Senior Seminar – Sibling Studies SYNTHESIS

Abstract Detecting early signs of autism is essential for timely diagnosis and initiation of effective inter-

ventions. Several research groups have initiated pro-

spective studies of high-risk populations including

infant siblings, to systematically collect data on early

signs within a longitudinal design. Despite the potential

advantages of prospective studies of young children at

high-risk for autism, there are also significant meth-

odological, ethical and practical challenges. This paper

outlines several of these challenges, including those

related to sampling (e.g., defining appropriate com-

parison groups), measurement and clinical implications

(e.g., addressing the needs of infants suspected of

having early signs). We suggest possible design and

implementation strategies to address these various

challenges, based on current research efforts in

the field and previous studies involving high-risk

populations.

Keywords Early identification Æ Screening Æ Longitudinal studies Æ Prospective studies Æ Infant Æ Autism Æ Child development Æ Siblings

Please note that the opinion and assertions contained herein are the private opinions of the authors and are not to be con- strued as official or as representing the views of the National Institute of Child Health and Human Development, the National Institute of Mental Health, or the National Institutes of Health.

L. Zwaigenbaum (&) Department of Paediatrics, McMaster Children’s Hospital at McMaster University, PO Box 2000, Hamilton, Ontario L8N 3Z5, Canada e-mail: [email protected]

A. Thurm Division of Pediatric Translational Research and Treatment Development, National Institute of Mental Health, Bethesda, MD, USA

W. Stone Departments of Pediatrics and Psychology & Human Development, Vanderbilt University, Nashville, TN, USA

G. Baranek Division of Occupational Science, University of North Carolina at Chapel Hill, Chapel Hill, NC, USA

S. Bryson Departments of Pediatrics and Psychology, Dalhousie University, Halifax, NS, USA

J. Iverson Department of Psychology, University of Pittsburgh, Pittsburgh, PA, USA

A. Kau Center for Developmental Biology and Perinatal Medicine, National Institute of Child Health and Human Development, Bethesda, MD, USA

A. Klin Yale Child Study Centre, Yale University, New Haven, CT, USA

C. Lord Department of Psychology, University of Michigan, Ann Arbor, MI, USA

R. Landa Department of Psychiatry and Behavioral Sciences, Kennedy Krieger Institute, Baltimore, MD, USA

J Autism Dev Disord (2007) 37:466–480

DOI 10.1007/s10803-006-0179-x

123

O R I G I N A L P A P E R

Studying the Emergence of Autism Spectrum Disorders in High-risk Infants: Methodological and Practical Issues

Lonnie Zwaigenbaum Æ Audrey Thurm Æ Wendy Stone Æ Grace Baranek Æ Susan Bryson Æ Jana Iverson Æ Alice Kau Æ Ami Klin Æ Cathy Lord Æ Rebecca Landa Æ Sally Rogers Æ Marian Sigman

Published online: 4 August 2006 � Springer Science+Business Media, Inc. 2006

Introduction

Overview

Several reviews over the past decade have highlighted

the importance of early recognition and specialized

intervention for improving outcomes for children with

autism spectrum disorders (ASD) (Dawson & Oster-

ling, 1997; Rogers, 1996; Smith, Groen, & Wynn, 2000).

Although recent service registry (Croen, Grether,

Hoogstrate, & Selvin, 2002) and population-based data

(Yeargin-Alsopp, et al., 2003) suggest that more chil-

dren are being diagnosed prior to age 4 years than in

the past, a formal diagnosis may still lag years behind

the time when parents initially identify concerns

(Coonrod & Stone, 2004; Howlin & Moore, 1997;

Siegel, Pliner, Eschler, & Elliott, 1988). As a result,

interest has increased in identifying and raising

awareness regarding the characteristics of ASD present

at young ages (Bryson, Zwaigenbaum & Roberts, 2004;

Landa, 2003). In addition to improving outcomes,

earlier diagnosis allows parents the opportunity to

receive counseling regarding current estimates of

recurrence risk in autism, which they may take into

account in future family planning. Research to date

supports the conclusions that one can: (1) reliably

diagnose as young as 24 months (Lord, 1995; Stone

et al., 1999); and (2) observe the behavioral markers of

autism well before 24 months (e.g., Dahlgren & Gill-

berg, 1989; Ohta, Nagai, Hara, & Sasaki, 1987; Rogers

& DiLalla, 1990).

Most of the work aimed at identifying early signs of

ASD has been retrospective, focusing on early behav-

ioral evidence of the disorder in children who have

already received a diagnosis. The most common

methods used to gather information about earlier

behaviors have been retrospective reports from parents

and analysis of early home videotapes. Although

research using these approaches has supported clinical

efforts aimed at earlier detection, many questions

regarding early signs, their timing, and their underlying

developmental mechanisms remain. Prospective

research into the early development of ASD in

high-risk infants is an exciting new frontier, and can

potentially answering these questions more systemati-

cally, while avoiding some of the biases associated with

retrospective designs. In this paper, we outline the

theoretical advantages and general feasibility of pro-

spective studies of young children at high-risk for

ASD, and acknowledge and discuss the significant

methodological, ethical and practical challenges that

accompany these studies. Issues discussed include the

design of high-risk studies, selection of comparison

groups, measurement of developmental delay and

deviance, generalizability, and clinical interpretation of

findings.

Identifying Early Signs of Autism using

Retrospective Designs

Retrospective parental reports offer a unique window

into early behaviors of children with ASD, as parents

have the advantage of observing their children’s

behavior over time and across a variety of settings.

Investigators report a wide range of symptoms that are

more common in children with autism under the age of

24 months than similar-aged children with develop-

mental delays or mental retardation (DD). Early

symptoms associated with autism cross several devel-

opmental domains, including social behavior (Dahlgren

& Gillberg, 1989; De Giacomo & Fombonne, 1998;

Hoshino et al., 1982; Ohta et al., 1987; Young, Brewer,

& Pattison, 2003), communication (Dahlgren &

Gillberg, 1989; De Giacomo & Fombonne, 1998; Ohta

et al., 1987; Young et al., 2003), affective expression

(Dahlgren & Gillberg, 1989; De Giacomo &

Fombonne, 1998; Hoshino et al., 1982), and sensory

hypo- and hypersensitivities (Dahlgren & Gillberg,

1989; De Giacomo & Fombonne, 1998; Hoshino et al.,

1982). These findings have been very important in

guiding further research aimed at identifying early signs

of ASD. However, a number of factors limit parents’

ability to provide accurate descriptions of early

behaviors. First, a parent’s incidental observations

regarding the subtle social and communicative differ-

ences that characterize young children with autism may

be limited compared to systematic assessment by

trained clinicians (Stone, Hoffman, Lewis, & Ousley,

1994). Moreover, their tendency to use compensatory

strategies to elicit their child’s best behaviors (with or

without their awareness) may affect their behavioral

descriptions (Baranek, 1999). Retrospective parental

R. Landa Department of Psychiatry and Behavioral Sciences, Johns Hopkins School of Medicine, Baltimore, MD, USA

S. Rogers Department of Psychiatry and Behavioral Sciences, University of California at Davis, Sacramento, CA, USA

M. Sigman Departments of Psychology and Psychiatry, University of California, Los Angeles, CA, USA

J Autism Dev Disord (2007) 37:466–480 467

123

reports may be also be prone to errors and distortions of

recall, especially when one asks parents to remember

behaviors that occurred many years ago. In particular,

having already received a diagnosis of autism for their

child, parents may bias their reports toward behaviors

that are consistent with the diagnosis. A recent retro-

spective study overcame some of these problems by

gathering information about behaviors under

24 months from parents of preschoolers before they had

received a diagnosis (Wimpory, Hobson, Williams, &

Nash, 2000). However, limitations of this methodology

remain, as retrospective reports are not generally

informative on the issue of whether differences in early

social and communicative development are best char-

acterized by delayed emergence, reduced frequency, or

truly abnormal or deviant quality of fundamental skills

such as joint attention.

A second strategy for obtaining retrospective

information about characteristics of autism present

before 24 months is the analysis of early videotapes of

children made by their parents. This approach has

significant strengths relative to retrospective parental

reports: it allows the observation of behaviors as they

occur in familiar and natural settings, and enables

objective rating of behavior by unbiased observers.

However, this methodology is not without its limita-

tions, the foremost being that parents record these

tapes to preserve family memories, rather than to

document their child’s behavior across a variety of

settings. As a result, tapes from different families will

naturally vary as a function of the quality of the

recording, the activities and settings that were

recorded, and the length of time the child is visible.

Moreover, if children do not behave as expected (or

desired), parents may re-record taped segments until

they obtain the desired response. Efforts to standardize

tapes across families can be extremely difficult and

time intensive (Baranek, 1999). Moreover, most

studies employing home videotapes have used children

with typical development (TD) rather than those with

DD as comparison groups, which limits the extent of

our knowledge about autism-specific deficits. Behav-

iors found to differentiate children with ASD from

children with TD under 24 months by at least two

studies are: responding to name (Baranek, 1999;

Osterling & Dawson, 1994; Osterling, Dawson, &

Munson, 2002), looking at others (Adrien et al., 1993;

Maestro et al., 2002; Osterling & Dawson, 1994;

Osterling et al., 2002), smiling at others (Adrien et al.,

1993; Maestro et al., 2002), and motor stereotypies

(Adrien et al., 1993; Baranek, 1999). Only two studies

published to date compared behaviors of children with

ASD with those of children with DD younger than

24 months; these found that children with ASD were

less likely to respond to their name (Baranek, 1999;

Osterling et al., 2002) and to look at others (Osterling

et al., 2002). Notably, analysis of home videos also

highlights that there is may be a subgroup of ‘‘late

onset’’ children whose early behavioral symptoms are

less apparent (Osterling et al., 2002; Werner, Dawson,

Osterling, & Dinno, 2000).

Potential Advantages of Prospective Studies

Retrospective parental reports and analyses of home

videos can help guide the development of early iden-

tification and screening procedures (as argued by Fil-

ipek et al., 1999), but these procedures must ultimately

be validated empirically in prospective studies, with

sufficient follow-up of both screen positive and screen

negative children to allow estimates of sensitivity and

specificity. In fact, prospective studies of high-risk

infants (which, until recently, have been rare in autism)

may also identify novel behavioral (and biological)

markers that show the way forward in developing more

effective early identification and screening measures.

Prospective studies are not subject to recall biases, they

can be designed to examine specific constructs of

interest, and they can provide comparable data col-

lection points and methods across children. Perhaps,

most importantly, these studies allow collection of data

longitudinally across different ages, which can foster

our understanding of developmental trajectories and

the impact of early delays in one domain (e.g., social

orienting) on the subsequent development of another

(e.g., language).

High-risk samples have informed studies of other

neurodevelopmental and neuropsychiatric conditions,

including language/reading disorders (Carroll &

Snowling, 2004), attention deficit hyperactivity disor-

der (Faraone, Biederman, Mennin, Gershon, & Tsu-

ang, 1996), bipolar affective disorder (Chang, Steiner,

& Ketter, 2000; Geller, Tillman, Craney, & Bolhofner,

2004), and schizophrenia (Schubert & McNeil, 2004).

Prospective studies of siblings and offspring of affected

probands have generated significant insights regarding

premorbid development and predictors of illness in

these high-risk groups. For example, children with

schizophrenic parents have attention and verbal

memory deficits, gross motor delays, and dysfunction

of smooth-pursuit eye movements (Erlenmeyer-Kim-

ling, 2000; Schubert & McNeil, 2004), and children with

a parent or sibling with dyslexia have greater difficulty

with phonological processing than age-matched low-

risk controls, despite normal early language develop-

ment (Carroll & Snowling, 2004). Notably, these

468 J Autism Dev Disord (2007) 37:466–480

123

studies generally focus on group differences between

high- and low-risk children rather than the association

between early markers and outcome status, because of

insufficient power and/or follow-up. In contrast,

autism can be diagnosed in early childhood, so out-

comes can be determined after a relatively short

follow-up period. Hence, one can study autism pro-

spectively much more easily (i.e., with fewer resources

and with less risk of sample loss) than an adult-onset

disorder such as schizophrenia.

Prospective Studies in Autism: Siblings and other

High-Risk Groups

Several populations at increased risk of ASD that can

be identified in early childhood: children with early

signs of autism or developmental delays (DD) identi-

fied through population screening, children at

increased risk of autism due to specific medical diag-

noses or genetic anomalies, and the main focus of this

paper, infants with an older sibling with ASD.

At least two research groups have studied early

signs of autism in high-risk samples identified by

population screening. Charman et al., (1997) and

Swettenham et al., (1998), reported on a high-risk group

of children who failed the Checklist for Autism in

Toddlers (CHAT; Baron-Cohen, Allen, & Gillberg,

1992), a screening measure administered at 18 months

of age. Detailed assessment of social, communication

and play skills was completed at 20 months, and

diagnostic outcomes were assessed at age of three and a

half. Children subsequently diagnosed with autism were

compared to those subsequently diagnosed with devel-

opmental delay based to their 20-month skills. At this

early age, the children with autism spent less time

looking at adults during free play (Swettenham et al.,

1998), were less likely to look at the face of an adult

feigning distress (Charman et al., 1997), showed less

gaze switching between people and objects (Charman

et al., 1997; Swettenham et al., 1998), and showed less

imitation (Charman et al., 1997). Wetherby et al., (2004)

followed a group of children who had failed communi-

cation screening using the Communication and Sym-

bolic Behavior Scales Developmental Profile (CSBS

DP, Wetherby & Prizant, 2002). They obtained video-

tapes of the CSBS Behavior Sample at a mean age of

18–21 months for children who received later diagnoses

of autism, DD, or who were typically developing.

Specific features that differentiated children with

autism from the other two groups include social-

communication behaviors (e.g., reduced eye gaze,

coordination of gaze with other nonverbal behaviors,

directing attention, responding to name, and unusual

prosody) and repetitive body and object use. Notably,

the content of the initial screen (i.e., children are se-

lected based on a particular profile of early signs), may

introduce sampling biases, and the fact that data are

only collected from the point of first screening onward

limits the age range over which autism can be studied.

There are also children at increased risk for autism

due to medical risk factors, such as Fragile X syndrome

(Rogers, Wehner & Hagerman et al., 2001), specific

chromosome abnormalities (Xu, Zwaigenbaum, Szat-

mari & Scherer, 2004), tuberous sclerosis (Bolton &

Griffiths, 1997), and prenatal exposure to valproic acid

(Williams et al., 2001) or thalidomide (Stromland et al.,

2002). However, these specific risk factors are all rel-

atively rare and would be difficult to study in large

numbers, and may be associated with unique clinical

features that may not generalize to other children with

autism.

There is growing interest in studying infant siblings

of children with ASD, who are arguably the most

clearly defined high-risk group available. Notably,

Baron-Cohen et al., (2002) originally developed the

CHAT screening algorithm based on items that, at

18 months, were atypical in four siblings subsequently

diagnosed with autism. More recent reports by Pilow-

sky and colleagues (Pilowski, Yirmiya, Shalev, &

Gross-Tsur, 2003; Pilowski, Yirmiya, Doppelt, Gross-

Tsur, & Shalev, 2004) support the feasibility of study-

ing early development in siblings and Zwaigenbaum

et al., (2005) reported several behavioral markers

which, at 12 months, predict a subsequent diagnosis of

autism in a sibling sample. In addition, Landa & Gar-

rett-Mayer (2006) report developmental levels and

trajectories in that differentiate infant siblings later

diagnosed with autism spectrum disorder, beginning at

6 months of age. Autism is associated with the highest

relative risk in siblings, compared to general popula-

tion of all the neuropsychiatric disorders (Szatmari,

Jones, Zwaigenbaum & MacLean, 1998). Previous

studies found rates of autism in siblings of children

with autism range from 3% to 5%, which is at least

20 times higher than rates of autism in the general

population (Bailey, Phillips & Rutter, 1996; Simonoff,

1998; Szatmari et al., 1998). In fact, estimates of

recurrence risk (that is, the risk to later-born children)

may be as high as 8.6% when one child in the family

has autism, and 35% when two siblings have autism

(Ritvo et al., 1989). Notably, these risk estimates may

be somewhat conservative, as they come mainly from

studies conducted over 20 years ago, using more

restrictive diagnostic criteria (DSM-III).

