the Attention and reactions Relative to both groups, ASD looked less at the
(2004a) DD 34 (18:16) 44.8 (5.3) examiner compared to a examiner and showed less concern in the distress
TD 39 (30:9) 27.1 (8.9) situation where the examiner relative to no the neutral condition.
displays neutral affect

Bacon et al. (1998) HiASD 32 (26:6) 55.8 (14.3) Non-social orienting, two Orienting, response to distress Relative to other groups, LoASD showed deficits
LoASD 51 (47:4) 62.0 (17.9) social situations involving across all measures, whereas HiASD were
LD 42 (29:13) 53.9 (12.5) simulated distress by the moderately impaired in prosocial behaviours
DD 39 (28:11) 55.8 (14.6) examiner. Social referencing differentiated the ASD group
TD 29 (18:11) 55.7 (13.4) from all others

Sigman et al. ASD 30 (27:3) 42.4 (11.2) MA Simulated distress/fear by the Response to distress Relative to both groups, ASD were significantly less
(1992) DD 30 (18:12) 41.6 (10.7) parent or the examiner attentive to adults’ fear or distress and less likely
TD 30 (27:3) 19.83 (8.2) to show social referencing
367

Table 4 (continued)
Study Group n (M:F) Age (sd) Matching Stimuli Measures Key Group differences

Sigman et al. ASD 22 (20:2) 51.3 (11.1) MA Video exposure to an infant Behavioural and cardiac responses ASD primarily differed in their responses to
(2003) DD 22 (14:8) 47.3 (13.1) crying or playing; interaction strangers where they were less interactive and
with mother or stranger; showed less arousal
separation and reunion with
the caregiver

Dawson et al. ASD 16 (13:3) 49.6 (12.1) VMA Free-play; structured Type of activity, gaze, and affective ASD engaged in non-social activities more than TD
(1990) TD 16 (13:3) communicative demands; face behaviour ASD were less likely to combine smiles with
to face interactions sustained eye gaze and to smile in response to the
mother’s smile
Mothers of ASD smiled less and were less likely to
smile in response to their children’s smiles

Kuhl et al. ASD 29 (26:3) 45.3 MA Consonant-vowel syllables Auditory preference and ERP Relative to both groups, ASD did not exhibit the
(2005) TDa 29 (22:7) 27.7 & CA /ba/ and /wa/; infant directed typical mismatch negativity response to change in
TDb 15 (13:2) 48.3 speech and a non-speech speech syllables.
analogue Relative to both groups, ASD preferred to listen to
non-speech analogues of infant directed speech,
whereas TD showed the opposite pattern
Preference for non-speech analogues correlated with
ASD symptoms

ASD, children with autism; HiASD, high functioning children with autism, LoASD; low-functioning children with autism; PDD, children with pervasive developmental disorder; DD, children with developmental
delay; TD, typically-developing children; n (M:F), number of participants and ratio of males to females; Age (sd), mean age in months and standard deviation; CA, group matching based on chronological age; MA,
group matching based on mental age; VMA, group matching based on verbal mental age; ESCS, Early Social Communication Scales (Mundy et al., 1986)
368

Table 5. Executive function

Study Group n (M:F) Age (sd) Matching Stimuli Measures Group differences

Dawson et al. (2002b) ASD 72 (60:12) 43.5 (4.3) MA Ventro-medial tasks: DNMS; Accuracy and number of trials to No group difference
DD 34 (18:16) 44.8 (5.3) dorsolateral prefrontal tasks: A reach a criterion Ventromedial tasks but not dorsolateral prefrontal
TD 39 (28:11) 27.1 (8.9) not B, spatial reversal tasks correlated with joint attention measures

Griffith et al., (1999) ASD 18 (15:3) 50.7 (6.7) MA Working memory and response Various measures including Very few differences where ASD performed better
DD 17 (10:7) 50.6 (9.2) inhibition: accuracy than the DD
A not B, object retrieval, spatial In both groups performance correlated with mental
reversal, boxes age.
A few measures correlated with joint attention in
both groups.

McEvoy et al. (1993) ASD 17 (10:7) 60.6 (12.9) MA A not B, DNMS, spatial reversal, Various measures including There was a ceiling effect for the A not B and
DD 13 (7:6) 50.3 (12.8) alternation accuracy DNMS tasks, and a floor effect for alternation.
TD 16 (10:6) 37.9 (20.0) Relative to both groups, ASD showed more errors
on the spatial reversal task.
In all groups, performance correlated with joint
attention

Dawson et al. (2001) ASD 20 (19:1) 64.6 (15.1) MA Visual paired comparison, Accuracy No group differences
DS 19 (16:3) 65.3 (16.5) DNMS
TD 20 (19:1) 30.9 (14.4)

ASD, children with autism; DD, children with developmental delay; DS, children with Down Syndrome; TD, typically-developing children; n (M:F), number of participants and ratio of males to females; Age (sd),
mean age in months and standard deviation; MA, group matching based on mental age;
A not B, spatial reversal, object retrieval, boxes, alternation, visual paired comparison, and DNMS (delayed non-match to sample) are various tasks assessing executive functions. See cited publications for details.
369

Infancy and autism at the crossroads: where do function (Hughs et al., 1997; 1999; Ozonoff et al.,
we go from here? 1993), and central coherence (Happé et al., 2001).
Hence, understanding the precursors of these
characteristics in infants can help us understand
Despite significant advances in research on the underlying mechanisms, which extend to
infancy and research on developmental disorders, unaffected relatives and potentially to typical
the two areas have rarely been combined. Models development.
of developmental disorders have often focused on
fractionating and describing patterns of strength Emerging finding from sibling studies
and weaknesse in a given disorder, leaving behind
the bulk of theoretical advances in our The emerging literature on infant siblings can be
understanding of the emergent nature of some broadly divided based on whether or not infant
brain functions during development (Johnson et siblings were followed up to the point in time at
al., 2005; Karmiloff-Smith, 1998). A recent field which some of them received a diagnosis. Studies
of investigation focusing on infant siblings of that have followed up the infants from 4-6-months
children with autism has offered promise in to the stage of formal diagnosis around 2-3 years
contributing to our knowledge of the nature of this provide information regarding the predictors or
developmental condition. In this section, we will precursors of autism. In these studies, a few
argue that integrating theoretical and predictor measures have been used longitudinally
methodological advances of functional brain to examine which factors differentiate infant
development in infancy adds substantial value to siblings who develop autism from siblings who do
this research approach, the results of which will be not, or from siblings of unaffected children (Table
fruitful to understanding of both typical as well as 6). When overall IQ was used as a predictor,
atypical development. We begin first with a review infants later diagnosed with autism showed
of the small but growing literature on infant differences on both verbal and non-verbal
siblings. measures as early as 14 months (Landa and
Garrett-Mayer, 2006). Another battery of risk
Why infant siblings? markers, the Autism Observation Scale for Infants
(Zwaigenbaum et al., 2005), focused on precursors
Research on infant siblings of children diagnosed of the abilities used in diagnosing older children,
with autism (infant siblings) is motivated by the including response to name, eye contact, social
emergent nature of autism symptoms and the reciprocity, and imitation. The battery also
premise that neural plasticity may provide the includes items examining basic visual and motor
basis for early intervention, which is likely to lead skills. Here too, infants later diagnosed with
to improved outcomes. Further, it is hoped that autism were clearly differentiated around 12
studying infant siblings will reveal the primary months on a variety of social and non-social
deficits in autism before symptoms are measures. Another approach has focused on
compounded by atypical interactions with the specific predictors such as joint attention (Sullivan
social and physical world, and before et al., 2007) or language (Mitchell et al., 2006).
compensatory strategies and systems cloud the Consistent with the other studies, infants later
basic processing difficulties. diagnosed with autism show differences in these
In fact, the study of siblings offers measures as early as 12 months of age. On the
opportunities to understand not only why autism other hand, none of the studies have, as of yet,
emerges in some cases and not in others, but also found reliable predictors of a later diagnosis within
to potentially explain variations associated with the first year.
the autism phenotype. As mentioned earlier, Other studies have not followed up the infants
genetic relatives of individuals with autism until the age of 2-3 years, but still found
including siblings, who do not have an autism differences in groups of siblings of children with
diagnosis, share certain characteristics with their autism relative to groups of infants who do not
affected relatives. This extended phenotype have affected siblings. These studies cannot
includes aspects of face processing (Dalton et al., provide information about the extent to which
2007; Dawson et al., 2002c), theory of mind performance on certain measures is a predictor of
(Baron-Cohen and Hammer, 1997), executive autism but they do provide information regarding
370