The risk to relatives of individuals with autism also

extends beyond the traditional boundaries of the

J Autism Dev Disord (2007) 37:466–480 469

123

autistic spectrum (Bailey, Palferman, Heavey, & Le

Couteur, 1998). Family members have higher rates of

certain psychiatric and developmental disorders, com-

pared to individuals with no family history of autism

(Landa, Folstein & Isaacs, 1991; Landa et al., 1992;

Pickles et al., 2000; Piven, Palmer, Jacobi, Childress, &

Arndt, 1997; Smalley, McCracken, & Tanguay, 1995;

Yirmiya & Shaked, 2005). As well, because of the early

age of diagnosis of the proband, one can ascertain sib-

lings in early infancy or even prenatally, making it

possible to study early neurodevelopmental mecha-

nisms, and to partially avoid (or at least to systemati-

cally measure) the impact of potentially confounding

environmental factors. Infant sibling research offers

unique opportunities to study the neural origins and

developmental cascade that leads to autism, potentially

providing new insights into its neurobiology, improved

methods of early detection, and earlier opportunities for

intervention.

In August 2003, the National Alliance for Autism

Research (NAAR) and the National Institute of Child

Health and Human Development (NICHD) co-spon-

sored a workshop for researchers engaged in the study

of populations of young children at high-risk for autism,

particularly siblings of children with autism. Despite the

theoretical advantages and exciting opportunities

associated with this research design, there are clearly

significant methodological, ethical and practical chal-

lenges facing researchers studying young children at

high-risk for autism. In the remainder of this paper we

outline several of these challenges, including those

related to sampling (e.g., recruitment of adequately

sized samples, determining inclusion/exclusion criteria

for high-risk infants and appropriate comparison

groups), measurement (e.g., selection of constructs and

measures) and clinical implications (e.g., clinical man-

agement of infants who appear to have early signs of

ASD). We suggest possible design and implementation

solutions for these various challenges, based on current

research efforts in the field and previous studies

involving high-risk populations. These issues have

implications not only for research with infant siblings,

but also for research in other aspects of early charac-

terization and diagnosis of autism.

Issues Related to Sampling

Sample Size

High-risk studies in other fields (e.g., schizophrenia,

dyslexia) have generally been designed to compare

siblings with controls on a group basis, without

knowing the ultimate outcomes of individual siblings

(Carroll & Snowling, 2004; Erlenmeyer-Kimling,

2000). Initial infant sibling studies of ASD by Pilowski

and colleagues (2003, 2004) also focus on group com-

parisons. However, if the main objective of a sibling

study is to identify early markers that are predictive of

a specific diagnosis, then individual outcomes become

important, and one must power the sample size with

reference to the expected number of participants who

will have the diagnosis of interest (i.e., not the total

number of infants enrolled).

The required sample size for these studies will

depend on the specific research question posed. A few

issues are considered for illustration. First, if one

defines the outcome of interest more broadly (e.g.,

language delay), there will be a larger number of sib-

lings with that outcome, potentially making it easier to

detect differences between ‘affected’ and ‘unaffected’

siblings. However, predictors of secondary outcomes

such as language delay may not generalize outside of an

autism sibling sample, limiting the clinical utility of such

findings. A second issue to consider is the strength of

the association between the predictor variables and

outcome under study (e.g., the sensitivity and specificity

of early markers for the subsequent diagnosis of aut-

ism), which will influence the power to detect a rela-

tionship. However, the investigator may sometimes

select predictor variables on a theoretical basis, so the

actual strength of the relation between predictor and

outcome variables may be difficult to estimate with

confidence. A third variable to consider is the number

of outcomes/variables being studied; for example,

contrasting siblings by more than two outcomes (e.g.,

ASD versus developmental delay versus typical devel-

opment), or examining the effects of stratification

variables (e.g., gender) within and across groups may be

of interest, but will require even larger samples sizes.

Due to limitations on numbers of infants born to

older siblings with autism within specified geographic

regions, studies may maximize their efforts to collect

data in a timely fashion by establishing collaborations

across multiple sites and utilizing a common set of

core assessment measures. Such collaborations accel-

erate the process of identifying early predictors of

outcome by increasing the collective sample size so

that investigators can address more refined questions

about outcomes and predictors. Although collabora-

tions between research groups require additional ef-

fort and resources to support the necessary steps of

ensuring consistency in methods and measures, as well

as inter-rater reliability for observations, these proce-

dures allow the examination of consistency of findings

across sites, ensure the fidelity of assessment

470 J Autism Dev Disord (2007) 37:466–480

123

measures, and facilitate future attempts to replicate

findings.

Inclusion/Exclusion Criteria

Decisions regarding inclusion/exclusion criteria for

siblings also depend on the goals of the study. One

major consideration is whether to include probands

and/or siblings with conditions associated with autism

(e.g., tuberous sclerosis, Fragile X syndrome) and those

with other medical risk factors that may predispose the

infant to developmental problems (e.g., low birth

weight, perinatal injuries). If the main goal is to study

early signs and neurodevelopmental mechanisms of

autism in high-risk infants, then there may be some

flexibility in whether to exclude probands and siblings

with known risk factors. Children with such risk factors

may differ developmentally from other children with

autism, an interesting and clinically relevant issue to

explore. However, if a major goal is to identify phe-

notypes and endophenotypes that ‘run true’ in families

for subsequent genetic linkage studies, or to estimate

recurrence rates associated with ‘idiopathic autism,’

then studies may need to exclude cases with known risk

factors.

Issues Related to Study Design

Within a high-risk (or sibling) design, several decisions

need to be made, including enrollment age for siblings,

selection of comparison groups, and approach to out-

come assessment. Scientific and practical consider-

ations must guide each of these decisions.

Enrollment Age of Siblings

The main strength of the high-risk design is the

potential to study ASD earlier than would be possible

by ascertaining children at the time of diagnosis, which

rarely occurs before age 2 years. There may be

advantages to starting assessment of high-risk infants

either during the first or the second year of life,

depending on the particular focus of the study.

Although studying autism in the first year of life is

largely uncharted territory, this strategy may be an

optimal way to learn about atypical patterns of infant

development that underlie later manifestations of

autism. The few extant findings examining children in

the first and second year of life indicate increased

subtlety of impairments at earlier ages and a number of

measurement challenges. For instance, two studies that

focused specifically on children younger than

12 months (Baranek, 1999; Werner et al., 2000) found

that children with autism show reduced social orienting

compared to typically developing children, but in

general, find fewer differences between the two groups

than analyses of videos taken at 12 months or later. As

well, preliminary data from ongoing sibling studies find

that behavioral risk markers more readily distinguish

autism at 12 months than at 6 months (Zwaigenbaum

et al., 2005). Studying children during the first year of

life presents tremendous opportunities to examine

early neurodevelopmental mechanisms that may

determine later impairments and developmental tra-

jectories in autism (for example, social orienting and

gaze monitoring; Moore & Corkum, 1998; Phillips,

Wellman & Spelke, 2002). Moreover, preliminary

findings that atypical brain growth (Courchesne, Car-

per & Akshoomoff, 2003) may predate behavioral

differences in autism emphasize that studies that target

high-risk infants earlier in life may yield unique data on

early markers. Another advantage of enrolling siblings

at 6 months or younger is the potential to reduce the

problem of biased sampling (and inflated recurrence

risk estimates) resulting from over-referral of parents

who have behavioral concerns.

Prospective studies of toddlers at high-risk of autism

starting in the second year of life are also informative.

Recruitment at this age can include high-risk children

other than siblings (e.g., population screening on the

basis of delays in communication skills; Wetherby et al.,

2004), allowing comparison of children with ASD

across different ascertainment routes, helping to ensure

the generalizability of findings. As well, while it may be

easier to study basic developmental mechanisms in the

first year of life, a more substantial empirical basis exists

for studying behavioral markers and early signs of

autism in the second year. These studies may lead to the

development of new screening measures (or validation

of existing measures), and generate educational strat-

egies to help improve early detection of autism in the

general community, such as in the ‘First Words’ initia-

tive (Wetherby et al., 2004). Recruiting siblings during

the second year of life may also be less resource-

intensive than recruiting younger infants.

Frequency of Assessments

The optimal age interval to detect the onset of autistic

symptoms and/or regression remains an empirical

question. Research on typically-developing populations,

as well as research on children with developmental

disorders, indicates that there are ‘‘critical periods’’ for

development of skills typically delayed or absent in

autism—such as between the ages of 6 and 18 months

J Autism Dev Disord (2007) 37:466–480 471

123

when social-communicative behaviors such as joint

attention skills, pointing, and imitation are consolidating

(Corkum & Moore, 1998). Multiple assessments within

such critical period could be extremely informative for

the timing and developmental sequence of these

impairments. However, there are potential trade-offs

between the rich detail afforded by frequent assessment,

and the cost, burden on parents, and potential practice

effects on some standardized measures. Frequent

assessments may be most feasible using naturalistic

observations that do not lead to test-related learning,

such as videotaped maternal-infant interaction samples

to track social development (Hsu & Fogel, 2003), speech

samples or vocabulary checklists to track language

development (Tsao, Liu & Kuhl, 2004) or parent diaries

or report forms tracking the emergence of behaviors

such as gestures (Crais, Day, & Campbell, 2004). The use

of parent questionnaires or diaries and video or audio-

taped behavior samples from home can facilitate data

collection. One can reserve standardized assessment of

language, cognition and adaptive function for ‘‘land-

mark’’ evaluations at less frequent intervals, depending

on study design and the minimal allowable testing

interval on particular tests. Some studies might combine

microanalysis of the development and emergence of

early social-communication processes (i.e., frequent

quantitative and qualitative analyses of operationally

defined, spontaneously occurring behaviors) with

macroanalysis of developmental trajectories in broad

domains of functioning.

Comparison Groups

If one of the goals of a high-risk study is to identify

early autism-specific markers, then comparison groups

are essential to control for potential confounding

variables and to minimize potential sources of bias. If

early markers identified in high-risk samples are to be

useful in the general population community samples or

clinically referred samples (i.e., to guide first- and

second-level screening and surveillance), it is impor-

tant to know not only whether these markers can dis-

tinguish autism from typical development, but also

whether they distinguish autism from language delays

and/or other developmental delays. One should base

the selection of comparison groups and matching

variables in sibling studies on the populations to which

the research findings will be applied.

In some ways, subgroups of the sibling sample itself

are ‘‘built in’’ comparison groups of infants who will

have outcomes other than ASD. Based on previous

family studies in autism, we might anticipate that in

addition to the 5–8% of siblings who develop ASD,

approximately 10–20% will exhibit milder impair-

ments, including language delay (Bailey et al., 1998;

Folstein et al., 1999; Murphy et al., 2000), leaving about

70% to develop typically. Comparing siblings who

develop ASD, to siblings who do not, controls for two

important factors: (1) the potential impact of exposure

to an older sibling with ASD (and to related psycho-

social stressors on the family); and (2) the possible

expectation bias of increased risk of ASD on the part

of the examiner (i.e., it may be difficult to maintain

blinding to sibling status) or parent rater. However,

there are also important limitations to this approach,

not the least of which is the possibility of misclassifi-

cation error at the point of the initial outcome assess-

ment. For example, some siblings classified as

‘‘typically developing’’ based on standardized mea-

sures may in fact have mild impairments that may

become more apparent at a later age. As well, some

children who are classified as delayed or as having

symptoms of a ‘‘broader autism phenotype’’ may later

receive a diagnosis of an ASD (particularly Asperger’s

syndrome) later on. Misclassification errors will tend to

minimize differences between groups and reduce

power. Other groups of typically developing and

developmentally delayed children may also include

some who would be classified differently as they get

older, but this is more likely to be an issue for siblings

of children with ASD because of their genetic liability.

As well, siblings with developmental delays may not be

representative of other children with delays. In par-

ticular, although siblings of children with autism are

not known to be at higher risk of global cognitive delay

unless they also have an ASD (Fombonne, Bolton,

Prior, Jordan, & Rutter, 1997; Szatmari et al., 1993),

they may have specific language impairments (Dawson

et al., 2002; Landa & Garrett-Mayer, 2006).

Thus, in addition to siblings of children with autism

who do not develop an ASD, one should consider

additional comparison groups. For example, some

studies may benefit from having low-risk groups that

control for the effects of being a later-born child, such

as infant siblings of typically developing children with

no family history of ASD. Including groups to serve as

controls for the developmental delays that often

accompany ASD is also important to consider,

although the selection and recruitment of such groups

is a challenge (Szatmari, Zwaigenbaum & Bryson,

2004). With the exception of children with identified

syndromes (who are unlikely to be representative),

even seemingly high-risk populations, such as siblings

of children with developmental delay, may include a

relatively small proportion who will ultimately receive

a diagnosis of developmental delay and may not cover

472 J Autism Dev Disord (2007) 37:466–480

123

the full spectrum of delays that one might observe in an

unselected sample (Crow & Tolmie, 1998). One may

find delays of a broader range of severity among

infants referred to early intervention programs due to

constitutional and/or psychosocial risk factors (Allen,

1993), and among infants attending a neonatal follow-

up clinic due to prematurity (e.g., Bucher, Killer,

Ochsner, Vaihinger, & Fauchere, 2002). Alternatively,

if the study follows children starting at a sufficiently

advanced age, then one can utilize a comparison group

of children ascertained directly by developmental or

communication delays (for example, through popula-

tion screening; Wetherby et al., 2004). However, a

substantial proportion of children identified due to this

type of delay in the second year of life may ultimately

receive a diagnosis of ASD (Robins, Fein, Barton, &

Green, 2001), so group comparisons may not be valid

or robust until one follows samples to an age at which

diagnostic classification is relatively stable (i.e., at least

3 years of age).

Once one select comparison groups, one should

consider other potential confounds between risk sta-

tus and outcome measures as potential matching

variables (Jarrold & Brock, 2004; Szatmari et al.,

2004). Such confounds might include age, gender

(since autism and language delays are more prevalent

among boys than girls), and birth order (since early

infant behaviour may be influenced by exposure to

older siblings). One should also consider matching on

parental education and/or socioeconomic status.

Although neither factor is known to affect rates of

autism, each may influence rates of other relevant

outcomes such as developmental delays and behav-

ioral disorders.

Outcome Assessment

Several ongoing studies of young infants use end-

points of at least 3 years of age, although investigators

may determine and communicate diagnoses to the

family before this time (Zwaigenbaum et al., 2005).

This approach is consistent with evidence that the

stability of autism spectrum diagnoses increases sig-

nificantly by this age (Lord & Risi, 2000). Ideally, the

diagnostician should be blind to the child’s group and

previous evaluation data to reduce expectation biases.

Diagnosis should also be based on expert assessment

using standardized measures (e.g., the Autism Diag-

nostic Interview—Revised and the Autism Diagnostic

Observation Schedule) and best clinical judgment

based on ICD-10 or DSM-IV-TR criteria. There is

currently very little published concerning the agree-

ment between the ADI-R and ADOS (either in

combination or singly) with clinical diagnosis based

on DSM-IV (de Bildt et al., 2004), although it is well-

established that diagnostic agreement in general cor-

relates positively with the experience of the clinician

(Stone et al., 1999; Volkmar et al., 1994). In that

regard, one should also consider the additional step of

having expert clinicians review all available clinical

data and then reach a consensus best estimate diag-

nosis (as is done in some genetic studies; see Ma-

cLean et al., 1999). However, we do not yet know

whether clinical experience with older preschool

children will ensure stability of autism diagnoses in

toddlers. At present, there is little data on the sensi-

tivity and specificity of measures such as the ADI-R

and ADOS in children under age 2–3 years, so the

interpretation of these measures requires careful

clinical judgment (Lord & Risi, 2000; Moore &

Goodson, 2003).

Issues Regarding Measures

Constructs for Measurement

Given the hypothesis that high-risk infants have

increased rates of language disorders, impaired cogni-

tive abilities, atypical social behaviors and other fea-

tures of the broader autism phenotype, assessments

should measure development across multiple domains

over time in order to capture the breadth of outcomes.

A comprehensive developmental approach grounded

in a thorough intellectual ability assessment is neces-

sary, as one needs to consider constructs such as play,

imitation, language and social interaction in the con-

text of the young child’s cognitive abilities. Develop-

mental assessment should include measures of

expressive and receptive language, adaptive behavior

and overall cognitive profile (see Klin, Chawarska,

Rubin & Volkmar, 2004 for a review). Although it is

challenging to find cognitive tests that include sufficient

nonverbal as well as verbal components, and that one

can use across a reasonable developmental range

without floor or ceiling effects, such assessments will

allow for outcomes such as mental retardation and

specific language impairment to be distinguished from

autism spectrum disorders. One of the challenges at

the outset is that most available measures are designed

to detect quantitative delays in early development

(e.g., smaller vocabularies, lower age equivalent scores

in various areas of cognition) but not atypical or

qualitatively abnormal or deviant patterns of skill

development (e.g., splinter skills, atypical develop-

mental sequence) that may ultimately be more specific

J Autism Dev Disord (2007) 37:466–480 473

123

to autism. Data on developmental trajectories of lan-

guage and cognitive skills may ultimately be more

informative than profiles from any single point in time,

another advantage of studying autism in high-risk

samples using a longitudinal design.

Because the existing diagnostic criteria for autism

(APA, 2000) are not necessarily suitable for diagnosing

very young children, evaluation of young children for

signs of autism (or related communication or social

problems) must include assessment of underlying

developmental constructs. For example, early charac-

teristics of autism evident in children younger than

2 years may likely include subtle deficits such as vari-

able eye gaze, inconsistent joint attention skills,

reduced vocal and/or motor imitation, and repetitive or

abnormal use of objects (Zwaigenbaum et al., 2005).

These behaviors or skill deficits may be markers for

disrupted underlying mechanisms, such as attentional

control, executive functioning, preferential orientation

to social stimuli, social motivation, face processing and

auditory processing (Volkmar, Lord, Bailey, Schultz, &

Klin, 2004). Measures of some of these analogue skills

including joint attention (Mundy, Sigman & Kasari,

1990) and imitation (Rogers, 1999; Stone, Ousley &

Littleford, 1997) have become quite refined. In addi-

tion, some investigators have reported measures of

face processing (Dawson & Zanolli, 2003) and eye-

gaze tracking (Chawarska, Klin, & Volkmar, 2003) in

very young children. However, the field generally lacks

well-validated measurements for most neuropsycho-

logical processes in very young children. Although this

presents an initial challenge to prospective studies of

autism, high-risk samples may provide the necessary

developmental substrate to evaluate innovative mea-

sures focused on early impairments and underlying

mechanisms in autism. Moreover, longitudinal studies

that assess the persistence and developmental pro-

gression of atypical behaviors and skills deficits offer a

significant advantage over previous cross-sectional

research.

Measures of Delay or Deviance

Early indicators of autism may present more as the

absence of expected behaviors rather than as the

presence of obvious behavioral aberrations. Measures

that ‘‘press’’ for social or communication behaviors

that are often delayed or deviant in children with

ASD would seem appropriate for assessing high-risk

infants. For example, the Autism Observation Scale

for Infants-AOSI; Zwaigenbaum et al., 2005),

recently developed for the purpose of assessing early

signs of ASD across a range of developmental

domains, adopts this approach. Similarly, the Com-

munication and Symbolic Behavior Scales-Develop-

mental Profile includes several specific play

interactions that press for early social communicative

behaviors, including measures of joint attention

(Wetherby & Prizant, 2002). However, a brief period

of observation in a research lab may not easily cap-

ture the range of contexts and facilitating/interfering

conditions that influence these behaviors in everyday

situations. As a result, one should obtain information

about the persistence, quality and frequency of social

responses from parental report as well as observation,

with special attention paid to how one elicits the

responses and how much parental prompting, sup-

ports, and accommodations are required. For in-

stance, in addition to level and type of

communication, examination of the rate of commu-

nicative behaviors during ‘‘typical’’ social situations

may be informative (Charman et al., 2005). One also

needs data on the quality and context of observed

behaviors to complement simple frequency counts.

Contributions of infant development experts may be

critical for identifying measures that capture the

variability of typical infant development with respect

to social-communicative behaviors.

Measurement of Atypical Behaviors

Measurement of atypical behaviors in young children

is also challenging. The types of unusual behaviors

seen in very young children with autism, such as

seeking or avoiding specific types of sensory responses

and input and repetitive motor behaviors, are partic-

ularly difficult to measure because they vary in pre-

sentation within and across children. Stereotypic

motor behaviors may also be less frequent at very

young ages, at least by parental report (Stone et al.,

1994) and may be difficult to distinguish from the

normal rhythmic movements observed in typically

developing infants (Thelen, 1981a; b). What is pre-

dictive of autism may not simply be the type of

behavior, but rather, the persistence, quality, fre-

quency and contexts under which the behavior is ob-

served—but determining this will require careful

quantitative and qualitative analysis, and appropriate

comparison groups. There are very little normative

data on the development of sensory preferences in

typical infants against which to compare the sensory

behaviors of infants at increased risk for ASD. It is

essential that measures of repetitive behaviors and

sensory interests be normed in typically developing

infants so one can meaningfully interpret the signifi-

cance of findings in high-risk populations.

474 J Autism Dev Disord (2007) 37:466–480

123

Issues Related to Generalizability

Potential Differences between Participants and

Non-Participants

Given that investigators may ultimately use the findings

from prospective studies of high-risk samples to assist

with identification of early signs of autism in the general

population, it is important to consider which factors

may influence participation rates. First, specific con-

cerns may motivate parents to enroll their younger in-

fants. This selection factor does not necessarily imply

that early development in this group will differ from

that of other infants subsequently diagnosed with aut-

ism. However, if parents are more sensitive to atypical

development in one domain compared to another (e.g.,

verbal language versus motor imitation), this factor may

bias the phenotypic distribution of participating infants.

This bias is most explicit in high-risk studies that use

specific screening tools to identify their participants

(Wetherby et al., 2004), but may also be an issue in

sibling studies. Given that early concerns may influence

participation rates, one must interpret estimates of

recurrence risk from infant sibling studies cautiously.

Second, the characteristics of the proband (e.g., level of

function, severity of symptoms) may influence parents’

perceptions of risk and hence, likelihood of participat-

ing. Similarly, parents with other children or relatives

with autism or autism-related conditions may also per-

ceive greater risk in their infants. Finally, other family

characteristics can influence research participation rates

in general, such as socioeconomic status, parental edu-

cation, and family composition (e.g., single versus two-

parent family, number of siblings).

Potential Differences between Siblings who

Develop ASD and other Children with ASD

Children with ASD ascertained through an affected

sibling may differ from other children with ASD. For

example, differences in genetic factors, that is, genes

that lead to higher recurrence rates, may influence the

clinical expression of autism. Notably, a slightly higher

rate of the BAP occurs in extended relatives when

there are two affected children in a sibship (Szatmari

et al., 2000). Differences in early development may

also result from the very fact that there is already a

child in the family with a diagnosis of autism. For the

second affected child, this may lead to earlier recog-

nition of symptoms and initiation of intervention, as

well as differences in parent–child interactions, influ-

enced both by parents’ previous experience with aut-

ism and the added stress of parenting an older child

with special needs. Parents often raise the question as

to whether some behaviors may result from interac-

tions with the older sibling with autism. However, most

available data on early markers of autism in young

children point to the absence of typical social-commu-

nicative behaviors—which would be less influenced by

interaction with siblings—rather that the presence of

atypical, potentially learned behaviors (Baranek, 1999;

Dawson & Osterling, 1997; Rogers & DiLalla, 1990).

Notably, cross-sectional studies have failed to identify

differences in autistic symptoms or level of function

between children with autism who have a sibling with

autism, and children with autism who do not (Cuccaro

et al., 2003). Comparing developmental trajectories

unfolding into later stages of childhood of children

with ASD ascertained through sibling studies with

those of children referred early (e.g., under 2–3 years)

for diagnostic assessment may shed further light on

potential differences between the two groups.

Clinical Issues Related to following High-Risk Infants

Addressing Concerns

Assessment and identification of possible early mark-

ers of autism have important clinical implications for

individual participants and their families. Discussing

and responding to clinical concerns are, inevitably,

major components (and major responsibilities) in the

day-to-day operation of prospective studies of high-risk

infants, and present challenging clinical and ethical

issues. First, what are clinically sensitive, yet scientifi-

cally rigorous, approaches to eliciting parents’ con-

cerns? How do we best collect information about

parental impressions across a broad range of domains

or test hypotheses about specific early signs without

creating concerns or raising parental anxiety? One

approach is to ask open-ended questions about par-

ticular developmental domains (e.g., ‘‘Describe your

child’s play interests.’’). This approach may yield richer

information than a checklist of atypical behaviors.

Opportunities to observe the child’s naturally occur-

ring behaviors and responses to experimentally

designed presses also reduce the potential burden on

parents to be the sole source of information on early

signs.

Second, how do researchers communicate concerns

that arise from their assessments? The involvement of

an experienced clinician is critical for this aspect of the

project. Providing feedback to parents regarding stan-

dardized measures of language, motor, and cognitive

development is relatively straightforward when the

J Autism Dev Disord (2007) 37:466–480 475

123

child’s performance is consistent with age expectations;

however, one requires clinical expertise to interpret

and communicate assessment results when delays are

found. It may also be possible to share some observa-

tions made during administration of experimental

measures. However, interpretation of the severity of

developmental delays or of performance on specific

experimental tasks in relation to future risk of autism is

much more difficult. At the outset of a study, the

relation between early findings and risk of autism is

generally unknown, and only descriptive feedback is

possible. On the other hand, as the number of children

who complete the study protocol to outcome assess-

ment increases, data accumulate regarding the predic-

tive validity of early markers. At what point is there an

ethical obligation to share this information with par-

ticipants? This issue should be given careful consider-

ation in the study development and design, and in some

situations may warrant ongoing consultation with an

independent ethics committee.

Third, how do researchers respond to concerns that

parents communicate spontaneously? Handling

parental concerns about their infants or toddlers

clearly requires clinical sensitivity and acumen. One

must acknowledge concerns and treat them with

appropriate seriousness, even if their implications for

course and/or prognosis are unknown.

Clinical Diagnosis

Ethical standards dictate that researchers follow cur-

rent best practice in dealing with diagnostic issues in

high-risk samples. When children meet DSM-IV cri-

teria for ASD, the diagnosis must be communicated to

parents in a timely way to ensure that they can obtain

appropriate services for their child. In some cases a

clinical diagnosis is appropriate before one schedules

the child’s formal outcome assessment. This procedure

may have an impact on outcome assessment itself, to

the extent that early intervention accelerates skill

development and reduces symptoms. However, one

can still complete an independent (and optimally

blind) assessment of diagnosis (e.g., at age 3 years),

thus providing an opportunity to assess stability of

early diagnoses in this group. Moreover, if researchers

fail to communicate diagnoses when criteria are met,

parents may seek support elsewhere and elect to opt

out of studies. Selective drop-out of children with

diagnoses may prove to be a greater threat to longi-

tudinal research in high-risk samples than the effects of

early intervention.

Longitudinal studies of referred samples indicate

that the vast majority of children receiving an expert

clinical diagnosis of ASD at 24 months retain that

diagnosis in an independent assessment at age 3 years

(Lord, 1995; Stone et al., 1999), but there may still be

some diagnostic changes up until age 7 (Charman et al.,

2005). However, only a small percentage of children

with ASD are referred prior to 24 months and children

with the most severe symptoms may predominate, an

ascertainment bias that may inflate estimates of diag-

nostic stability. A high-risk sample may be more likely

to include children across the full range of ASD

severity, including those with milder or more subtle

symptoms.

There may even be instances in which a child

appears to meet DSM-IV criteria for ASD even earlier

than 2 years. However, there are currently few data on

the stability of diagnoses made prior to 2 years, and as

noted earlier, there are no guidelines on how to

interpret scores on standardized measures such as the

ADI-R and ADOS, or even how to interpret DSM-IV

criteria for children in this age group. Boundaries

between ‘‘early markers’’ (those atypical behaviors

which have a statistical association with a later diag-

nosis of autism) and ‘‘diagnostic markers’’ (atypical

behaviors which provide evidence that DSM-IV crite-

ria are currently met) are ill-defined. Although this

situation presents a clinical dilemma in current studies

of high-risk infants, prospective research in this area

provides a unique opportunity to develop diagnostic

criteria that are more developmentally appropriate for

this age group.

The current emphasis on avoiding delays in diag-

nosis places a strong focus on children with ASD who

are missed by early identification and screening efforts

(false negatives). However, particularly as we begin to

test the limits of our clinical experience regarding

assigning diagnosis to toddlers with strong evidence of

autism, we must also consider the significance of mis-

classification errors in the opposite direction (i.e., false

positives, children who do not retain a stable diagnosis

of autism or move in an out of ASD) (Charman et al.,

2005). Although children with other developmental

conditions may also benefit from early referral to

intervention services, clinical best practice requires

careful follow-up to at least an age where diagnostic

stability is better established, and sensitive but open

discussion at the time of diagnosis regarding possible

change in status over time.

Intervention Referrals

Given the discrepancy between accelerating knowl-

edge concerning early behavioral markers for autism

and the lack of proven interventions for children

476 J Autism Dev Disord (2007) 37:466–480

123

under the age of 2 years, combined with the notion

that earlier intervention is highly desirable to maxi-

mize the chances of a positive outcome (Lord &

McGee, 2001), the process of referring families for

intervention is complex. To fulfill ethical require-

ments, informed consent must address what will occur

when study measures indicate that a child has a sig-

nificant problem, the criteria for which should be

specified a priori (Chen, Miller & Rosenstein, 2003).

Clinicians are obliged to refer children for treatment

when they believe it is clinically necessary for facili-

tating the child’s development as well as providing

support to the parents. Developmental services have a

responsibility to offer interventions targeting chil-

dren’s specific needs (mandated by law in the U.S.)

However, local providers of early intervention may

have limited experience in delivering interventions

specialized to the social-communicative needs of

children younger than age 2 years, indicating a critical

need for further research in this area. One should

carefully document any intervention received by

participants with respect to modality (targeted skills/

ideology), setting (home based versus clinic/center

based), and, critically, intensity (hours per week) to

try to factor such interventions into outcome analyses.

Notably, we currently lack efficacy data for inter-

ventions targeting early signs of autism in this age

group, so it will be difficult to determine whether the

interventions change developmental trajectories or

whether gains related to the natural unfolding of

developmental processes. Controlled clinical trials of

interventions that target the specific deficits of autism

yet are developmentally appropriate to young infants

and toddlers are essential to resolve this issue. How-

ever, until such data are available, the absence of a

clear-cut standard of care for at-risk children will

leave a significant degree of ambiguity regarding

appropriate and ethical referral decisions, a situation

which investigators note in other samples at high-risk

of psychopathology (Heinssen, Perkins, Appelbaum,

& Fenton, 2001, p. 572).

Summary and Future Directions

In summary, general recommendations for the field

with respect to high-risk research include the need to

pay critical attention to methodological rigor as well as

human subjects concerns and practicalities in engaging

families in research, retaining their research partici-

pation, and ethically considering appropriate parental

involvement and feedback. Specific recommendations

include a careful consideration of issues related to

recruitment and sampling, the need to follow infant

participants closely during ‘‘critical’’ age periods (6–

18 months), the need to consider current knowledge

limitations in making decisions about clinical concerns,

diagnoses, and referrals, and the need to use appro-

priate comparison groups.

Other recommendations include collaboration

across research groups to achieve adequate samples for

successful data analysis of siblings who develop autism.

Given the small sample sizes of families in any one

geographic area and the low recurrence risk estimates,

collaboration among research groups greatly expedite

studies examining the development of younger siblings

with autism. To facilitate productive collaboration,

research groups should attempt to use consistent

diagnostic methodologies as well as at least some

common core measures.

Another avenue for maximizing the efforts of

studying high-risk samples is to include researchers

from disciplines such as genetics, neurobiology,

developmental psychology as well as ethicists.