the extended autism phenotype which appears to Towards an understanding of the emergent
be present in infancy. Three studies of 4-6 month- nature of autism
old infant siblings (Cassel et al., 2007; Merin et al.,
2007; Yirmiya et al., 2006) showed subtle It is appropriate to begin the discussion by
differences in the behaviour of infant siblings in a addressing what the current understanding of
task where the mother interacts with the infant, development can contribute towards solving the
then freezes and displays a neutral, expressionless puzzle of the emergent nature of autism: On the
face, and then finally resumes interaction (still face one hand, the human brain undergoes substantial
episode). Two of these studies found that infant and maximal development parentally and
siblings showed differences in their affect during postnatally in the early years, with the clear
the still face episode (Cassel et al., 2007; Yirmiya emergence of precursors of many adult skills.
et al., 2006). The third study did not find the same Autism, however, frequently associated with
differences in affect but did report differences in serious atypical brain and behavioural organization
visual fixation during the episode, where infant in childhood onwards, appears to confer only
siblings looked more to the mother’s mouth subtle symptoms early in life. Why then are the
relative to her eyes (Merin et al., 2007). deficits not so striking in infancy as they are later
Interestingly, differences between infant siblings in childhood and adulthood? There are a number
and control infants also extend to low-level visual of plausible explanations. First, the core deficits
processing. One study found that infant siblings could be specific to social abilities, such as theory
showed enhanced sensitivity to luminance of mind, which are acquired late in development.
contrasts processed in the magnocellular pathway Consistent with certain perspectives, the most
(McCleery et al., in press). Older siblings above sophisticated parts of the social brain network may
14 months show somewhat clearer differences in a be silent early in infancy and only later come
number of social-communicative measures as well online. Autism, according to this explanation, is a
as cognitive and motor measures (detailed in Table case of failure of these later developing systems to
7). Of course follow-up data for these studies are mature. Alternatively, it is the possible that
necessary to examine the extent to which the group theserious symptoms emerge as a result of
differences between infant siblings and control interactions among several more subtle brain and
infants predict any of the characteristics associated behavioural abnormalities early in life. Yet another
with autism later on. possibility is that there may be qualitatively
Taken together, and despite the small number different pathways through which infants come to
of studies to date, some aspects of these data are exhibit the autism phenotype. These different
already revealing. Studies on infant siblings have pathways may be difficult to detect in group
confirmed that the expression of autism in the studies. The last two possibilities are consistent
early years is subtle, at least as far as overt with the notion of probabilistic epigenesis, where
symptoms or delays in the expected developmental development is the result of bi-directional
milestones, including some of the precursors to interactions among brain structure and function
social-communicative difficulties. Infant siblings (c.f., Johnson 2005). Entertaining these
who do not receive a diagnosis in the second or possibilities is essential as these different
third year may share some of the early perspectives would motivate different research
characteristics of siblings who go on to develop questions including what to look for, what to
autism, suggesting that the presence of specific expect, and what theoretical and practical
social or non-social difficulties early on in not conclusions are drawn. We will now discuss these
sufficient for the disorder to emerge. Although alternatives in detail, incorporating some existing
much more research is required before drawing controversies in research on older children and
firm conclusions, these new findings can be adults. We begin with the first possibility, which is
integrated with findings from other approaches to that autism symptoms are subtle in infancy
motivate future directions in this area. because the core deficit in this disorder relates to
later developing skills such as theory of mind.
371
Table 6. Studies that followed groups of infant siblings at various ages in early infancy until the stage when some received an autism diagnosis
Studya Groupb n (M:F) Predictor ages Predictor measures Outcome Outcome measures Key results
ages
Zwaigenbaum et ASD 19 6, 12 Risk markers (AOSI); 24 Diagnosis (ADOS) At 6 months, ASD differed on only one measure of
al. (2005) Sibling 46 IQ (Mullen); temperament (activity level)
language (CDI); visual At 12 months, the groups differed on several
attention behavioural markers from the AOSI and showed
(disengagement); delayed language, prolonged latency to disengage
temperament (IBQ) visual attention, and atypical temperament
Mitchell et al. ASD 15 12, 18 Language (CDI) 24 Diagnosis (ADOS) Both at 12 and 18 months, the ASD group showed
(2006) Sibling 82 delays in understanding and producing words and
TD 49 gestures.
Bryson et al. ASD 9 (5:4) 6, 12, Risk markers (AOSI); 18, 24, 36 Diagnosis Six children showed a serious decrease in IQ between
(2007) 18, 24 IQ (Mullen/ Bayley) (DSM-IV, ADOS, 12 and 24/36 months whereas the remaining 3 were
ADI), IQ within the average range.
(Mullen/Bayley) The earliest signs of autism included social-
communicative problems, atypical sensorimotor and
temperamental characteristics.
Loh et al. (2007) ASD 8 (3:5) 12, 18 Observation of motor 36 Diagnosis ASD showed more frequent and atypical stereotyped
Sibling 9 (3:6) behaviours during semi- (DSM-IV, ADOS, motor movements and postures.
TD 15 (10:5) structured play ADI)
Landa and ASD 24 6, 14 IQ (Mullen) 24 Diagnosis (ADOS), At 6-m the ASD group did not differ on any subscales
Garrett-Mayer LD 11 language (PLS, but at 14 months they differed on all except visual
(2006) TD 52 CDI), IQ (Mullen) reception.
At 24-m the ASD group was further differentiated from
the LD group based on language scores.
The ASD group showed a slower increase in scores
over time.
372
Table 6 (continued)
Studya Groupb n (M:F) Predictor ages Predictor measures Outcome Outcome measures Key results
ages
Sullivan et al. ASD 16 (14:2) 14, 24 Various measures of 30/36 Diagnosis (ADOS), ASD showed subtle differences in joint attention at
(2007) BAP 8 (7:1) response to joint IQ (Mullen) 14m, but the differences were clearer and more stable
Sibling 27(12:15) attention across measures at 24-m
Joint attention performance predicted language
outcomes for the entire group
Gamliel et al. Sibling 39 (13:25) 4, 14 IQ (Bayley) 24, 36, 54 IQ (Kaufman), A subgroup showed difficulties in various cognitive and
(2007); Yirmiya TD 39 (14:25) language (RDLS, linguistic domains (only one child received an
et al. (2007) CELF-P) Autism diagnosis).
At 4 years, the group was within typical levels of
functioning with the exception of a few siblings
showing some difficulties in language.
a
Some studies overlap in their samples. bThe groups refer to the classification of infants at the time of diagnosis; for studies with different n over time, the largest is reported
ASD, children with autism; LD, children with language delay; Sibling, siblings of children with autism who did not themselves receive a diagnosis; BAP, children exhibiting the characteristics
of the broader autism phenotype; TD, typically-developing children with no family history of autism; n (M:F), number of participants and ratio of males to females; Predictor ages, ages in
months around which the predictor measures were collected; Outcome ages, ages in months around which the outcome measures were collected; AOSI, Autism Observation Scale for Infants
(Zwaigenbaum et al., 2005); Mullen, Mullen Scales of Early Learning (Mullen, 1995); Bayley, Bayley Scales of Infant Development (Bayley, 1993); CDI, McArthur Communicative
Development Inventory (Fenson et al., 1993); IBQ, Infant Behavior Questionnaire (Rothbart, 1981); ADOS, Autism Diagnostic Observation Schedule (Lord et al., 2000), ADI, Autism
Diagnostic Interview (Lord et al., 1994); DSM-VI, clinical diagnosis based on the DSM-IV (APA, 1994); PLS, Preschool Language Scale (Zimmerman et al., 2002); Kaufman, Kaufman
Assessment Battery for Children (Kaufman and Kaufman, 1983); RDLS, Reynell Developmental Language Scales (Reynell, J. K., & Grubber, C. P., 1990), CELF-P, Clinical Evaluation of
Language Fundamentals-Preschool (Wiig et al., 1992)
373
Table 7. Studies of the extended phenotype in autism within the first two years
Study Group n (M:F) Age (sd) Measures Key results
McCleery et al. Sib-ASD 13 (5:8) 6.07 (0.14) Contrast sensitivity related to magnocellular Sib-ASD showed greater contrast sensitivity related to the
(in press) TD 26 (5:8) 6.07 (0.12) (luminance) and parvocellular (chromatic) magnocellular pathway but did not differ in their contrast
visual pathways sensitivity related to the parvocallular pathway
Cassel et al. Sib-ASD 11 6.1 (0.3) Emotional reaction in the still face episode Sib-ASD smiled less during the overall episode
(2007) TD 19 for overall sample
Merin et al. Sib-ASD 31 (15:16) 6.06 (0.3) Emotional reaction and behavioural eye More infants in the Sibs-ASD group gazed predominantly at the
(2007) TD 24 (16:8) 5.9 (0.3) tracking before, during, and after the still mother’s mouth relative to her eyes
face episode
Iverson and Sib-ASD 21 (6:15) Around 5, 14 Developmental milestones; production of A significant percentage of Sib-ASD showed delays including
Wozniak (2007) TD 18 (8:10) rhythmic limb movements postural babbling, pointing, first words, and sitting.
variability Sib-ASD showed postural instability
Yirmiya et al. Sib-ASD 30 (8:13) 20.2 (3.2) Synchrony of interaction, emotional reaction Sib-ASD were less synchronous in their interactions
(2006) TD 31 (13:18) 19.6 (2.8) in the still face, response to name Sib-ASD were less upset during the still face
More infants in the Sibs-ASD group responded to their name
Toth et al. (2007) Sib-ASD 42 (21:21) 20.3 (2.2) Immediate and deferred imitation, functional No group differences in imitation or play
TD 20 (13:7) 22.4 (2.8) and symbolic play; IQ (Mullen); social Sib-ASD had lower scores on the receptive language subscale of
communication (CSBS), autism screening the Mullen and a lower composite score on the CSBS.
(ADOS) No group differences on the ADOS except on items of pointing and
responsive social smile.
Presmanes et al. Sib-ASD 46 (26:20) 15.4 (3.0) Different combinations of joint attention Sib-ASD showed less responding to joint attention for moderately
(2007) TD 35 (24:11) 15.5 (2.9) cues, e.g., gaze, gaze and head movement; redundant cues but not for highly redundant ones
IQ (Mullen); autism characteristics (STAT) No differences on the Mullen except on the visual reception
subscale
No differences on the STAT
Sib-ASD, siblings of children with autism; TD, typically-developing children with no family history of autism; n (M:F), number of participants and ratio of males to females; Age (sd), mean age
and standard deviation around which the measures were taken, Mullen, Mullen Scales of Early Learning (Mullen, 1995); CSBS, Communication and Symbolic behavior Scale (Wetherby et
al., 2002), ADOS, Autism Diagnostic Observation Schedule (Lord et al., 2000), STAT, Screening Tool for Autism in Two-year-olds (Stone et al., 2004)
374