Although autism clinical researchers may lead these

studies, geneticists and neuroscientists could use

early phenotypic and endophenotypic data to narrow

their search for gene locations and brain mecha-

nisms. Contributions from experts in normative

development may enhance infant sibling studies by

providing guidance in developing measures suitable

for infants as well as evaluating variability in

behaviors and in specific skill development in the

first year of life. Ethicists may be necessary for

designing studies that maximize data collection while

ensuring participants and family members engaged in

such research have a favorable risk-benefit ratio

(Chen et al., 2003).

The methodological and clinical concerns that are

specific to research with samples at high-risk for the

development of autism continue to evolve, particularly

as one identifies and tests behavioral (and biological)

markers at younger ages. As research with infant sib-

lings begins to validate early manifestations of autism

empirically, and consequently early diagnostic mea-

surements improve, both research questions and design

will narrow in focus and guide the development of

more refined guidelines for such investigations.

Acknowledgments We thank the National Alliance for Autism Research (NAAR) and the National Institute of Child Health and Human Development (NICHD) for supporting the work- shop where we initially formulated the ideas outlined in this paper, and to NAAR, NICHD and the National Institute of Mental Health (NIMH) for supporting our ongoing collaborative research. In particular, we thank Dr. Andy Shih and Dr. Eric London at NAAR for their support and guidance. We also thank

J Autism Dev Disord (2007) 37:466–480 477

123

the investigators who have joined the ‘Baby Sibs’ Research Consortium since the initial inception of this paper, including Drs. Alice Carter, Leslie Carver, Kasia Chawarska, John Constantino, Karen Dobkins, Deborah Fein, Daniel Menninger, Helen Tager-Flusberg, and Nurit Yirmiya for their valuable in- sights and outstanding commitment. We also thank the scientific advisors to the consortium, including Drs. Anthony Bailey, Peter Mundy, Peter Szatmari, Steve Warren and Marshalyn Yeargin- Allsopp. Dr. Zwaigenbaum is supported by a New Investigator Fellowship from the Canadian Institute of Health Research.

References

Adrien J. L., Lenoir P., Martineau J., Perrot A., Hameury L., Larmande C., & Sauvage, D. (1993). Blind ratings of early symptoms of autism based upon family home movies. Journal of the American Academy of Child and Adolescent Psychiatry, 32, 617–626.

Allen, M. C. (1993). The high-risk infant. Pediatric Clinics of North America, 40, 479–490.

American Psychiatric Association (2000). Diagnostic and statis- tical manual of mental disorders (4th ed., text revision). Washington, DC: Author.

Bailey, A., Phillips, W., & Rutter, M., (1996). Autism: Towards an integration of clinical, genetic, neuropsychological, and neurobiological perspectives. Journal of Child Psychology and Psychiatry, 37, 89–126.

Bailey, A., Palferman, S., Heavey, L., & Le Couteur, A. (1998). Autism: The phenotype in relatives. Journal of Autism and Developmental Disorders, 28, 369–392.

Baranek, G. (1999). Autism during infancy: A retrospective video analysis of sensory-motor and social behaviors at 9– 12 months of age. Journal of Autism and Developmental Disorders, 29, 213–224.

Baron-Cohen, S., Allen, J., & Gillberg, C. (1992). Can autism be detected at 18 months? The needle, the haystack, and the CHAT. British Journal of Psychiatry, 161, 839–843.

Bolton, P. F., & Griffiths, P. D. (1997). Association of tuberous sclerosis of temporal lobes with autism and atypical autism. Lancet, 349, 392–395.

Bryson S, Zwaigenbaum L, & Roberts W. (2004). The early detection of autism in clinical practice. Pediatrics and Child Health, 4, 219–221.

Bucher, H. U., Killer, C., Ochsner, Y., Vaihinger, S., & Fauchere, J. C. (2002). Growth, developmental milestones and health problems in the first 2 years in very preterm infants com- pared with term infants: A population based study. Euro- pean Journal of Pediatrics, 161, 151–156.

Carroll, J. M., Snowling, M. J. (2004). Language and phonolog- ical skills in children at high risk of reading difficulties. Journal of Child Psychology and Psychiatry, 45, 631–640.

Chang, K. D., Steiner, H., & Ketter, T. A. (2000). Psychiatric phenomenology of child and adolescent bipolar offspring. Journal of the American Academy of Child and Adolescent Psychiatry, 39, 453–460.

Charman, T., Taylor, E., Drew, A., Cockeril, H, Brown, J., & Baird G. (2005). Outcome at 7 years of children diagnosed with autism at age 2: Predictive validity of assessments conducted at 2 and 3 years of age and pattern of symptom change over time. Journal of Child Psychology and Psychiatry, 46, 500–513.

Charman, T., Swettenham, J., Baron-Cohen, S., Cox, A., Baird, G., Drew, A. (1997). Infants with autism: An investigation of empathy, pretend play, joint attention, and imitation. Developmental Psychology, 33, 781–789.

Chawarska, K., Klin, A., & Volkmar, F. (2003). Automatic attention cueing through eye movement in 2-year-old chil- dren with autism. Child Development, 74, 1108–1122.

Chen, D. T., Miller, F. G., & Rosenstein, D. L. (2003). Ethical aspects of research into the etiology of autism. Mental Retardation and Developmental Disabilities, 9, 48–53.

Coonrod, E. E., & Stone, W. L. (2004). Early concerns of parents of children with autistic and nonautistic disorders. Infants & Young Children, 17, 258–268.

Corkum, V., & Moore, C. (1998). The origins of joint visual attention in infants. Developmental Psychology, 34, 28–38.

Courchesne, E., Carper, R., & Akshoomoff, N. (2003). Evidence of brain overgrowth in the first year of life in autism. Journal of the American Medical Association, 290, 337–344.

Crais, E., Douglas, D. D., & Campbell, C. C. (2004). The inter- section of the development of gestures and intentionality. Journal of Speech, Language and Hearing Research, 47, 678– 694.

Croen, L. A., Grether, J. K., Hoogstrate, J., & Selvin, S. (2002). The changing prevalence of autism in California. Journal of Autism and Developmental Disorders, 32, 207–215.

Crow, Y. J., & Tolmie, J. L. (1998). Recurrence risks in mental retardation. Journal of Medical Genetics, 35, 177–182.

Cuccaro, M. L., Shao, Y., Bass, M. P., Abramson, R. K., Ra- van, S. A., Wright, H. H., Wolpert, C. M., Donnelly, S. L., & Pericak-Vance, M. A. (2003). Behavioral comparisons in autistic individuals from multiplex and singleton fami- lies. Journal of Autism and Developmental Disorders, 33, 87–91.

Dahlgren, S. O., & Gillberg, C. (1989). Symptoms in the first two years of life. A preliminary population study of infantile autism. European Archives of Psychiatry And Neurological Sciences, 238, 169–174.

De Giacomo, A., Fombonne, E. (1998). Parental recognition of developmental abnormalities in autism. European Journal of Child and Adolescent Psychiatry, 7, 131–136.

Dawson, G., & Osterling, J. (1997). Early intervention in autism. In M. J. Guralnick (Ed.), The effectiveness of early inter- vention. Baltimore: Brooks.

Dawson, G., Webb, S., Schellenberg, G. D., Dager, S., Fried- man, S., Aylward, E., & Richards, T. (2002). Defining the broader phenotype of autism: Genetic, brain, and behav- ioral perspectives. Developmental Psychopathology, 14(3), 581–611.

Dawson, G., & Zanolli, K. (2003). Early intervention and brain plasticity in autism. Novartis Foundation Symposium, 251, 266–274.

Erlenmeyer-Kimling, L. (2000). Neurobehavioral deficits in off- spring of schizophrenic parents: Liability indicators and predictors of illness. American Journal of Medical Genetics, 97, 65–71.

de Bildt, A., Sytema, S., Ketelaars, C., Kraijer, D., Mulder, E., Volkmar, F., & Minderaa, R. (2004). Interrelationship between Autism Diagnostic Observation Schedule-Generic (ADOS-G), Autism Diagnostic Interview-Revised (ADI-R), and the Diagnostic and Statistical Manual of Mental Disor- ders (DSM-IV-TR) classification in children and adolescents with mental retardation. Journal of Autism and Develop- mental Disorders, 34, 129–137.

Faraone, S. V., Biederman, J., Mennin, D., Gershon, J., Tsuang, M. T. (1996). A prospective four-year follow-up study of children at risk for ADHD: Psychiatric, neuropsychological, and psychosocial outcome. Journal of the American Acad- emy of Child and Adolescent Psychiatry, 35, 1449–1459.

Filipek, P. A., Accardo, P. J., Baranek, G. T., Cook, E. H. Jr., Dawson, G., Gordon, B., Gravel, J. S., Johnson, C. P.,

478 J Autism Dev Disord (2007) 37:466–480

123

Kallen, R. J., Levy, S. E., Minshew, N. J., Ozonoff, S., Prizant, B. M., Rapin, I., Rogers, S. J., Stone, W. L., Teplin, S., Tuchman, R. F., & Volkmar, F. R. (1999). The screening and diagnosis of autistic spectrum disorders. Journal of Autism and Developmental Disorders, 29, 439–484.

Folstein, S. E., Santangelo, S. L., Gilman, S. E., Piven, J., Landa, R., Lainhart, J., Hein, J., & Wzorek, M. (1999). Predictors of cognitive test patterns in autism families. Journal of Child Psychology and Psychiatry, 40, 1117–1128.

Fombonne, E., Bolton, P., Prior, J., Jordan, H., & Rutter, M. (1997). A family study of autism: Cognitive patterns and levels in parents and siblings. Journal of Child Psychology and Psychiatry, 38, 667–683.

Geller, B., Tillman, R., Craney, J. L., & Bolhofner, K. (2004). Four-year prospective outcome and natural history of mania in children with a prepubertal and early adolescent bipolar disorder phenotype. Archives of General Psychiatry, 61, 459–467.

Heinssen, R. K., Perkins, D. O., Appelbaum, P. S., & Fenton, W. S. (2001). Informed consent in early psychosis research: National Institute of Mental Health Workshop, November 15, 2000. Schizophrenia Bulletin, 27, 571–584.

Hoshino, Y., Kumashiro, H., Yashima, Y., Tachibana, R., Watanabe, M., & Furukawa, H. (1982). Early symptoms of autistic children and its diagnostic significance. Folia Psy- chiatrica Neurologica Japan, 36, 367–374.

Howlin, P., & Moore, A. (1997). Diagnosis of autism: A survey of over 1200 patients in the UK. Autism, 1, 135–162.

Hsu, H. C., & Fogel, A. (2003). Stability and transitions in mother-infant face-to-face communication during the first 6 months: A microhistorical approach. Developmental Psy- chology, 39, 1061–1082.

Jarrold, C., & Brock, J. (2004). To match or not to match? Methdological issues in autism-related research. Journal of Autism and Developmental Disorders, 34, 81–86.

Klin, A., Chawarska, K., Rubin, E., & Volkmar, F. R. (2004). Clinical assessment of toddlers at risk of autism. In R. DelCarmen-Wiggins, & A. Carter (Eds.), Handbook of infant and toddler mental health assessment (pp. 311–336). Oxford: Oxford University Press.

Landa, R., Folstein, S., & Isaacs, C (1991). Spontaneous narra- tive discourse performance of parents of autistic individuals. Journal of Speech and Hearing Research, 34, 1339–1345.

Landa, R., Piven, J., Wzorek, M., Gayle, J., Chase, G., Folstein, S. (1992). Social language use in parents of autistic indi- viduals. Psychological Medicine, 22, 245–254.

Landa, R. (2003). Early identification of autism spectrum disor- ders. Exceptional Parent, 33, 60–63.

Landa, R., & Garrett-Mayer, L. (2006). Development in infants with autism spectrum disorders: A prospective study. Jour- nal of Child Psychology and Psychiatry, 47, 629–638.

Lord, C. (1995). Follow-up of two-year-olds referred for possible autism. Journal of Child Psychology and Psychiatry, 36, 1365–1382.

Lord, C., & McGee, J. P. (2001). Educating children with autism. Washington, DC: National Academy Press.

Lord, C., & Risi, S. (2000). Diagnosis of autism spectrum disorders in young children. In A. M. Wetherby, & B. M. Prizant (Eds.), Autism spectrum disorder: A transac- tional developmental perspective. (pp. 11–30). Baltimore: Brookes Publishing.

MacLean, J. E., Szatmari, P., Jones, M. B., Bryson, S. E., Mahoney, W. J., Bartolucci, G., Tuff, L. (1999). Familial factors influence level of functioning in pervasive develop- mental disorder. Journal of the American Academy of Child and Adolescent Psychiatry, 382, 746–753.

Maestro, S., Muratori, F., Cavallaro, M. C., Pei, F., Stern, D., Golse, B., & Palacio-Espasa, F. (2002). Attentional skills during the first 6 months of age in autism spectrum disorder. Journal of the American Academy of Child and Adolescent Psychiatry, 41, 1239–1245.

Moore, V., & Goodson, S. (2003). How well does early diagnosis of autism stand the test of time? Autism, 7, 47–63.

Moore, C., & Corkum, V. (1998). Infant gaze following based on eye direction. British Journal of Developmental Psychology, 16, 495–503.

Mundy, P., Sigman, M., & Kasari, C. (1990). A longitudinal study of joint attention and language development in autistic children. Journal of Autism and Developmental Disorders, 20, 115–128.

Murphy, M., Bolton, P. F., Pickles, A., Fombonne, E., Piven, J., & Rutter, M. (2000). Personality traits of the relatives of autistic probands. Psychological Medicine, 30, 1411–1424.

Ohta, M., Nagai, Y., Hara, H., & Sasaki, M. (1987). Parental perception of behavioral symptoms in Japanese autistic children. Journal of Autism and Developmental Disorders, 17, 549–563.

Osterling, J. A., & Dawson, G. (1994). Early recognition of children with autism: A study of first birthday home video- tapes. Journal of Autism and Developmental Disorders, 24, 247–257.

Osterling, J. A., Dawson, G., & Munson, J. A. (2002). Early recognition of 1-year-old infants with autism spectrum dis- order versus mental retardation. Development and Psycho- pathology, 14, 239–251.

Phillips, A. T., Wellman, H. M., & Spelke, E. S. (2002). Infants’ ability to connect gaze and emotional expression to inten- tional action. Cognition, 85, 53–78.

Pickles, A., Starr, E., Kazak, S., Boston, P., Papanikolaou, K., Bailey, A., Goodman, R., & Rutter, M. (2000). Variable expression of the autism broader phenotype: Findings from extended pedigrees. Journal of Child Psychology and Psy- chiatry, 41, 491–502.

Piven, J., Palmer, P., Jacobi, D., Childress, D., & Arndt, S. (1997). Broader autism phenotype: Evidence from a family history study of multiple-incidence families. American Journal of Psychiatry, 154, 185–190.

Pilowsky, T., Yirmiya, N., Doppelt, O., Gross-Tsur, V., & Shalev, R. S. (2004). Social and emotional adjustment of siblings of children with autism. Journal of Child Psychology and Psychiatry, 45, 855–865.

Pilowsky, T., Yirmiya, N., Shalev, R. S., & Gross-Tsur, V. (2003). Language abilities of siblings of children with autism. Journal of Child Psychology and Psychiatry, 44, 914–925.

Ritvo, E. R., Jorde, L. B., Mason-Brothers, A., Freeman, B. J., Pingree, C., Jones, M. B., McMahon, W. M., Pterson, P. B., Jenson, W., & Mo, A. (1989). The UCLA-University of Utah epidemiologic survey of autism: Recurrence risk esti- mates and genetic counseling. American Journal of Psychi- atry, 146, 1032–1036.

Robins, D. L., Fein, D., Barton, M. L., & Green, J. A. (2001). The Modified-Checklist for Autism in Toddlers: The modi- fied checklist for autism in toddlers: An initial study inves- tigating the early detection of autism and pervasive developmental disorders. Journal of Autism and Develop- mental Disorders, 31, 131–144.

Rogers, S. J. (1996). Brief report: Early intervention in autism. Journal of Autism and Developmental Disorders, 26, 243– 246.

Rogers, S. J. (1999). An examination of the imitation deficit in autism. In J. Nadel, & G. Butterworth (Eds.), Imitation in infancy. Cambridge studies in cognitive perceptual

J Autism Dev Disord (2007) 37:466–480 479

123

development. (pp. 254–283). New York, NY, US: Cambridge University Press.

Rogers, S. J., & DiLalla, D. L. (1990). Age of symptom onset in young children with pervasive developmental disorders. Journal of the American Academy of Child and Adolescent Psychiatry, 29, 863–872.

Rogers, S. J., Wehner, D. E., & Hagerman, R. (2001). The be- havioral phenotype in fragile X: Symptoms of autism in very young children with fragile X syndrome, idiopathic autism, and other developmental disorders. Journal of Develop- mental Behavioral Pediatrics, 22, 409–417.

Schubert, E. W., & McNeil, T. F. (2004). Prospective study of neurological abnormalities in offspring of women with psy- chosis: Birth to adulthood. American Journal of Psychiatry, 161, 1030–1037.

Siegel, B., Pliner, C., Eschler, J., & Elliott, G. R. (1988). How children with autism are diagnosed: Difficulties in identifi- cation of children with multiple developmental delays. Journal of Developmental and Behavioral Pediatrics, 9, 199– 204.

Simonoff, E. (1998). Genetic counseling in autism and pervasive developmental disorders. Journal of Autism and Develop- mental Disorders, 28, 447–456.

Smalley, S. L., McCracken, J., & Tanguay, P. (1995) Autism, affective disorders, and social phobia. American Journal of Medical Genetics, 60, 19–26.

Smith, T., Groen, A. D., & Wynn, J. W. (2000). Randomized trial of intensive intervention for children with pervasive devel- opmental disorder. American Journal of Mental Retardation, 105, 269–285.

Stone, W. L., Hoffman, E. L., Lewis, S. E., & Ousley, O. Y. (1994). Early recognition of autism: Parental reports vs. clinical observation. Archives of Pediatrics and Adolescent Medicine, 148, 174–179.

Stone, W. L., Lee, E. B., Ashford, L., Brissie, J., Hepburn, S. L., Coonrod, E. E., & Weiss, B. H. (1999). Can autism be diagnosed accurately in children under 3 years? Journal of Child Psychology and Psychiatry, 40, 219–226.

Stone, W. L., Ousley, O. Y., & Littleford, C. D. (1997). Motor imitation in young children with autism: What’s the object? Journal of Abnormal Child Psychology, 25, 475–485.

Stromland, K., Sjogreen, L., Miller, M., Gillberg, C., Wentz, E., Johansson, M., Nylen, O., Danielsson, A., Jacobsson, C., Andersson, J., & Fernell, E. (2002). Mobius sequence–a Swedish multidiscipline study. European Journal of Paedi- atric Neurology, 6, 35–45.

Swettenham, J., Baron-Cohen, S., Charman, T., Cox, A., Baird, G., Drew, A., Rees, L., & Wheelwright, S. (1998). The fre- quency and distribution of spontaneous attention shifts between social and nonsocial stimuli in autistic, typically developing, and nonautistic developmentally delayed infants. Journal of Child Psychology and Psychiatry, 39, 747–753.

Szatmari, P., Jones, M. B., Tuff. L., Bartolucci, G., Fisman, S., & Mahoney, W. (1993). Lack of cognitive impairment in first- degree relatives of children with pervasive developmental disorders. Journal of the American Academy of Child and Adolescent Psychiatry, 32, 1264–1273.

Szatmari, P., Jones, M. B., Zwaigenbaum, L., & MacLean, J. E. (1998). Genetics of autism: Overview and new directions. Journal of Autism & Developmental Disorders, 28, 351–368.

Szatmari, P., Maclean, J. E., Jones, M. B., Bryson, S. E., Zwai- genbaum, L., Bartolucci, G., Majoney, W. J., & Tuff, L.

(2000). The familial aggregation of the lesser variant in biological and nonbiological relatives of PDD probands: A family history study. Journal of Child Psychology and Psy- chiatry, 41, 579–586.

Szatmari, P., Zwaigenbaum, L., & Bryson, S. (2004). Conducting genetic epidemiology studies of autism spectrum disorders: Issues in matching. Journal of Autism and Developmental Disorders, 34, 49–57.

Thelen, E. (1981a). Kicking, rocking, and waving: Contextual analysis of stereotyped behaviour in normal infants. Animal Behaviour, 29, 3–11.

Thelen, E. (1981b). Rhythmical behavior in infancy: An etho- logical perspective. Developmental Psychology, 17, 237–257.

Tsao, F. M., Liu, H. M., & Kuhl, P. K. (2004). Speech perception in infancy predicts language development in the second year of life: A longitudinal study. Child Development, 75(4), 1067–1084.

Volkmar, F. R., Klin, A., Siegel, B., Szatmari, P., Lord, C., Campbell, M., Freeman, B. J., Cicchetti, D. V., Rutter, M., Kline W, et al. (1994). Field trial for autistic disorder in DSM-IV. American Journal of Psychiatry, 151, 1361– 1367.

Volkmar, F., Lord, C., Bailey, A., Schultz, R., & Klin, A. (2004). Autism and pervasive developmental disorders. Journal of Child Psychology and Psychiatry, 45(1), 135–170.

Werner, E., Dawson, G., Osterling, J., & Dinno, N. (2000). Brief report: Recognition of autism spectrum disorder before one year of age: A retrospective study based on home video- tapes. Journal of Autism and Developmental Disorders, 30, 157–162.

Wetherby, A., & Prizant, B. (2002). Communication and sym- bolic behavior scales developmental profile-first normed edition. Baltimore, MD: Paul H.Brookes.

Wetherby, A. M., Woods, J., Allen, L., Cleary, J., Dickinson, H., & Lord, C. (2004). Early indicators of autism spectrum disorders in the second year of life. Journal of Autism and Developmental Disorders, 34, 473–493.

Williams, G., King, J., Cunningham, M., Stephan, M., Kerr, B., & Hersh, J. H. (2001). Fetal valproate syndrome and autism: Additional evidence of an association. Developmental Medicine and Child Neurology, 43, 202–206.

Wimpory, D. C., Hobson, R. P., Williams, J. M. G., & Nash, S. (2000). Are infants with autism socially engaged? A study of recent retrospective parental reports. Journal of Autism and Developmental Disorders, 30, 525–536.

Xu, J., Zwaigenbaum. L., Szatmari, P., & Scherer S. (2004). Molecular cytogenetics of autism: Current status and future directions. Current Genomics, 5, 347–364.

Yeargin-Allsopp, M., Rice, C., Karapurkar, T., Doernberg, N., Boyle, C., & Murphy, C. (2003). Prevalence of autism in a US metropolitan area. JAMA, 289, 49–55.

Yirmiya, N., & Shaked, M. (2005). Psychiatric disorders in par- ents of children with autism: A meta-analysis. Journal of Child Psychology and Psychiatry, 46, 69–83.

Young, R. L., Brewer, N., & Pattison C. (2003). Parental iden- tification of early behavioural abnormalities in children with autistic disorder. Autism, 7, 125–143.

Zwaigenbaum L, Bryson S, Rogers, T., Roberts W, Brian J, & Szatmari P. (2005). Behavioral markers of autism in the first year of life. International Journal of Developmental Neuro- sciences, 23, 143–152.

480 J Autism Dev Disord (2007) 37:466–480

123

39

EW RESEARCH

N

8 D

The Broader Autism Phenotype in Infancy: When Does It Emerge?

Sally Ozonoff, PhD, Gregory S. Young, PhD, Ashleigh Belding, BA, Monique Hill, MA, Alesha Hill, BA, Ted Hutman, PhD, Scott Johnson, PhD, Meghan Miller, PhD, Sally J. Rogers, PhD, A.J. Schwichtenberg, PhD,

Marybeth Steinfeld, MD, Ana-Maria Iosif, PhD

Objective: This study had 3 goals, which were to examine the following: the frequency of atypical development, consistent with the broader autism phenotype, in high-risk infant siblings of children with autism spectrum disorder (ASD); the age at which atypical development is first evident; and which developmental domains are affected. Method: A prospective longitudinal design was used to compare 294 high-risk infants and 116 low-risk infants. Participants were tested at 6, 12, 18, 24, and 36 months of age. At the final visit, outcome was classified as ASD, Typical Development (TD), or Non-TD (defined as elevated Autism Diagnostic Observation Schedule [ADOS] score, low Mullen Scale scores, or both). Results: Of the high-risk group, 28% were classified as Non-TD at 36 months of age. Growth curve models demonstrated that the Non-TD group could not be distinguished from the other groups at 6 months of age, but differed significantly from the Low-Risk TD group by 12 months on multiple measures. The Non-TD group demonstrated atypical development in cognitive, motor, language, and social domains, with differences particularly prominent in the social-communication domain. Conclusions: These results demonstrate that features of atypical development, consis- tent with the broader autism phenotype, are detectable by the first birthday and affect develop- ment in multiple domains. This highlights the necessity for close developmental surveillance of infant siblings of children with ASD, along with implementation of appropriate interventions as needed. J. Am. Acad. Child Adolesc. Psychiatry, 2014;53(4):398–407. Key Words: autism spec- trum disorder, broader autism phenotype, siblings, social-communication, infancy

he broader autism phenotype (BAP) is a constellation of subclinical characteristics

T that are seen at elevated rates in family

members of children with autism spectrum dis- order (ASD).1 It is generally agreed that the BAP encompasses features related to the core diag- nostic domains of ASD, such as language delays and deficits, social difficulties, and rigidity of personality or behavior.2,3 Most previous studies have examined the BAP in parents and school-age siblings of children with ASD2,3; few have inves- tigated BAP features in infancy and toddlerhood,

This article is discussed in an editorial by Dr. John R. Pruett, Jr. on page 392.

Clinical guidance is available at the end of this article.

Supplemental material cited in this article is available online.

JOURN www.jaacap.org

ownloaded for Anonymous User (n/a) at Florida International Universit For personal use only. No other uses without perm

so it is not clear when these differences in behavior first develop and can be detected.

For questions that require precise timing of onset, prospective studies provide an optimal experimental design, because they do not rely solely on parent report, which can be subject to recall errors and other biases. In the past decade, prospective studies of high-risk infants have proliferated. Most commonly, the individuals at increased risk for ASD studied thus far are later- born siblings of children with ASD. Such infant sibling study designs often compare high-risk samples to low-risk infants with no family his- tory of ASD. Although several dozen such studies have been published, most focus on describing the early development and predictive early risk signs of infants who ultimately develop ASD.4,5

Other infant sibling studies have reported differ- ences between high- and low-risk groups in a variety of domains, including eye contact, joint

AL OF THE AMERICAN ACADEMY OF CHILD & ADOLESCENT PSYCHIATRY VOLUME 53 NUMBER 4 APRIL 2014

y – Florida state consortium from ClinicalKey.com by Elsevier on January 09, 2019. ission. Copyright ©2019. Elsevier Inc. All rights reserved.

BROADER AUTISM PHENOTYPE IN INFANCY

Down

attention, and nonverbal reasoning, but did not follow the infants long enough to know whether these differences were early signs of ASD or might instead index other types of atypical out- comes, including the BAP.6-9

Only a few infant sibling studies have specif- ically focused on describing early signs of the BAP.10-15 These investigations follow participants until age 3 years, determine which children develop ASD, and remove them from the larger high-risk group before analyses (because, by definition, the BAP and ASD are mutually exclu- sive). Several studies, most involving small sam- ples, have found significant differences between high-risk non-ASD groups and low-risk control individuals early in life, on tasks of response to joint attention at 14 months (n ¼ 8)10 and social referencing at 18 months (n ¼ 30),11 as well as on parent report measures of temperament as early as 7 months (n ¼ 12).12 Early differences in parent-reported temperament in high-risk siblings without ASD have also been reported in a much larger sample at 24 months of age (n ¼ 104).13 In a comprehensive study examining multiple domains of development, 40 high-risk siblings without ASD outcomes were, as a group, below average in expressive and receptive language, overall IQ, adaptive behavior, and social commu- nication skills at 18 to 27 months.14 In addition, parents reported social impairments on a ques- tionnaire by 13 months of age. A recent large study followed 170 high-risk children, none of whom were diagnosed as having ASD at age 3 years.15 A cluster analysis identified a subgroup (19% of the high-risk sample) that had elevated scores on the Autism Observation Scale for Infants at 12 months of age. At age 3, this cluster de- monstrated lower scores than low-risk controls on independent social-communication and cognitive measures. Taken together, these and other studies strongly suggest that behavioral and develop- mental features consistent with the BAP emerge early in life.

Most published sibling studies have been cross- sectional and/or focused on whether group dif- ferences are evident at a single age. Only 1 study thus far has examined longitudinal trajectories of development, following a cohort of 37 high- risk children from 4 months to 7 years of age.16

At 7 years, the researchers split their high-risk group into 2 subgroups, 1 group with BAP fea- tures (40%) and 1 group without, and then examined their cognitive and language trajec- tories in the preschool years (4–54 months) using

JOURNAL OF THE AMERICAN ACADEMY OF CHILD & ADOLESCENT PSYCHIATR VOLUME 53 NUMBER 4 APRIL 2014 loaded for Anonymous User (n/a) at Florida International University – Florida state

For personal use only. No other uses without permission. Copyrig

growth curve analysis. They found that language scores were different for the BAP group as early as 14 months, but that cognitive scores did not differentiate the group from the low-risk controls at any age. The current study took a similar approach, examining development longitudinally from 6 to 36 months in high- and low-risk infants (n ¼ 294 and n ¼ 116, respectively) and looking for the earliest inflection point at which the tra- jectories diverge from one of typical to atypical development. The current study is the largest sample to date that examines BAP features longi- tudinally. We focus on several domains of early development: social-communication, language, nonverbal cognitive, and fine motor abilities.

The studies reviewed above have taken 1 of 2 approaches when studying the BAP. Some have studied all children in the high-risk group, after excluding those with an ASD outcome, looking for differences from low-risk infants.14 Others have classified an “atypical” outcome group, us- ing varying criteria at varying outcome ages, and then examined whether this “atypical” subgroup differs from low-risk controls at earlier ages than when the groups were defined.10,12,16 This latter approach is the one used in the current study. It is clear that there is substantial heterogeneity within the high-risk group; virtually all previous studies find that atypical development or BAP-like fea- tures are present in only a subset of siblings of children with ASD.2,3,17 Therefore, studying all high-risk siblings without ASD outcomes risks the possibility of obscuring potential differences that may be evident in a subgroup. Using a de- finition similar to other recent investigations,10,12

we identified a group of high-risk children with non-typical developmental outcomes at 36 months of age. We then used growth curve analysis to examine when non-typical development could first be detected. We studied multiple areas of development, extending more broadly than the BAP (e.g., social-communication, but also cogni- tion and motor skills), to examine in which do- mains non-typical development was evident.

METHOD Participants The sample reported in this article was drawn from a larger longitudinal study of infant siblings of children with ASD (High-Risk group) or children with typical development (Low-Risk group), recruited at 2 sites (University of California, Davis [UC Davis] and Uni- versity of California, Los Angeles [UCLA]) during 2 phases of grant funding (2003–2008 and 2008–2013).

Y

www.jaacap.org 399 consortium from ClinicalKey.com by Elsevier on January 09, 2019. ht ©2019. Elsevier Inc. All rights reserved.

OZONOFF et al.

The sole inclusion criterion for the High-Risk group was status as a younger sibling of a child with ASD. Diagnosis of the affected older sibling was confirmed by meeting ASD criteria on both the Autism Diagnostic Observation Schedule (ADOS) and the Social Communication Questionnaire (SCQ).18,19 Exclusion criteria for the High-Risk group included birth before 36 weeks of gestation and a known genetic disorder (e.g., fragile X syndrome) in the older affected sibling. The primary inclusion criterion for the Low-Risk group was status as a younger sibling of a child (or children) with typical development. Low-risk status of all older siblings was confirmed by an intake screening ques- tionnaire and scores below the ASD range on the SCQ. Exclusion criteria for the Low-Risk group were as fol- lows: birth before 36 weeks of gestation; develop- mental, learning, or medical conditions in any older sibling; and ASD in any first-, second-, or third-degree relative. All participants with complete data at the 36- month outcome visit were included.