Is there a unitary core deficit in autism? impairment on other symptoms is less marked.
This does not imply that the social-communicative
In the previous section, we discussed several deficits are the sole characteristics of autism.
approaches to autism in older children and adults Associated deficits are clearly present in the
that seek a core characterization of autism, where majority of cases. These characteristics, however,
other deficits are secondary or peripheral to the will be present in variable degrees and may be less
primary one. These differing approaches have striking in some individuals relative to social-
given rise to several debates including whether communicative impairment. From a scientific point
deficit in theory of mind or alternatively in of view, even if certain characteristics of autism
executive function should be viewed as the core are not universal, while others are (by virtue of
one. This debate is similar to more general one in having defined the disorder as such), the associated
the area of developmental disorders concerning the impairments still require an explanation, which a
utility of fractionating cognitive abilities as a single core cognitive deficit falls short of
function of the degree of impairment found in each providing.
domain (Karmiloff-Smith, 1998). Beyond these issues of definition of the
Several caveats render this frequently disorder, several lines of evidence appear to
recurring debate in developmental disorders less challenge the notion of a single core deficit in
interesting. One of these caveats is related to the autism. Research on other developmental disorders
behavioural characterization of the autism has challenged the notion of selective deficits both
phenotype. Research studies have increasingly at the behavioural and genetic levels. The emerging
benefited from relying on more standardized picture has favoured the view that multiple and
clinical definitions (American Psychiatric sometimes subtle deficits across domains lead to
Association, 1994; World Health Organization, atypical patterns of development with some
1994) and diagnostic tools (Lord et al., 1994,2000) domains being more affected than others
for autism. The conceptualization of the disorder (Karmiloff-Smith, 1998). There is also evidence
has become broader, encompassing “non-classic” for strong overlap between autism and other
cases where a diagnosis may or may not be developmental disorders such as Specific
warranted given the individual’s overall cognitive Language Impairment, implying that common
and functional ability. In parallel, the diagnostic aetiology may lead to qualitatively different
tools have become more focused on symptoms that phenotypic conditions depending on the presence
are likely to be common to most individuals. These and severity of other risk factors (Bishop, 2003).
instruments have attempted to reduce variations in Furthermore, population studies indicate that the
clinical judgement, which may lead to highly characteristics of autism tend to cluster differently
heterogeneous groups in research studies, by in the general population making it unlikely that a
ensuring that the defining characteristics of autism, there is a single explanation for the triad (Happé et
namely social-communicative symptoms are al., 2006).
present in all cases, albeit varying enormously in Turning to the proposed core deficits from the
severity and in comorbid symptoms. In the interest developmental models of autism presented earlier,
of more homogeneous samples, the definition of there is reasonable evidence that problems in social
autism frequently excludes symptoms which orienting are present early on. There is also
overlap with other disorders including the presence preliminarily evidence that impairments in
of repetitive and stereotyped behaviours, or those attentional disengagement are expressed between 6
related to overall cognitive or language delay. As a and 12 months of age. Hence, these two
result, autism like several other developmental mechanisms are potentially viable candidates for a
disorders is diagnosed primarily on the basis of core deficit. However, neither mechanism can be
behavioural impairment in a specific area. described as directly causal, since (a) both co-occur
While this narrowing of the criteria is towards the end of the first year of life, and (b)
motivated by reasonable objectives in terms of problems with disengagement occur in various
specificity of the clinical definition and developmental disorders including ones with
homogeneity of the samples, from a theoretical markedly different social profiles from the one
point of view, there is doubt on whether it is found in autism.
sensible to look for core deficits in other domains, In sum, the search for core deficits at the
after having a priori excluded individuals whose genetic and cognitive levels has, as yet, proved
375