Participants were enrolled before 18 months of age (age at enrollment: mean ¼ 6.7 months, SD ¼ 5.2 months; 76% were enrolled by 9 months or earlier). Depending on age of study entry, data were collected at up to 5 ages: 6, 12, 18, 24, and 36 months. At the 36-month visit, participants were classified into 1 of 3 algorithmically defined outcome groups: ASD, Typical Development (TD), and Non-Typical Development (Non-TD). Table 1 provides algorithmic group defini- tions, which were developed by the Baby Siblings Research Consortium, a network of researchers study- ing very young children at risk for ASD (Chawarska et al., unpublished data, November 2013).

Given this article’s focus on the BAP, which by definition is a characteristic of family members of a child with ASD, the small groups of Low-Risk partici- pants with ASD (n ¼ 4) or Non-TD (n ¼ 27) outcomes were not included in analyses. The final sample with complete 36-month data included in the study were 51 participants classified with ASD (17.4% of the High- Risk group; n ¼ 8 females), 83 with Non-TD outcomes (28.2% of the High-Risk group; n ¼ 32 females), and

TABLE 1 Algorithmic Group Outcome Definitions

Outcome Classification

ASD At o Mee

Typical Development Doe No No ADO

Non-Typical Development Doe Two One ADO

Note: ADOS ¼ Autism Diagnostic Observation Schedule; ASD ¼ autism spectru specified.

JOURN 400 www.jaacap.org

Downloaded for Anonymous User (n/a) at Florida International Universit For personal use only. No other uses without perm

276 with TD outcomes, who were further stratified into High-Risk TD (n ¼ 160; n ¼ 90 females) and Low-Risk TD (n ¼ 116; n ¼ 53 females). Of the 83 participants in the Non-TD sample, 66 were classified into this group because of elevated ADOS alone, 9 were classified because of low Mullen Scale scores alone (8 had at least 1 Mullen Scale score that was �2 SD below the mean, and 1 had �2 Mullen Scale scores that were �1.5 SD below mean), and 8 were classified as Non-TD because of both elevated ADOS and low Mullen Scale scores (7 had at least 1 Mullen Scale score that was �2 SD below the mean, and 1 had �2 Mullen Scale scores that were �1.5 SD below the mean).

Measures The study was conducted under the approval of both sites’ institutional review boards. Infants were assessed by examiners who were unaware of group membership.

Autism Diagnostic Observation Schedule.18 This is a semi-structured, standardized interaction and obser- vation tool that measures symptoms of autism. It has 2 empirically derived cutoffs, 1 for ASD and 1 for Autistic Disorder. Because data collection occurred before the publication of newer ADOS algorithms, the CommunicationþSocial Total algorithm score was used.19 Psychometric studies report high interrater reliability and agreement in diagnostic classification (autism versus non-ASD). The ADOS was used to confirm older sibling diagnosis and to determine infant outcome at 36 months of age (Table 1).

Mullen Scales of Early Learning.20 This is a stan- dardized developmental test for children from birth to 68 months. Four subscales were administered: Fine Motor, Visual Reception, Expressive Language, and Receptive Language. Scores are expressed in raw score points, which can also be converted to T-scores and age equivalents using published normative data. An overall score, the Early Learning Composite, is also obtained. The Mullen Scale subscales have excellent internal consistency (median ¼ 0.91) and test–retest reliability (median ¼ 0.84). This test was used to

Criteria

r above the ASD cutoff of the ADOS and ts DSM-IV-TR criteria for Autistic Disorder or PDD-NOS s not meet criteria for ASD classification and more than 1 Mullen Scale subtest �1.5 SD below mean and Mullen Scale subtest �2 SD below mean and S >3 points below ASD cutoff s not meet criteria for ASD classification and or more Mullen Scale subtests �1.5 SD below mean and/or or more Mullen subtests �2 SD below mean and/or S �3 points below ASD cutoff m disorder; PDD-NOS ¼ pervasive developmental disorder not otherwise

AL OF THE AMERICAN ACADEMY OF CHILD & ADOLESCENT PSYCHIATRY VOLUME 53 NUMBER 4 APRIL 2014

y – Florida state consortium from ClinicalKey.com by Elsevier on January 09, 2019. ission. Copyright ©2019. Elsevier Inc. All rights reserved.

BROADER AUTISM PHENOTYPE IN INFANCY

Down

measure cognitive functioning at each visit and to determine outcome status at 36 months. Ongoing administration and scoring fidelity procedures were implemented to ensure that there were minimal cross- examiner and cross-site differences.

Examiner-Rated Social Engagement. At the end of the session, examiners rated 3 behaviors using a 3-point scale (1 ¼ rare, 2 ¼ occasional, 3 ¼ frequent), as fol- lows: frequency of eye contact; frequency of shared affect; and overall social responsiveness. These 3 scores were summed to create a social engagement composite score (ranging from 3 to 9). In a previous study, this measure was able to distinguish infants with typical versus atypical development by 12 months of age.21

Clinical Best Estimate Outcome Classification. At the end of the 36-month visit, examiners classified each child into 1 of 6 Clinical Best Estimate (CBE) categories: ASD, BAP, Behavior Problems, Global Developmental Delay, Speech–Language Problems, or Typical Devel- opment. In contrast to the algorithmic groups (ASD, TD, Non-TD) that were empirically determined for the current analyses, the CBE classifications were clinically defined. Children classified with ASD met DSM-IV-TR criteria for autistic disorder or pervasive developmental disorder not otherwise specified (PDD-NOS). Children classified as BAP displayed social-communication diffi- culties that were judged to be below the ASD threshold. Children classified as having ADHD concerns displayed high activity level, poor attention, or disruptive behavior, beyond what would be expected for devel- opmental level. Children classified clinically with Global Developmental Delay had low scores across multiple cognitive and motor domains. Children classified as having Speech–Language Problems displayed immature speech patterns or low language levels in isolation (no accompanying social or cognitive difficulties). All other participants were classified as having Typical Development.

Statistical Analysis Mixed-effects linear models were used to estimate patterns of change in Mullen Scale raw scores and to test whether group was related to the initial level or rate of change in these variables.22 All core models included fixed effects for group (ASD, Non-TD, High- Risk TD, and Low-Risk TD), the linear effect of age (centered at 6 months), and the interaction between group and age. To account for the correlated nature of the data, the core models included 2 random effects for child-specific intercepts and slopes. Additional fixed terms (for the quadratic effect of age, the interaction of the quadratic effect of age with group, gender, phase, site, etc.) were also added to the core model and tested. These terms were retained in the models only if they were significant. For the models with a significant quadratic effect of age, we also included random effects for the quadratic age. For some of those models, there was little variability left in the intercepts, so only the

JOURNAL OF THE AMERICAN ACADEMY OF CHILD & ADOLESCENT PSYCHIATR VOLUME 53 NUMBER 4 APRIL 2014 loaded for Anonymous User (n/a) at Florida International University – Florida state

For personal use only. No other uses without permission. Copyrig

child-specific slopes were retained. A similar modeling strategy was used to analyze the Examiner-Rated So- cial Engagement composite scores (with age centered at 6 months) and ADOS social-communication scores (with age centered at 18 months). Further details on the mixed-effects models are presented in Supplement 1, available online.

All tests were 2-sided, with a ¼ 0.05. Residual ana- lyses and graphical diagnostics determined that the model assumptions were adequately met. Analyses were implemented using PROC MIXED in SAS Version 9.3.23

RESULTS Table 2, Table S1 (available online), and Figure 1 summarize the results of the mixed-effects models for Mullen Scale raw scores. At baseline (6 months of age), all 4 groups had comparable values on all 4 scales. The Low-Risk TD group demon- strated a sharp increase in raw scores with age on all Mullen Scales. The High-Risk TD group had significantly slower growth over time than the Low-Risk TD group on the Expressive and Receptive Language scales, but not on the Visual Reception and Fine Motor scales. At 36 months, the 2 TD groups had comparable Visual Recep- tion and Fine Motor scores, but the High-Risk TD group showed significantly lower levels of Expressive Language (by 1.1 points) and Recep- tive Language (by 1.7 points). The ASD group showed a significantly slower rate of change than both TD groups on all 4 scales and was signifi- cantly different from both groups by 12 months of age. Of primary interest for this article, the Non-TD group’s performance was intermediate between the ASD and both TD groups. The Non- TD group had lower rates of growth than both TD groups, resulting in significant differences from them by 12 months of age on all scales except Fine Motor. The differences from both TD groups were modest at 12 months (differences from Non-TD ranged from 0.3 to 1.5 points across scales for Low-Risk TD and 0.2 to 1.1 points for High-Risk TD) but amplified over time (at 36 months, differences from Non-TD ranged from 3.4 to 4.7 points in Low-Risk TD and from 2.8 to 3.5 points in High-Risk TD).

At 6 months of age, all 4 groups had similar Examiner-Rated Social Engagement composite scores (Table 2 and Figure 2). The 2 TD groups exhibited significant growth over time, and the ASD group showed a sharp decrease in scores with age. The Non-TD group had a flat trajectory, with significant differences from the Low-Risk

Y

www.jaacap.org 401 consortium from ClinicalKey.com by Elsevier on January 09, 2019. ht ©2019. Elsevier Inc. All rights reserved.

TABLE 2 Parameter Estimates (SE) for Mixed-Effects Regression Models Predicting Mullen Scale Raw Scores, Examiner-Rated Social Engagement, and Autism Diagnostic Observation Schedule (ADOS) Social-Communication Scores

Model Term

Mullen Scale Examiner-Rated

Social Engagement ADOS

Social-CommunicationEL RL VR FM

Estimated trajectory for Low-Risk TD group Baseline 6.24 (0.15)*** 6.69 (0.22)*** 9.52 (0.18)*** 9.28 (0.18)*** 7.87 (0.18)*** 2.42 (0.26)*** Linear age effect (year) 11.42 (0.18)*** 14.81 (0.55) *** 12.84 (0.30) *** 12.98 (0.39)*** 1.13 (0.34)** e0.48 (0.82) Quadratic age effect (year) months) — e1.49 (0.21)*** e0.28 (0.10)** e1.26 (0.16)*** e0.29 (0.11)* e0.08 (0.49)

Estimated difference between ASD and Low-Risk TD group Baseline e0.30 (0.29) 0.82 (0.42) 0.35 (0.32) e0.35 (0.35) e0.23 (0.37) 8.46 (0.70)*** Linear age effect (year) e3.60 (0.34)*** e9.32 (1.02)*** e3.11 (0.38)*** e0.63 (0.73) e2.12 (0.59)*** e6.83 (2.51)** Quadratic age effect (year) months) — 2.09 (0.39)*** — e0.67 (0.30)* e0.46 (0.20)* 5.94 (1.56)***

Estimated difference between Non-TD and Low-Risk TD groups Baseline 0.31 (0.24) 0.04 (0.35) 0.03 (0.27) e0.13 (0.29) e0.37 (0.31) 2.24 (0.39)*** Linear age effect (year) e1.94 (0.28)*** e3.43 (.86)*** e1.42 (.32)*** e0.17 (0.61) e0.53 (0.51) e1.33 (1.23) Quadratic age effect (year) months) — .61 (0.33) — e0.45 (0.25) e0.13 (0.17) 1.75 (0.73)*

Estimated difference between High-Risk TD and Low-Risk TD groups Baseline 0.32 (0.21) 0.49 (0.31) 0.15 (0.23) 0.15 (0.26) e0.26 (0.29) 1.10 (0.34)** Linear age effect (year) e0.56 (0.23)* e2.03 (0.73)** e0.22 (0.27) e0.67 (0.52) e0.38 (0.47) e2.02 (1.05) Quadratic age effect (year) months) — 0.46 (0.27) — .15 (0.21) e0.17 (0.15) 1.01 (0.62)

Note: Baseline is 18 months for ADOS and 6 months for all other variables. ASD ¼ autism spectrum disorder; EL ¼ Expressive Language; FM ¼ Fine Motor; RL ¼ Receptive Language; SE ¼ standard error; TD ¼ typically developing; VR ¼ Visual Reception. *p < .05; **p < .01; ***p < .001.

JO U R N A L O F TH

E A

M E RIC

A N A

C A D EM

Y O F C

H ILD

& A

D O LES

C EN

T P S Y C H IA

TR Y

4 0 2

w w w .jaacap.org

V O LU

M E 5 3

N U M BER

4 A PRIL

2 0 1 4

O Z O N O FF

e t a l.

D ow

nloaded for A nonym

ous U ser (n/a) at F

lorida International U niversity – F

lorida state consortium from

C linicalK

ey.com by E

lsevier on January 09, 2019. F

or personal use only. N o other uses w

ithout perm ission. C

opyright © 2019. E

lsevier Inc. A ll rights reserved.

FIGURE 1 Estimated trajectories for Mullen Scales. ASD ¼ autism spectrum disorder; TD ¼ typically developing.

BROADER AUTISM PHENOTYPE IN INFANCY

Down

TD group evident starting at 12 months and resulting in 36-month scores that were signi- ficantly lower over time than both TD groups (by w1 point) but higher than the ASD group (by w2 points).

At 18 months (the first visit in which the ADOS was administered), there were significant group differences on the social-communication algo- rithm score, with the Low-Risk TD group dem- onstrating lower scores than the High-Risk TD (by 1 point), Non-TD (by 2 points), and ASD (by 8 points) groups (Table 2 and Figure 2). The Low- Risk TD group demonstrated a stable trajectory over time, whereas the High-Risk TD group exhibited a slight decrease over time. The Non- TD group showed a significant quadratic effect of age. At 36 months, the 2 TD groups had comparable scores (1.5 and 1.9, respectively), whereas the Non-TD and ASD group showed significantly higher scores (estimated values 5.7 and 13.1, respectively). Again, as with the Mullen Scale, the scores and longitudinal trajectories of the Non-TD group fell intermediate between the TD and ASD groups.

Table 3 depicts the correspondence between the empirically derived algorithmic classifications (ASD, TD, Non-TD) and clinical judgment (CBE

JOURNAL OF THE AMERICAN ACADEMY OF CHILD & ADOLESCENT PSYCHIATR VOLUME 53 NUMBER 4 APRIL 2014 loaded for Anonymous User (n/a) at Florida International University – Florida state

For personal use only. No other uses without permission. Copyrig

outcome classification) at 36 months. The perfect correspondence between the 2 classifications for the ASD group is secondary to the algorithmic definition, which requires a clinical diagnosis of ASD. The Non-TD group had a significantly higher rate of classifications of BAP, ADHD concerns, Global Developmental Delay, and Speech–Language Problems and significantly lower rate of Typical Development classifications than both the High-Risk and Low-Risk TD groups (Fisher’s exact test, p < .001). The most common clinical classification for the Non-TD group was BAP, with more than one-third of the sample falling in this category. Three Non- TD participants received a CBE rating of ASD but did not meet the algorithmic criteria (e.g., did not have an ADOS score over the ASD cutoff), resulting in their classification as Non- TD. Interestingly, almost 40% of the Non-TD group was judged by examiners to have a CBE outcome of typical development, despite the elevated ADOS scores or lowered Mullen Scale scores that classified them empirically in the Non- TD group.

In secondary analyses, we added to the core models and tested terms for gender, site, funding phase, and, for those models with significant

Y

www.jaacap.org 403 consortium from ClinicalKey.com by Elsevier on January 09, 2019. ht ©2019. Elsevier Inc. All rights reserved.

FIGURE 2 Estimated trajectories for Examiner-Rated Social Engagement composite and Autism Diagnostic Observation Schedule (ADOS) social-communication algorithm score. ASD ¼ autism spectrum disorder; TD ¼ typically developing.