unsuccessful. This is clearly illustrated by our subtle impairments in several systems where the
review of finding in the previous section, where no typical developmental constraints are altered. A
single account appears to capture all manifestations more fruitful approach is to focus on interactions
of the autism profile. From a developmental between systems within individuals. The
perspective, even if a cognitive core deficit existed, development of autism in infancy is best
it would be very hard to detect by later childhood understood in terms of multiple risk factors, where
due to the compounding of, and compensatory the presence and the severity of each risk factor as
strategies applied to, different secondary well as their interactions would explain the
consequences of the original cause. However, it resulting phenotype.
remains possible that there will be a core deficit in Considering some concrete examples will
the functional development of the brain, and that clarify this point. An early deficit in orienting to
this could explain the full triad of impairments seen socially relevant stimuli may be necessary but not
in autism. This core deficit could have immediate sufficient for autism to emerge. The deficit in
multiple causes at the perceptual/cognitive level, as social orienting would result in decreased input
well as a range of developmental consequences. from socially relevant stimuli. This deficit,
The emergence of autism as a result of complex however, would be compounded and amplified by
interactions the presence of other difficulties, such as those
found in attentional disengagement. A problem
Brain development is a dynamic and self- with flexibly switching attention between different
organizing system (c.f., Johnson 2005). stimuli would result in ‘locking’ onto certain
Consequently, the division between primary and irrelevant aspects of the diminished input. The
secondary functional abnormalities is misleading. infant in this case, would not only receive
For instance neuroanatomical abnormalities in an decreased input from social stimuli, but attentional
earlier developing structure will have constraints would impose qualitatively different
consequences for later developing structures. forms of input, namely focal and irrelevant ones.
Abnormalities in these later developing systems This would suggest that infants who exhibit a
may then exert further atypicality on the earlier combination of disengagement difficulties with
developing systems if they provide feedback decreased social orienting would be at higher risk
projections to them. Hence, although the functions for autism than infants who exhibit one of these
mediated by later developing abnormal structures difficulties in isolation. Several concrete
can be described as being caused by the primary predictions can be further examined in terms of
abnormalities, these in turn causally modify the more specific behavioural outcomes. For instance,
earlier structures and their functions through we might expect that the severity of early deficit in
feedback projections. For example, atypical social orienting and/or in disengagement to
modulation of early visual processing areas by top- correlate with the severity of impairment in face
down feedback would have far-reaching processing.
implications in modifying other perceptual and Another attractive hypothesis is the extent to
cognitive processes which rely on these top-down which social orienting deficits are themselves
systems. These consideration stress the importance specific to social stimuli. An alternative view is
of studying autism at the earliest ages that that infants who go on to develop autism differ in
symptoms appear, since here we have the most their spontaneous preference for the degree of
chance to study brain development while the contingency in events, whether social or non-
atypical trajectory is only slightly deviant from the social. Consistent with findings from young
typical course, and before compensatory and children with autism (reviewed in Table 3), those
confounding effects have influenced further infants would show an atypical preference for
development. perfect rather than high, but imperfect, contingency
In short, no models that involve a single core and as a result, they are less attracted to social
deficit in autism have yet explained the range of stimuli because the latter tend to be imperfect.
data available, and we should not necessarily Infancy research with at-risk infants prior to the
assume that there will be a unitary developmental onset of symptoms would help to disentangle these
cause of this kind. Instead, it is possible that the competing alternatives.
functions, which later have been described as core Instead of seeking to affirm that a core deficit
cognitive deficits, arise from widespread albeit is present or not prior to the onset of the symptoms,
376

research focusing on autism in infancy needs to design fails to capture qualitatively different routes
examine interactions among several systems within that lead to the autism phenotype.
individuals. Such preliminary integrative accounts Several lines of evidence suggest that the early
of how several risk factors could explain developmental course of autism is in fact
subsequent competence are essential in explaining heterogeneous. A handful of “classic” cases of
the resulting adult profiles. Elements of such autism have been described in infants around the
preliminary developmental accounts can be first year of life, where the typical manifestations
examined rigorously and verified independently or of autism symptoms are clear (Dawson et al., 2000;
in combination, but must be integrated Klin et al., 2004). On the other hand, in other
theoretically. This would shift our understanding of infants early signs may be very difficult to detect.
autism towards a coherent picture of the dynamics Different developmental profiles of children with
through which the adult phenotype is reached. autism have also been hypothesized (Palomo et al.,
With these considerations in mind, we now turn to 2006; Sigman et al., 2004; Werner and Dawson,
a related issue of explaining the emergent nature of 2005; Werner et al., 2005). For some, deficits are
autism in view of individual variability in early apparent early on and arrested development
developmental trajectories. continues. In contrast, other children achieve some
developmental milestones on the typical
Heterogeneity: merely methodological nuisance? developmental schedule but then exhibit loss of
some abilities and a developmental regression.
We have described earlier that the most common Interestingly, heterogeneity in expression of early
and unifying feature of autism is the presence of symptoms appears to map out on less
social-communicative impairment. Beyond this, heterogeneous outcomes. An earlier age of
cases vary enormously in severity and in comorbid recognition correlates with more classic phenotypic
symptoms. For instance, two cases meeting an autism outcomes, whereas later recognition of
autism classification cut-off may include a child early symptoms is associated with more variable
who is non-verbal as well as a child whose manifestations, even if the severity of these
language is advanced, albeit peculiar. In fact, the symptoms does not differ (Chawarska et al., 2007).
conception of autism as a spectrum suggests that Hence, this variability is likely to be the result
this diagnostic label includes individuals differing of dynamic and probabilistic interactions over
greatly in functional impairment. This striking development, and the systematic study of these
heterogeneity has been considered a daunting variations can offer important clues towards
issue, not only for autism research but also more understanding the emergent nature of autism. This
generally in developmental disorders. On understanding will depend on developing
methodological grounds such individual appropriate theoretical and methodological models
differences in the resulting profiles are frequently capable of handling individual differences in
discounted as random noise, where heterogeneity is phenotypic expression. What are the implications
regarded as an explanation in itself. However, of this for the study of infants at risk for autism?
phenotypic expression in developmental disorders Infancy research has traditionally relied on group
is considerably more variable than that seen in methodology. In fact, some classic paradigms such
typical development, suggesting that not all of the as habituation yield binary looking time measures
variability is random. In other words, atypical for which there is little confidence in individual
development is inherently more variable than infants’ results. In the study of at-risk siblings, it is
typical development, reflecting different ways of essential that existing techniques are adapted to
brain and behavioural reorganization. In the boost confidence in individual results, which
previous section we argued that autism symptoms would, in turn, be fruitful for infancy research in
are subtle early on because of the emergent nature general. Furthermore, as we argued earlier, cross-
of the disorder, where several risk factors interact task correlations within individual infants offer the
over time, resulting in the phenotype found in older opportunity to examine individual differences in
children and adults. Another possibility, which is cognitive profiles. Another possibility is to
not mutually exclusive, is that early symptoms are examine variations in performance within tasks
difficult to detect because they are variable across across different infants. The latter requires
different individuals. It is possible that a group individual and continuous measures, which would
377

clarify the extent of consistency or variability of American Psychiatry Association (1994).

each measure across individuals. Diagnostic and statistical manual (4th ed).
Washington, DC: APA Press.
Anderson, C. J., Colombo, J. and Jill Shaddy, D.
Concluding remarks (2006) Visual scanning and pupillary responses
in young children with Autism Spectrum
Recent advances in autism and in infancy research Disorder. J Clin Exp Neuropsychol, 28, 1238-
and the recent integration of these directions hold a 1256.
lot of promise for solving some of the most Arndt, T. L., Stodgell, C. J. and Rodier, P. M.
difficult puzzles in autism. To understand the (2005) The teratology of autism. Int J Dev
underlying causes of this condition and the process Neurosci, 23, 189-199.
through which it emerges, any convincing model Autism Genome Project (2007) Mapping autism
must encompass not only the symptoms which risk loci using genetic linkage and
define the disorder but also other characteristics of chromosomal rearrangements. Nat Genet., 39,
the profile. 319-328.
This is best achieved through making use of Bailey, A., Le Couteur, A., Gottesman, I., Bolton,
existing developmental models to motivate P., Simonoff, E., Yuzda, E. and Rutter, M.
theoretical accounts of how and why risk factors, (1995) Autism as a strongly genetic disorder:
alone or in combination, exert an impact on the evidence from a British twin study. Psychol
resulting phenotype. It is equally important to Med, 25, 63-77.
understand protective factors, as this is likely to Baird, G, Simonoff, E, Pickles, A, Chandler, S,
offer important clinical implications, if it turns out Loucas, T, Meldrum, D, Charman, T. (2006)
that simple environmental manipulations might Prevalence of disorders of the autism spectrum
have important consequences in the window of in a population cohort of children in South
maximal plasticity. A theoretical account of this Thames: the Special Needs and Autism Project
sort requires the development of continuous and (SNAP). Lancet, 15, 210-215.
individually valid measurements, both of predictors Baron-Cohen, S., Allen, J. and Gillberg, C. (1992)
and outcomes, capable of capturing individual Can autism be detected at 18 months? The
differences. This approach is likely to be needle, the haystack, and the CHAT. Br J
productive not only for understanding autism, but Psychiatry, 161, 839-843.
could also offer insights into the mechanisms Baron-Cohen, S. and Hammer, J. (1997) Parents of
which drive and constrain typical development. children with Asperger Syndrome: what is the
cognitive phenotype? Journal of Cognitive
Neuroscience 9, 548-554.
Acknowledgment Baron-Cohen, S., Leslie, A. M. and Frith, U.
(1985) Does the autistic child have a "theory of
We would like to thank Leah Kaminsky for her mind"? Cognition, 21, 37-46.
help with reference materials, and Teodora Gliga, Baron-Cohen, S., Ring, H. A., Wheelwright, S.,
Atsuhi Senju, and Dick Aslin for helpful comments Bullmore, E. T., Brammer, M. J., Simmons, A.
on an earlier version of this manuscript. This work and Williams, S. C. (1999) Social intelligence
was supported by the UK Medical Research in the normal and autistic brain: an fMRI study.
Council Programme Grant G9715587. Eur J Neurosci, 11, 1891-1898.
Bayley, N. (1993). Bayley Scales of Infant
References Development. 2nd ed. The Psychological
Corporation.
Adolphs, R. (2003) Cognitive neuroscience of Behrmann, M., Thomas, C. and Humphreys, K.
human social behaviour. Nat Rev Neurosci, 4, (2006) Seeing it differently: visual processing
165-178. in autism. Trends Cogn Sci, 10, 258-264.
Allison, T., Puce, A. and Mccarthy, G. (2000) Bishop, D. V. M. (2003) Autism and specific
Social perception from visual cues: role of the language impairment: categorical distinction or
STS region. Trends Cogn Sci, 4, 267-278. continuum? . IN Bock, G. and Goode, J. (Eds.)
Autism: neural basis and treatment possibilities.
Chichester, John Wiley.
378