OZONOFF et al.

gender effects, the interactions between gender and group and between gender and age. There was no phase effect, and site was significant only in the model predicting receptive language (the UCLA sample scored 0.4 points higher than the UC Davis sample, p < .05, but the difference was so small that it is unlikely to be clinically mean- ingful). Gender was a significant predictor for all Mullen Scales except Receptive Language, with girls demonstrating slightly higher Visual Reception scores than boys (0.5 point, p < .05). For Expressive Language and Fine Motor, there was a significant gender-by-group interaction, driven by girls in the ASD group, who scored lower than boys on these scales, whereas girls in the other 3 groups scored w0.5 point higher than boys on the same scales. There were no gender, phase, or site effects in the model pre- dicting ADOS social-communication score. For the Examiner-Rated Social Engagement compos- ite, there was a significant gender-by-group in- teraction, driven again by the girls in the ASD group, who scored 2 points lower than the boys (p < .001), whereas girls in the other 3 groups scored similarly to boys of the same group. For this variable, there was a phase effect, with phase 2 children scoring about 0.2 points higher than phase 1 children (p ¼ .03). The interaction be- tween gender and age was not significant in any of the models considered.

DISCUSSION This study focused on developmental aspects of the BAP, exploring the frequency of non-typical development in high-risk infant siblings, the

JOURN 404 www.jaacap.org

Downloaded for Anonymous User (n/a) at Florida International Universit For personal use only. No other uses without perm

age at which atypical development was first evident, and which developmental domains were affected. We found that 28% of the high-risk cohort demonstrated atypical development (not including ASD) at 36 months of age, as defined by elevated ADOS scores (within 3 points of the ASD cutoff), low Mullen Scale scores, or both. Working backwards from this age, we used growth curve models to determine when these differences in development could first be detected. On the Mullen Scales of Early Learning and the examiner ratings of social engagement, the Non-TD group was not distinguishable from any other group at 6 months, but differed significantly from the Low-Risk TD group by 12 months of age, devi- ating from typical development as early as the group with ASD. At 18 months, the earliest age at which the ADOS was administered, the Non-TD group was already obtaining significantly higher scores than the Low-Risk TD group.

The aspects of atypical development that dis- tinguished the Non-TD group from the Low-Risk group occurred in all domains assessed in this study (cognition, motor, language, and social development) but were most prominent in the social-communication domain. Of the Non-TD group, 90% demonstrated social-communication difficulties (as defined by an ADOS score within 3 points of the ASD cutoff), including reduced eye contact, infrequent social initiations with unfamiliar persons, repetitive vocalizations, and delayed onset of gestures, speech, and play. Iso- lated language and cognitive delays (e.g., low Mullen Scale scores alone) were relatively rare, seen in only 10% of the Non-TD group, as pre- vious studies have also found.24 When such

AL OF THE AMERICAN ACADEMY OF CHILD & ADOLESCENT PSYCHIATRY VOLUME 53 NUMBER 4 APRIL 2014

y – Florida state consortium from ClinicalKey.com by Elsevier on January 09, 2019. ission. Copyright ©2019. Elsevier Inc. All rights reserved.

TABLE 3 Clinical Best Estimate Classifications at 36 Months by Algorithmic Group

Clinical Best Estimate, n (%) ASD

(n ¼ 51) Non-TD (n ¼ 83)

High-Risk TD (n ¼ 160)

Low-Risk TD (n ¼ 116)

Autism spectrum disorder 51 (100) 3 (4) 0 (0) 0 (0) Broader autism phenotype 0 (0) 29 (35) 10 (6) 0 (0) ADHD concerns 0 (0) 8 (10) 7 (4) 2 (2) Global developmental delay 0 (0) 5 (6) 2 (1) 0 (0) Speechelanguage problems 0 (0) 6 (7) 14 (9) 3 (3) Typical development 0 (0) 32 (39) 127 (79) 111 (96)

Note: ASD ¼ autism spectrum disorder; TD ¼ typically developing.

BROADER AUTISM PHENOTYPE IN INFANCY

Down

delays were evident, they occurred in combina- tion with elevated ADOS scores. Thus, the vast majority of the Non-TD group demonstrated the kinds of social-communication features that have been previously described in older siblings as consistent with the BAP. Interestingly, almost 40% of the Non-TD group was given a CBE rating of typical development by examiners, despite such elevated ADOS scores. We plan to further examine this subgroup to better understand what may lead to a clinical judgment of typicality, despite non-typical scores. An item analysis of the ADOS, for example, may reveal that high scores on certain items are not considered particularly concerning by clinicians, leading to a CBE of typical development, whereas high scores on other items (e.g., eye contact) are judged as consistent with the BAP.

One of the primary gaps in the literature motivating this research was the paucity of studies of BAP-like phenomena in very young siblings, with most previous investigations con- ducted on school-age siblings and parents. This results in a need to “translate” the types of defi- cits seen at older ages, and instruments used to measure them, into those appropriate for earlier stages of development. Some of this translation was straightforward, when the same instrument used with older siblings and parents could also be used with this young age group (e.g., the ADOS; the comprehensive review by Sucksmith et al.3 includes a list of previous studies and measures used). It was not clear at the start of this study whether the Mullen Scales would adequately index any cognitive delays that might be apparent. The findings here demonstrate that general developmental delays can occasionally be seen in very young siblings and that the Mullen Scales can detect them.

In future follow-up studies, as our sample reaches school-age, we plan to examine what proportion of the High-Risk children meet

JOURNAL OF THE AMERICAN ACADEMY OF CHILD & ADOLESCENT PSYCHIATR VOLUME 53 NUMBER 4 APRIL 2014 loaded for Anonymous User (n/a) at Florida International University – Florida state

For personal use only. No other uses without permission. Copyrig

definitions of the BAP used previously in older samples.1 Many definitions include behavioral features not seen in infancy or measured by our tasks, such as peer problems, pragmatic language difficulties, rigid inflexible behavior, anxiety, and depression. It is possible that the rate of atypical development will increase over time, and that some children in the High-Risk TD group who did not show atypicalities at 36 months or did not meet cutoffs for the Non-TD definition may be identified with a BAP-like phenotype as they are followed up longitudinally into the school years. Previous longitudinal studies have, in fact, reported a significant increase in the number of high-risk siblings identified with BAP-related difficulties at age 7 years compared to the pre- school years.25-27

The results reported here are largely consistent with a recently published study that used a dif- ferent type of prospective design.28 This research team analyzed the Avon Longitudinal Study of Parents and Children (ALSPAC) cohort, a very large, community-ascertained general population sample that followed children from before birth to age 11 years, obtaining parent reports of de- velopment (including a measure of ASD traits) at multiple ages. Bolton et al. found that parents reported differences in development within the first year of life that not only predicted later di- agnoses of ASD, but also a wider, subthreshold range of autistic-like behaviors potentially con- sistent with the BAP.28

A question that often arises is whether siblings like those in the Non-TD group, who have delays that are sub-threshold to ASD, should receive early intervention services or whether their de- lays will lessen over time without treatment. There are not, as yet, any well-controlled inter- vention studies that can help to answer this question, so we must turn to other sources. One answer to the question comes from the law involving early intervention services for children

Y

www.jaacap.org 405 consortium from ClinicalKey.com by Elsevier on January 09, 2019. ht ©2019. Elsevier Inc. All rights reserved.

Clinical Guidance

� Close to 50% of younger siblings of children with ASD develop in an atypical fashion. In the current study, 17% developed ASD, and another 28% showed delays or deficits in other areas of development or behavior.

� Differences in development are detectable using standardized assessment instruments by 12 months of age in many children.

� The most common development differences seen in younger siblings of children with ASD are delays in social-communication development (including reduced eye contact, extreme shyness with unfamil- iar persons, and delayed onset of gestures and speech). Some younger siblings also show delays in cognitive and motor abilities, as well as attentional and behavioral problems.

� Close developmental surveillance of infant siblings of children with ASD is necessary, along with implementation of appropriate interventions as needed.

OZONOFF et al.

under 3 years of age, the Individuals with Dis- abilities Education Act (IDEA), Part C, which states that young children with delays and those who are at high risk for developmental delays are entitled to assessments and intervention services. Thus, good clinical practice suggests that when children are showing atypical development, they and their families should be provided with in- formation about the child’s difficulties, clinical reports when practical, and referrals to local Part C service providers. The second response to this question about early intervention for BAP-like features comes from 2 long-term longitudinal studies of infant siblings, both of which demon- strated that children with early lagging trajec- tories continue to experience challenges after the preschool period and do not “catch up” to typi- cally developing peers.16,28

Which types of intervention should be pro- vided to this wide-ranging group of children? Certainly, no single approach or modality can be expected to fit a group whose difficulties range from severe hyperactivity, to mild-to-moderate intellectual impairment, to subthreshold symp- toms of ASD. Intervention approaches need to be chosen based on each child’s profile of strengths and weaknesses and each family’s goals and priorities. However, there are a range of choices available to early intervention professionals from a range of disciplines. Empirically supported, manualized, parent-implemented interventions for toddlers and preschoolers with behavior disorders, general delayed development, social- communicative symptoms related to autism, and difficulties with expressive communication are represented in the literature, and many of these can be carried out by professionals from a variety of disciplines.29-31

We will continue to follow our sample as they reach school age, to examine whether develop- mental difficulties identified at age 3 years per- sist, and whether new difficulties (e.g., learning disorders, anxiety) emerge over time. It is critical to better understand the long-term functional consequences of the early developmental pat- terns identified in the current study. The ulti- mate goal of this program of research is to determine whether monitoring and identifica- tion in the preschool years could be used to

JOURN 406 www.jaacap.org

Downloaded for Anonymous User (n/a) at Florida International Universit For personal use only. No other uses without perm

provide appropriate interventions that would reduce the number of high-risk siblings who display later difficulties. &

AL

y – is

Accepted December 24, 2013.

Drs. Ozonoff, Young, Miller, Rogers, Steinfeld, and Iosif, and Ms. Belding, Ms. M. Hill, and Ms. A. Hill are with the University of CaliforniaeDavis. Drs. Hutman and Johnson are with the University of CaliforniaeLos Angeles. Dr. Schwichtenberg is with Purdue University.

This study was supported by the National Institute of Mental Health grants R01 MH0638398 (S.O.) and U54 MH068172 (Marian Sigman, PhD [deceased]).

Drs. Iosif and Young served as the statistical experts for this research.

Editorial support for the preparation of this article was provided by Diane Larzelere, BA, University of California-Davis. The authors thank the children and families who participated in this longitudinal study.

Disclosure: Drs. Ozonoff, Young, Hutman, Johnson, Miller, Rogers, Schwichtenberg, Steinfeld, and Iosif, and Ms. Belding, Ms. M. Hill and Ms. A. Hill report no biomedical financial interests or potential conflicts of interest.

Correspondence to Sally Ozonoff, PhD, MIND Institute, University of California Davis Health System, 2825 50th Street, Sacramento CA 95817; e-mail: [email protected]

0890-8567/$36.00/ª2014 American Academy of Child and Adolescent Psychiatry

http://dx.doi.org/10.1016/j.jaac.2013.12.020

REFERENCES

1. Bolton P, Macdonald H, Pickles A, et al. A case-control family history

study of autism. J Child Psychol Psychiatry. 1994;35:877-900. 2. Bailey A, Palferman S, Heavey L, Le Couteur A. Autism: the

phenotype in relatives. J Autism Dev Disorder. 1998;28:369-392.

3. Sucksmith E, Roth I, Hoekstra RA. Autistic traits below the clinical threshold: re-examining the broader autism phenotype in the 21st century. Neuropsychol Rev. 2011;21: 360-389.

OF THE AMERICAN ACADEMY OF CHILD & ADOLESCENT PSYCHIATRY VOLUME 53 NUMBER 4 APRIL 2014

Florida state consortium from ClinicalKey.com by Elsevier on January 09, 2019. sion. Copyright ©2019. Elsevier Inc. All rights reserved.

BROADER AUTISM PHENOTYPE IN INFANCY

Down

4. Landa R, Garrett-Mayer E. Development in infants with autism spectrum disorders: a prospective study. J Child Psychol Psychi- atry. 2006;47:629-638.

5. Zwaigenbaum L, Bryson S, Rogers T, Roberts W, Brian J, Szatmari P. Behavioral manifestations of autism in the first year of life. Int J Dev Neurosci. 2005;23:143-152.

6. Bedford R, Elsabbagh M, Gliga T, et al. Precursors to social and communication difficulties in infants at-risk for autism: gaze following and attentional engagement. J Autism Dev Disord. 2012; 42:2208-2218.

7. Bhat AN, Galloway JC, Landa RJ. Social and non-social visual attention patterns and associative learning in infants at risk for autism. J Child Psychol Psychiatry. 2010;51:989-997.

8. Merin N, Young GS, Ozonoff S, Rogers SJ. Visual fixation patterns during reciprocal social interaction distinguish a subgroup of 6-month-old infants at-risk for autism from comparison infants. J Autism Dev Disord. 2007;37:108-121.

9. Stone WL, McMahon CR, Yoder PJ, Walden TA. Early social- communicative and cognitive development of younger siblings of children with autism spectrum disorders. Arch Pediatr Adolesc Med. 2007;161:384-390.

10. Sullivan M, Finelli J, Marvin A, Garrett-Mayer E, Bauman M, Landa R. Response to joint attention in toddlers at risk for autism spectrum disorder: a prospective study. J Autism Dev Disord. 2007;37:37-48.

11. Cornew L, Dobkins KR, Akshoomoff N, McCleery JP, Carver LJ. Atypical social referencing in infant siblings of children with autism spectrum disorders. J Autism Dev Disord. 2012;42:2611-2621.

12. Clifford SM, Hudry K, Elsabbagh M, Charman T, Johnson MH. Temperament in the first 2 years of life in infants at high-risk for autism spectrum disorders. J Autism Dev Disord. 2013;43:673-686.

13. Garon N, Bryson SE, Zwaigenbaum L, et al. Temperament and its relationship to autistic symptoms in a high-risk infant sib cohort. J Abnorm Child Psychol. 2009;37:59-78.

14. Toth K, Dawson G, Meltzoff AN, Greenson J, Fein D. Early social, imitation, play, and language abilities of young non-autistic siblings of children with autism. J Autism Dev Disord. 2007;37:145-157.

15. Georgiades S, Szatmari P, Zwaigenbaum L, et al. A prospective study of autistic-like traits in unaffected siblings of probands with autism spectrum disorder. JAMA Psychiatry. 2013;70:42-48.

16. Gamliel I, Yirmiya N, Jaffe DH, Manor O, Sigman M. Develop- mental trajectories in siblings of children with autism: cognition and language from 4 months to 7 years. J Autism Dev Disord. 2009;39:1131-1144.

JOURNAL OF THE AMERICAN ACADEMY OF CHILD & ADOLESCENT PSYCHIATR VOLUME 53 NUMBER 4 APRIL 2014 loaded for Anonymous User (n/a) at Florida International University – Florida state

For personal use only. No other uses without permission. Copyrig

17. Messinger D, Young GS, Ozonoff S, et al. Beyond autism: a Baby Siblings Research Consortium study of high-risk children at three years of age. J Am Acad Child Adolesc Psychiatry. 2013;52: 300-308.

18. Lord C, Risi S, Lambrecht L, et al. The Autism Diagnostic Obser- vation Schedule—Generic: a standard measure of social and communication deficits associated with the spectrum of autism. J Autism Dev Disord. 2000;30:205-223.

19. Rutter M, Bailey A, Lord C. Social Communication Questionnaire: Manual. Los Angeles: Western Psychological Services; 2003.

20. Mullen EM. Mullen Scales of Early Learning. Circle Pines, MN: American Guidance Service; 1995.

21. Ozonoff S, Iosif A, Baguio F, et al. A prospective study of the emergence of early behavioral signs of autism. J Am Acad Child Adolesc Psychiatry. 2010;49:258-268.

22. Laird NM, Ware JH. Random-effects models for longitudinal data. Biometrics. 1982;38:963-974.

23. SAS Institute. SAS/STAT Version 9.3. Cary, NC: 2002-2010. 24. Szatmari P, Jones MB, Tuff L, et al. Lack of cognitive impairment in

first-degree relatives of children with pervasive developmental disorders. J Am Acad Child Adolesc Psychiatry. 1993;32:1264-1273.