Bolton, P., Macdonald, H., Pickles, A., Rios, P., correlates of early symptoms of autism. Child
Goode, S., Crowson, M., Bailey, A. and Rutter, Dev, 69, 1276-1285.
M. (1994) A case-control family history study Dawson, G., Munson, J., Estes, A., Osterling, J.,
of autism. J Child Psychol Psychiatry, 35, 877- Mcpartland, J., Toth, K., Carver, L. and Abbott,
900. R. (2002a) Neurocognitive function and joint
Bryson, S., Landry, R., Czapinski, P., Mcconnell, attention ability in young children with autism
B., Rombough, V. and Wainwright, A. (2004) spectrum disorder versus developmental delay.
Autistic spectrum disorders: causal mechanisms Child Dev, 73, 345-358.
and recent findings in attention and emotion. Dawson, G., Osterling, J., Meltzoff, A. N. and
International Journal of Special Education, 19, Kuhl, P. K. (2000) Case study of the
14-22. development of an infant with autism from
Bryson, S.E., Zwaigenbaum, L., Brian, J., Roberts, birth to 2 years of age J App Devl Psychol, 21,
W., Szatmari, P., Rombough, V. and 299-313.
McDermott, C. (2007) A prospective case Dawson, G., Toth, K., Abbott, R., Osterling, J.,
series of high-risk infants who developed Munson, J., Estes, A. and Liaw, J. (2004a)
autism. J Autism Dev Disord, 37,12-24. Early social attention impairments in autism:
Cassel, T. D., Messinger, D. S., Ibanez, L. V., social orienting, joint attention, and attention to
Haltigan, J. D., Acosta, S. I. and Buchman, A. distress. Dev Psychol, 40, 271-83.
C. (2007) Early Social and Emotional Dawson, G., Webb, S., Schellenberg, G. D., Dager,
Communication in the Infant Siblings of S., Friedman, S., Aylward, E. and Richards, T.
Children with Autism Spectrum Disorders: An (2002b) Defining the broader phenotype of
Examination of the Broad Phenotype. J Autism autism: genetic, brain, and behavioral
Dev Disord, 37, 122-132. perspectives. Dev Psychopathol, 14, 581-611.
Chawarska, K., Klin, A. and Volkmar, F. (2003) Dawson, G., Webb, S. J., Carver, L., Panagiotides,
Automatic attention cueing through eye H. and Mcpartland, J. (2004b) Young children
movement in 2-year-old children with autism. with autism show atypical brain responses to
Child Dev, 74, 1108-1122. fearful versus neutral facial expressions of
Chawarska, K., Paul, R., Klin, A., Hannigen, S., emotion. Dev Sci, 7, 340-59.
Dichtel, L. E. and Volkmar, F. (2007) Parental Dawson, G., Webb, S. J., Wijsman, E.,
Recognition of Developmental Problems in Schellenberg, G., Estes, A., Munson, J. and
Toddlers with Autism Spectrum Disorders. J Faja, S. (2005) Neurocognitive and
Autism Dev Disord, 37, 62-72. electrophysiological evidence of altered face
Chawarska, K. and Volkmar, F. (2007) processing in parents of children with autism:
Impairments in monkey and human face implications for a model of abnormal
recognition in 2-year-old toddlers with Autism development of social brain circuitry in autism.
Spectrum Disorder and Developmental Delay. Dev Psychopathol, 17, 679-97.
Dev Sci, 10, 266–279. De Haan, M., Johnson, M. H. and Halit, H. (2003)
Dalton, K.M., Nacewicz, B.M., Alexander, A.L. Development of face-sensitive event-related
and Davidson, R.J. (2007) Gaze-fixation, brain potentials during infancy: a review. Int J
activation, and amygdala volume in unaffected Psychophysiol, 51, 45-58.
siblings of individuals with autism. Biol Dehaene-Lambertz, G., Dehaene, S. and Hertz-
Psychiatry, 61, 512-512. Pannier, L. (2002) Functional neuroimaging of
Dalton, K. M., Nacewicz, B. M., Johnstone, T., speech perception in infants. Science, 298,
Schaefer, H. S., Gernsbacher, M. A., 2013-2015.
Goldsmith, H. H., Alexander, A. L. and Deruelle, C., Rondan, C., Gepner, B. and Tardif, C.
Davidson, R. J. (2005) Gaze fixation and the (2004) Spatial frequency and face processing in
neural circuitry of face processing in autism. children with autism and Asperger syndrome. J
Nat Neurosci, 8, 519-526. Autism Dev Disord, 34, 199-210.
Dawson, G. and Adams, A. (1984) Imitation and Escalona, A., Field, T., Nadel, J. and Lundy, B.
social responsiveness in autistic children. J (2002) Brief report: imitation effects on
Abnorm Child Psychol, 12, 209-25. children with autism. J Autism Dev Disord, 32,
Dawson, G., Meltzoff, A. N., Osterling, J. and 141-144.
Rinaldi, J. (1998) Neuropsychological
379

Fenson, L., Dale, P.S., Reznick, S., Thal, D., Bates, Gergely, G. and Watson, J. S. (1999) Early Socio-
S., Hartung, J.P., Pethick, S., Reilly, J.S., Emotional Development: Contingency
(1993). MacArthur Communication Perception and the Social-Biofeedback. IN
Development Inventories. San Antonio, TX: Gergely, G. and Watson, J. (Eds.) Early social
Psychological Corporation. cognition: Understanding others in the first
Filipek, P. A., Accardo, P. J., Ashwal, S., Baranek, months of life Mahwah, NJ, Lawrence
G. T., Cook, E. H., Jr., Dawson, G., Gordon, Erlbaum.
B., Gravel, J. S., Johnson, C. P., Kallen, R. J., Griffith, E. M., Pennington, B. F., Wehner, E. A.
Levy, S. E., Minshew, N. J., Ozonoff, S., and Rogers, S. J. (1999) Executive functions in
Prizant, B. M., Rapin, I., Rogers, S. J., Stone, young children with autism. Child Dev, 70,
W. L., Teplin, S. W., Tuchman, R. F. and 817-832.
Volkmar, F. R. (2000) Practice parameter: Happé, F., Briskman, J. and Frith, U. (2001).
screening and diagnosis of autism: report of the Exploring the cognitive phenotype of autism:
Quality Standards Subcommittee of the weak "central coherence" in parents and
American Academy of Neurology and the siblings of children with autism: I.
Child Neurology Society. Neurology, 55, 468- Experimental tests. J Child Psychol Psychiatry,
479. 42,299-307.
Filipek, P. A., Accardo, P. J., Baranek, G. T., Happé, F., Ehlers, S., Fletcher, P., Frith, U.,
Cook, E. H., Jr., Dawson, G., Gordon, B., Johansson, M., Gillberg, C., Dolan, R.,
Gravel, J. S., Johnson, C. P., Kallen, R. J., Frackowiak, R. and Frith, C. (1996) 'Theory of
Levy, S. E., Minshew, N. J., Ozonoff, S., mind' in the brain. Evidence from a PET scan
Prizant, B. M., Rapin, I., Rogers, S. J., Stone, study of Asperger syndrome. Neuroreport, 8,
W. L., Teplin, S., Tuchman, R. F. and Volkmar, 197-201.
F. R. (1999) The screening and diagnosis of Happé, F. and Frith, U. (2006) The weak coherence
autistic spectrum disorders. J Autism Dev account: detail-focused cognitive style in
Disord, 29, 439-484. autism spectrum disorders. J Autism Dev
Fink, G.R., Halligan, P.W., Marshall, J.C., Frith, Disord, 36, 5-25.
C.D., Frackowiak, R.S. and Dolan, R.J. (1997) Happé, F., Ronald, A. and Plomin, R. (2006) Time
Neural mechanisms involved in the processing to give up on a single explanation for autism.
of global and local aspects of hierarchically Nat Neurosci, 9, 1218-1220.
organized visual stimuli. Brain,10, 1779-1791. Hill, E. L. (2004) Executive dysfunction in autism.
Freeseman, L.J., Colombo, J. and Coldren, J.T. Trends Cogn Sci, 8, 26-32.
(1993) Individual differences in infant visual Jarrold, C., Butler, D. W., Cottington, E. M. and
attention: four-month-olds' discrimination and Jimenez, F. (2000) Linking theory of mind and
generalization of global and local stimulus central coherence bias in autism and in the
properties. Child Dev., 64,1191-1203. general population. Dev Psychol, 36, 126-38.
Frick, J.E. and Colombo, J. (1996) Individual Hughes, C., Plumet, M.H. and Leboyer, M.. (1999)
differences in infant visual attention: Towards a cognitive phenotype for autism:
recognition of degraded visual forms by four- increased prevalence of executive dysfunction
month-olds. Child Dev, 67,188-204. and superior spatial span amongst siblings of
Frith, U. and Happé, F. (1994) Autism: beyond children with autism. J Child Psychol
"theory of mind". Cognition, 50, 115-32. Psychiatry, 40,705-718.
Gamliel, I., Yirmiya, N. and Sigman, M. (2007) Hughes, C., Leboyer, M. and Bouvard, M. (1997)
The development of young siblings of children Executive function in parents of children with
with autism from 4 to 54 months. J Autism Dev autism. Psychol Med., 27, 209-220.
Disord., 37,171-183. Iverson, J.M. and Wozniak, R.H. (2007).Variation
Gauthier, I. and Nelson, C. A. (2001) The in vocal-motor development in infant siblings
development of face expertise. Curr Opin of children with autism. J Autism Dev Disord,
Neurobiol, 11, 219-224. 37, 158-170.
Gepner, B. and Mestre, D. (2002) Postural Jarrold, C., Gilchrist, I. D. and Bender, A. (2005)
reactivity to fast visual motion differentiates Embedded figures detection in autism and
autistic from children with Asperger syndrome. typical development: preliminary evidence of a
J Autism Dev Disord, 32, 231–238.
380

double dissociation in relationships with visual behaviors associated with autism in high-risk
search. Dev Sci, 8, 344-351. infants: a pilot videotape analysis of a sibling
Jemel, B., Mottron, L. and Dawson, M. (2006) sample. J Autism Dev Disord, 37, 25-36.
Impaired face processing in autism: fact or Lord, C. (1995) Follow-up of two-year-olds
artifact? J Autism Dev Disord, 36, 91-106. referred for possible autism. J Child Psychol
Johnson, M. H. (2001) Functional brain Psychiatry, 36, 1365-1382.
development in humans. Nat Rev Neurosci, 2, Lord, C., Risi, S., Dilavore, P. S., Shulman, C.,
475-483. Thurm, A. and Pickles, A. (2006) Autism from
Johnson, M. H. (2005) Developmental Cognitive 2 to 9 years of age. Arch Gen Psychiatry, 63,
Neuroscience, Oxford, Blackwell. 694-701.
Johnson, M. H., Griffin, R., Csibra, G., Halit, H., Lord, C., Risi, S., Lambrecht, L., Cook, E.H., Jr.,
Farroni, T., De Haan, M., Tucker, L. A., Baron- Leventhal, B.L., DiLavore, P.C., Pickles, A.
Cohen, S. and Richards, J. (2005) The and Rutter, M. (2000) The Autism Diagnostic
emergence of the social brain network: Observation Schedule-Generic: A standard
evidence from typical and atypical measure of social and communicative deficits
development. Dev Psychopathol, 17, 599-619. associated with the spectrum of autism. J
Johnson, M. H. and Morton, J. (1991) Biology and Autism Dev Disord, 30, 205-223.
Cognitive Development: The Case of Face Lord, C., Rutter, M., and Le Couteur, A. (1994).
Recognition, Oxford, Blackwell. Autism diagnostic interview-revised: A revised
Jolliffe, T. and Baron-Cohen, S. (1997) Are people version of a diagnostic interview for caregivers
with autism and Asperger syndrome faster than of individuals with possible pervasive
normal on the Embedded Figures Test? J Child developmental disorders. J Autism Dev Disord,
Psychol Psychiatry, 38, 527-534. 24, 659-685.
Karmiloff-Smith, A. (1998) Development itself is Maestro, S., Muratori, F., Cavallaro, M. C., Pecini,
the key to understanding developmental C., Cesari, A., Paziente, A., Stern, D., Golse, B.
disorders. Trends Cog Sci, 2, 389-398. and Palacio-Espasa, F. (2005) How young
Kaufman, A. S. and Kaufman, N. L. (1983). children treat objects and people: an empirical
Kaufman assessment battery for children (K- study of the first year of life in autism. Child
ABC). Circle Pines, MN: American Guidance Psychiatry Hum Dev, 35, 383-396.
Service. Maestro, S., Muratori, F., Cavallaro, M. C., Pei, F.,
Klin, A., Chawarska, K., Paul, R., Rubin, E., Stern, D., Golse, B. and Palacio-Espasa, F.
Morgan, T., Wiesner, L. and Volkmar, F. (2002) Attentional skills during the first 6
(2004) Autism in a 15-month-old child. Am J months of age in autism spectrum disorder. J
Psychiatry, 161, 1981-8. Am Acad Child Adolesc Psychiatry, 41, 1239-
Klin, A., Jones, W., Schultz, R. and Volkmar, F. 1245.
(2003) The enactive mind, or from actions to Mars, A. E., Mauk, J. E. and Dowrick, P. W.
cognition: lessons from autism. Philos Trans R (1998) Symptoms of pervasive developmental
Soc Lond B Biol Sci, 358, 345-360. disorders as observed in prediagnostic home
Landa, R. and Garrett-Mayer, E. (2006) videos of infants and toddlers. J Pediatr, 132,
Development in infants with autism spectrum 500-504.
disorders: a prospective study. J Child Psychol McCleery, J.P., Allman, E., Carver, L.J. and
Psychiatry, 47, 629-638. Dobkins, K.R. (in press) Abnormal
Landry, R. and Bryson, S. E. (2004) Impaired Magnocellular Pathway Visual Processing in
disengagement of attention in young children Infants at Risk for Autism. Biol Psychiatry.
with autism. J Child Psychol Psychiatry, 45, Mcevoy, R. E., Rogers, S. J. and Pennington, B. F.
1115-1122. (1993) Executive function and social
Leslie, A. M. and Thaiss, L. (1992) Domain communication deficits in young autistic
specificity in conceptual development: children. J Child Psychol Psychiatry, 34, 563-
neuropsychological evidence from autism. 78.
Cognition, 43, 225-251. Merin, N., Young, G. S., Ozonoff, S. and Rogers,
Loh, A., Soman, T., Brian, J., Bryson, S.E., S. J. (2007) Visual Fixation Patterns during
Roberts, W., Szatmari, P., Smith, I.M. and Reciprocal Social Interaction Distinguish a
Zwaigenbaum, L. (2007). Stereotyped motor Subgroup of 6-Month-Old Infants At-Risk for
381

Autism from Comparison Infants. J Autism highly similar stimuli by adults with autism
Dev Disord, 37,108-121. during a perceptual learning task. J Child
Miller, M. T., Stromland, K., Ventura, L., Psychol Psychiatry, 39, 765-775.
Johansson, M., Bandim, J. M. and Gillberg, C. Presmanes, A.G., Walden, T.A., Stone, W.L. and
(2005) Autism associated with conditions Yoder, P.J. (2007) Effects of different
characterized by developmental errors in early attentional cues on responding to joint attention
embryogenesis: a mini review. Int J Dev in younger siblings of children with autism
Neurosci, 23, 201-19. spectrum disorders. J Autism Dev Disord.,
Milne, E., White, S., Campbell, R., Swettenham, J., 37,133-144.
Hansen, P. and Ramus, F. (2006) Motion and Reynell, J. K. and Grubber, C. P. (1990). Reynell
form coherence detection in autistic spectrum developmental language scale. Los Angeles:
disorder: Relationship to motor control and 2:4 Western Psychological Association.
digit ratio. J Autism Dev Disord, 36, 225-237. Richler, J., Bishop, S.L., Kleinke, J.R. and Lord, C.
Mitchell, S., Brian, J., Zwaigenbaum, L., Roberts, (2007) Restricted and repetitive behaviors in
W., Szatmari, P., Smith, I. and Bryson, S. young children with autism spectrum disorders.
(2006) Early language and communication J Autism Dev Disord, 37, 73-85.
development of infants later diagnosed with Ring, H. A., Baron-Cohen, S., Wheelwright, S.,
autism spectrum disorder. J Dev Behav Pediatr, Williams, S. C., Brammer, M., Andrew, C. and
27, S69-78. Bullmore, E. T. (1999) Cerebral correlates of
Muhle, R., Trentacoste, S. V. and Rapin, I. (2004) preserved cognitive skills in autism: a
The genetics of autism. Pediatrics, 113, e472- functional MRI study of embedded figures task
86. performance. Brain, 122 ( Pt 7), 1305-1315.
Mullen, E.M. (1995). Mullen Scales of Early Rogers, S. J., Hepburn, S. L., Stackhouse, T. and
Learning-AGS Edition. AGS Publishing, Circle Wehner, E. (2003) Imitation performance in
Pines, MN. toddlers with autism and those with other
Newschaffer, C. J., Fallin, D. and Lee, N. L. (2002) developmental disorders. J Child Psychol
Heritable and nonheritable risk factors for Psychiatry, 44, 763-781.
autism spectrum disorders. Epidemiol Rev, 24, Rothbart, M. K. (1981) Measurement of
137-153. temperament in infancy. Child Development,
Osterling, J. A., Dawson, G. and Munson, J. A. 52, 569-578.
(2002) Early recognition of 1-year-old infants Russell, J., Saltmarsh, R. and Hill, E. (1999) What
with autism spectrum disorder versus mental do executive factors contribute to the failure on
retardation. Dev Psychopathol, 14, 239-251. false belief tasks by children with autism? J
Ozonoff, S. and Jensen, J. (1999) Brief report: Child Psychol Psychiatry, 40, 859-868.
specific executive function profiles in three Sasson, N. J. (2006) The development of face
neurodevelopmental disorders. J Autism Dev processing in autism. J Autism Dev Disord, 36,
Disord, 29, 171-7. 381-394.
Ozonoff, S., Rogers, S.J., Farnham, J.M. and Schultz, R. T. (2005) Developmental deficits in
Pennington, B.F. (1993). Can standard social perception in autism: the role of the
measures identify subclinical markers of amygdala and fusiform face area. Int J Dev
autism? J Autism Dev Disord. 23, 429-441. Neurosci, 23, 125-141.
Palomo, R., Belinchon, M. and Ozonoff, S. (2006) Sigman, M., Dijamco, A., Gratier, M. and Rozga,
Autism and family home movies: a A. (2004) Early detection of core deficits in
comprehensive review. J Dev Behav Pediatr, autism. Ment Retard Dev Disabil Res Rev, 10,
27, S59-68. 221-233.
Pickles, A., Starr, E., Kazak, S., Bolton, P., Spencer, J., O'brien, J., Riggs, K., Braddick, O.,
Papanikolaou, K., Bailey, A., Goodman, R. and Atkinson, J. and Wattam-Bell, J. (2000) Motion
Rutter, M. (2000) Variable expression of the processing in autism: evidence for a dorsal
autism broader phenotype: findings from stream deficiency. Neuroreport, 11, 2765-2767.
extended pedigrees. J Child Psychol Psychiatry, Stone, W. L., Coonrod, E.E., Turner, L. M., and
41, 491-502. Pozdol, S. L. (2004) Psychometric properties of
Plaisted, K., O'riordan, M. and Baron-Cohen, S. the STAT for early autism screening. J Autism
(1998) Enhanced discrimination of novel, Dev Disord, 34, 691-701.
382

Stone, W. L., Lee, E. B., Ashford, L., Brissie, J., World Health Organization. (1994). The ICD 10
Hepburn, S. L., Coonrod, E. E. and Weiss, B. classification of mental and behavioral
H. (1999) Can autism be diagnosed accurately disorders: Clinical descriptions and diagnostic
in children under 3 years? J Child Psychol guidelines. Geneva, Switzerland: World Health
Psychiatry, 40, 219-226. Organization.
Sullivan, M., Finelli, J., Marvin, A., Garrett-Mayer, Yirmiya, N., Gamliel, I., Pilowsky, T., Feldman,
E., Bauman, M. and Landa, R. (2007) Response R., Baron-Cohen, S. and Sigman, M. (2006)
to Joint Attention in Toddlers at Risk for The development of siblings of children with
Autism Spectrum Disorder: A Prospective autism at 4 and 14 months: social engagement,
Study. J Autism Dev Disord, 37, 37-48. communication, and cognition. J Child Psychol
Tarr, M. J. and Gauthier, I. (2000) FFA: a flexible Psychiatry, 47, 511-523.
fusiform area for subordinate-level visual Yirmiya, N., Gamliel, I., Shaked, M., Sigman, M.
processing automatized by expertise. Nat (2007) Cognitive and verbal abilities of 24- to
Neurosci, 3, 764-769. 36-month-old siblings of children with autism.
Toth, K., Dawson, G., Meltzoff, A.N., Greenson, J Autism Dev Disord., 37, 218-229.
J., Fein, D. (2007) Early social, imitation, play, Zimmerman, I.L., Steiner, V.G. and Pond, R.E.
and language abilities of young non-autistic (2002). Preschool language scaes-4. San
siblings of children with autism. J Autism Dev Antonio, TX: The Psychological Corporation.
Disord, 37, 145-157. Zwaigenbaum, L. (2001) Autistic spectrum
Tzourio-Mazoyer, N., De Schonen, S., Crivello, F., disorders in preschool children. Can Fam
Reutter, B., Aujard, Y. and Mazoyer, B. (2002) Physician, 47, 2037-2042.
Neural correlates of woman face processing by Zwaigenbaum, L., Bryson, S., Rogers, T., Roberts,
2-month-old infants. Neuroimage, 15, 454-461. W., Brian, J. and Szatmari, P. (2005)
Werner, E. and Dawson, G. (2005) Validation of Behavioral manifestations of autism in the first
the phenomenon of autistic regression using year of life. Int J Dev Neurosci, 23, 143-152.
home videotapes. Arch Gen Psychiatry, 62, Zwaigenbaum L., Thurm A, Stone W, Baranek G,
889-895. Bryson S, Iverson J, Kau A, Klin A, Lord C,
Werner, E., Dawson, G., Munson, J. and Osterling, Landa R, Rogers S, Sigman M. (2007)
J. (2005) Variation in early developmental Studying the emergence of autism spectrum
course in autism and its relation with behavioral disorders in high-risk infants: methodological
outcome at 3-4 years of age. J Autism Dev and practical issues. J Autism Dev Disord. 37,
Disord, 35, 337-350. 466-480.
Wetherby, A. M., Woods, J., Allen, L., Cleary, J.,
Dickinson, H. and Lord, C. (2004) Early
indicators of autism spectrum disorders in the
second year of life. J Autism Dev Disord, 34,
473-493.
Wetherby, A. M., Allen, L., Cleary, J., Kublin, K.,
and Goldstein, H. (2002). Validity and
reliability of the communication and symbolic
behavior scales developmental profile with very
young children. J Speech Lang and Hearing
Res, 45, 1202-1218.
Wiig, E. H., Secord, W., and Semel, E. (1992).
CELF-preschool: Clinical evaluation of
language fundamentals – preschool version.
San Antonio: Psychological Corporation.
Wimpory, D. C., Hobson, R. P., Williams, J. M.
and Nash, S. (2000) Are infants with autism
socially engaged? A study of recent
retrospective parental reports. J Autism Dev
Disord, 30, 525-536.