25. Gamliel I, Yirmiya N, Sigman M. The development of young siblings of children with autism from 4 to 54 months. J Autism Dev Disord. 2007;37(1):171-183.

26. Yirmiya N, Gamliel I, Pilowsky T, Feldman R, Baron-Cohen S, Sigman M. The development of siblings of children with autism at 4 and 14 months: social engagement, communication, and cogni- tion. J Child Psychol Psychiatry. 2006;47:511-523.

27. Yirmiya N, Gamliel I, Shaked M, Sigman M. Cognitive and verbal abilities of 24-to 36-month-old siblings of children with autism. J Autism Dev Disord. 2007;37:218-229.

28. Bolton PF, Golding J, Emond A, Steer CD. Autism spectrum dis- order and autistic traits in the Avon Longitudinal Study of Parents and Children: precursors and early signs. J Am Acad Child Adolesc Psychiatry. 2012;51:249-260.

29. Wallace KS, Rogers SJ. Intervening in infancy: implications for autism spectrum disorders. J Child Psychol Psychiatry. 2010;51: 1300-1320.

30. Rogers SJ, Vismara L. Evidence-based comprehensive treatments for early autism. J Clin Child Adolesc Psychol. 2008;37:8-38.

31. Webster-Stratton CH, Reid MJ, Beauchaine T. Combining parent and child training for young children with ADHD. J Clin Child Adolesc Psychol. 2011;40:191-203.

Y

www.jaacap.org 407 consortium from ClinicalKey.com by Elsevier on January 09, 2019. ht ©2019. Elsevier Inc. All rights reserved.

OZONOFF et al.

SUPPLEMENT 1

Mixed-Effects Model Details. Mixed-effects regres- sion models were used to estimate individual patterns of change in Mullen Scale raw scores, Examiner-Rated Social Engagement Scores, and Autism Diagnostic Observation Schedule (ADOS) social-communication scores from 6 to 36 months, and to test the effects of diagnosis and covariates on the initial level and the rate of change in these variables. Change in these variables was assessed in the mixed-effects models with a term for age (centered at baseline 6 months). The models as- sume that each child’s individual path of growth followed the mean path, except for child-specific random effects that caused the initial level to be higher or lower and the rate of change (linear, quadratic) to be faster or slower.

The core set of models included fixed effects for diagnosis, age (centered at baseline), and the interaction between diagnosis and age. A second

JOURN 407.e1 www.jaacap.org

Downloaded for Anonymous User (n/a) at Florida International Universit For personal use only. No other uses without perm

set of models also included terms for the quadratic effect of age and the interaction of diagnosis with the quadratic effect of age. These interaction terms tested whether the rate of change in the variables varied across diagnosis. The most general core model included 3 random effects: a random intercept and random slopes for both the linear and the quadratic effect of age. These random ef- fects (describing the between-child variation) were assumed to follow a multivariate normal distri- bution. We used an unstructured covariance ma- trix, which is the most general structure, to model the dependence between the random effects. To model within-person variation, we assumed that the observed measurements differed from the child’s true trajectory by independent, identically distributed errors at each visit. Separate variances for this residual error (assessing the within-child variance) were estimated in each group, and tests to assess whether the within-child variances differed across diagnosis were performed.

AL OF THE AMERICAN ACADEMY OF CHILD & ADOLESCENT PSYCHIATRY VOLUME 53 NUMBER 4 APRIL 2014

y – Florida state consortium from ClinicalKey.com by Elsevier on January 09, 2019. ission. Copyright ©2019. Elsevier Inc. All rights reserved.

TABLE S1 Estimated Trajectories (Estimate, 95% CI) From the Mixed-Effects Models Predicting Mullen Scale Raw Scores, Examiner-Rated Social Engagement Scores, and Autism Diagnostic Observation Schedule (ADOS) Social- Communication Scores

Age

Estimated Scores (95% Confidence Intervals)

ASD (n ¼ 51)

Non-TD (n ¼ 83)

High-Risk TD (n ¼ 160)

Low-Risk TD (n ¼ 116)

Mullen Expressive Language 6 mo 5.9 (5.5e6.4) 6.5 (6.2e6.9) 6.6 (6.3e6.8) 6.2 (5.9e6.5) 12 mo 9.8 (9.4e10.3) 11.3 (10.9e11.6)** 12.0 (11.7e12.2) 11.9 (11.7e12.2)** 18 mo 13.8 (13.2e14.3) 16.0 (15.6e16.4)*** 17.4 (17.1e17.7) 17.7 (17.3e18.0)*** 24 mo 17.7 (16.9e18.4) 20.8 (20.2e21.3)*** 22.8 (22.4e23.3) 23.4 (22.9e23.9)*** 36 mo 25.5 (24.2e26.8) 30.2 (29.2e31.2)*** 33.7 (33.0e34.4) 34.8 (34.0e35.6)***

Mullen Receptive Language 6 mo 7.5 (6.8e8.2) 6.7 (6.3e7.3) 7.2 (6.8e7.6) 6.7 (6.3e7.1) 12 mo 10.4 (9.8e11.0) 12.2 (11.8e12.6)*** 13.3 (13.0e13.6) 13.7 (13.4e14.1)*** 18 mo 13.6 (12.8e14.5) 17.2 (16.6e17.9)*** 18.9 (18.5e19.4) 20.0 (19.5e20.5)*** 24 mo 17.2 (16.1e18.1) 21.8 (21.1e22.6)*** 24.0 (23.5e24.6) 25.6 (24.9e26.2)*** 36 mo 25.0 (23.5e26.5) 29.7 (28.6e30.8)*** 32.7 (31.9e33.5) 34.4 (33.4e35.3)***

Mullen Visual Reception 6 mo 9.9 (9.3e10.4) 9.5 (9.1e10.0) 9.6 (9.3e10.0) 9.5 (9.2e9.9) 12 mo 14.7 (14.2e15.1) 15.2 (14.9e15.5)** 15.9 (15.7e16.2) 15.9 (15.6e16.1)** 18 mo 19.3 (18.8e19.9) 20.7 (20.3e21.1)*** 22.0 (21.7e22.3) 22.1 (21.7e22.4)*** 24 mo 23.8 (23.1e24.6) 26.1 (25.5e26.6)*** 28.0 (27.6e28.4) 28.2 (27.7e28.7)*** 36 mo 32.4 (31.1e33.8) 36.4 (35.3e37.4)*** 39.5 (38.7e40.2) 39.9 (39.0e40.8)***

Mullen Fine Motor 6 mo 8.9 (8.4e9.5) 9.1 (8.7e9.6) 9.4 (9.1e9.8) 9.3 (8.9e9.6) 12 mo 14.6 (14.3e15.0) 15.1 (14.9e15.4)# 15.3 (15.1e15.5) 15.5 (15.2e15.7)#

18 mo 19.4 (18.9e19.8) 20.2 (19.9e20.6)** 20.6 (20.4e20.9) 21.0 (20.7e21.3)** 24 mo 23.1 (22.5e23.7) 24.5 (24.1e25.0)*** 25.4 (25.1e25.7) 25.9 (25.5e26.3)*** 36 mo 27.7 (26.5e29.0) 30.5 (29.5e31.4)*** 33.3 (32.7e33.9) 33.8 (33.1e34.6)***

Examiner-Rated Social Engagement Composite Score 6 mo 7.6 (7.0e8.2) 7.5 (7.0e7.9) 7.6 (7.2e8.0) 7.9 (7.5e8.3) 12 mo 7.2 (6.9e7.5) 7.8 (7.5e8.0)*** 8.0 (7.7e8.2) 8.4 (8.1e8.6)*** 18 mo 6.8 (6.5e7.1) 7.9 (7.7e8.2)*** 8.2 (8.0e8.4) 8.7 (8.5e8.9)*** 24 mo 6.5 (6.2e6.9) 8.0 (7.7e8.3)*** 8.4 (8.2e8.7) 8.9 (8.6e9.0)*** 36 mo 6.2 (5.8e6.6) 7.9 (7.7e8.2)*** 8.7 (8.5e8.9) 8.8 (8.6e9.0)***

ADOS Social-Communication Score 18 mo 10.9 (9.6e12.2) 4.7 (4.1e5.2)*** 3.5 (3.1e3.9) 2.4 (1.9e2.9)*** 24 mo 8.7 (7.5e9.9) 4.2 (3.7e4.7)*** 2.5 (2.2e2.8) 2.2 (1.7e2.6)*** 36 mo 13.1 (11.8e14.4) 5.7 (5.3e6.1)*** 1.9 (1.6e2.1) 1.5 (1.2e1.8)***

Note: ASD ¼ autism spectrum disorder; TD ¼ typically developing. #p < .07; *p < .05; **p < .01; ***p < .001 (for comparing Non-TD and Low-Risk TD).

JOURNAL OF THE AMERICAN ACADEMY OF CHILD & ADOLESCENT PSYCHIATRY VOLUME 53 NUMBER 4 APRIL 2014 www.jaacap.org 407.e2

BROADER AUTISM PHENOTYPE IN INFANCY

Downloaded for Anonymous User (n/a) at Florida International University – Florida state consortium from ClinicalKey.com by Elsevier on January 09, 2019. For personal use only. No other uses without permission. Copyright ©2019. Elsevier Inc. All rights reserved.

Contents lists available at ScienceDirect

Neuroscience and Biobehavioral Reviews

journal homepage: www.elsevier.com/locate/neubiorev

Changing conceptualizations of regression: What prospective studies reveal about the onset of autism spectrum disorder

Sally Ozonoffa,⁎, Ana-Maria Iosifb

a Department of Psychiatry and Behavioral Sciences, MIND Institute, University of California – Davis, 2825 50th Street, Sacramento CA, 95817, USA b Department of Public Health Sciences, University of California – Davis, Medical Sciences 1C, Davis CA, 95616, USA

A R T I C L E I N F O

Keywords: Autism spectrum disorder Onset patterns Regression Prospective studies

A B S T R A C T

Until the last decade, studies of the timing of early symptom emergence in autism spectrum disorder (ASD) relied upon retrospective methods. Recent investigations, however, are raising significant questions about the accuracy and validity of such data. Questions about when and how behavioral signs of autism emerge may be better answered through prospective studies, in which infants are enrolled near birth and followed longitudinally until the age at which ASD can be confidently diagnosed or ruled out. This review summarizes the results of recent studies that utilized prospective methods to study infants at high risk of developing ASD due to family history. Collectively, prospective studies demonstrate that the onset of ASD involves declines in the rates of key social and communication behaviors during the first years of life for most children. This corpus of literature suggests that regressive onset patterns occur much more frequently than previously recognized and may be the rule rather than the exception.

1. Introduction

The onset of behavioral signs of autism spectrum disorder (ASD) is usually conceptualized as occurring in one of two ways: an early onset pattern, in which children demonstrate delays and deviances in social and communication development early in life, and a regressive pattern, in which children develop largely as expected for some period and then experience a substantial decline in or loss of previously developed skills. While it was long believed that the majority of children with ASD demonstrated an early onset pattern, more recent studies suggest that regressive onset occurs more frequently than previously recognized (Brignell et al., 2017; Hansen et al., 2008; Kern et al., 2015; Pickles et al., 2009; Shumway et al., 2011; Thurm et al., 2014; for a review, see meta-analysis by Barger et al., 2013). Studies occasionally also identify a third onset pattern, that of developmental stagnation or plateau (Shumway et al., 2011), that is characterized by intact early skills that fail to progress or transform into more advanced developmental achievements. This onset pattern is distinct from regression, in that the child does not lose acquired skills, but instead fails to make expected gains.

1.1. Methods for measuring onset patterns

The most common procedure for collecting information about the

timing of early symptoms is retrospective parent report. A number of factors can influence report validity, including awareness of the child’s eventual diagnosis and knowledge of developmental milestones. It has long been understood that retrospective reports are subject to problems of memory and interpretation (Finney, 1981; Henry et al., 1994; Pickles et al., 1996), including in studies of ASD (Andrews et al., 2002). Mul- tiple studies have documented the ways in which recall problems and other biases can influence parent report. Changes in recall occur over time, with past events often reported to occur more recently, closer to the time of recollection, than they actually took place, an error called forward telescoping (Loftus and Marburger, 1983). Studies of children with ASD have demonstrated significant forward telescoping in parent report of milestones, resulting in parents being less likely to report re- gression and more likely to report early delays as their children grow older (Hus et al., 2011; Lord et al., 2004). A recent study from our research team (Ozonoff et al., 2018a) conducted longitudinal inter- views with parents about onset of ASD symptoms when their child was 2–3 years old (Time 1) and approximately 6 years old (Time 2). Sig- nificant forward telescoping was found in both age of regression and age when milestones were achieved. The correspondence between Time 1 and Time 2 parent report of onset was low (kappa = .38). One-quarter of the sample changed onset categories, most often due to parents not recalling a regression at Time 2 that they had reported at Time 1.

Analysis of home movies of children later diagnosed with ASD is

https://doi.org/10.1016/j.neubiorev.2019.03.012 Received 24 September 2018; Received in revised form 12 February 2019; Accepted 14 March 2019

⁎ Corresponding author. E-mail addresses: [email protected] (S. Ozonoff), [email protected] (A.-M. Iosif).

Neuroscience and Biobehavioral Reviews 100 (2019) 296–304

Available online 15 March 2019 0149-7634/ © 2019 Elsevier Ltd. All rights reserved.

T

another retrospective method used in research studies to study symptom emergence (Goldberg et al., 2008; Palomo et al., 2006). Video analysis may be a more objective procedure for documenting early symptoms than parent recall (Werner and Dawson, 2005) but it is labor- intensive and subject to other limitations, such as selective recording (e.g., tendency of parents to film positive behaviors). In a study from our team that compared classification of onset based on coding of fa- mily movies to onset type as recalled by parents (Ozonoff et al., 2011a), less than half of children whose home video displayed clear evidence of a major decline in social and communication behavior were reported to have had a regression by parents. Similarly, only 40% of participants with clear evidence of early delays and little evidence of skill decline on video were reported by parents to show an early onset pattern.

1.2. Prospective studies of onset

Questions about when and how behavioral signs of autism emerge may be better answered through prospective investigations, in which infants are recruited and enrolled near birth, prior to the advent of parent concerns, and then followed longitudinally through the window of developmental risk, until the age at which ASD can be confidently diagnosed or ruled out, usually 36 months. A few large general popu- lation cohorts have been studied prospectively to examine onset pat- terns (Brignell et al., 2017; Havdahl et al., 2018) but this study design is inefficient, since fewer than 2 in 100 participants will develop ASD (Centers for Disease Control and Prevention, 2018), making it difficult to achieve an appropriate sample size. Additionally, large prospective cohort studies must, of necessity, rely upon parent questionnaires and rarely provide the opportunity for in-person clinical assessments to verify diagnosis or onset pattern.

For this reason, most prospective investigations utilize high-risk samples in order to increase the number of ASD outcomes that are in- formative for study. The most widely used high-risk group has been later-born siblings of children with ASD, who are known to be at higher ASD risk than the general population (Constantino et al., 2010). Most investigations compare high-risk infants to lower-risk participants with no known family history of ASD in first-, second-, and sometimes third- degree relatives. This study design improves on retrospective methods in a number of important ways. Serial comprehensive assessments, in standardized testing contexts, are used to document the timing of symptom emergence, thus avoiding reliance on potentially fallible parent recall or non-representative home video. Assessments can utilize a wide range of tools, including eye tracking, EEG, and imaging, al- lowing broader investigations of symptom onset and testing of specific hypotheses. And while most retrospective studies recruit samples through clinics, which may influence the results by including more severely affected children, infant sibling studies avoid such potential biases by ascertaining participants via family history alone.

Several recent papers provide comprehensive reviews of the infant sibling literature (Bölte et al., 2013; Jones et al., 2014; Pearson et al., 2018; Szatmari et al., 2016). Here we focus on research reports of greatest relevance to symptom emergence, specifically those that study infants beginning in the first year of life on measures that are appro- priate for examining potential skill decline over time. Using a variety of different prospective methods, these studies have reported largely in- tact early development, followed by developmental declines and onset of symptoms around the first birthday and in the second year of life. For example, Zwaigenbaum et al. (2005), using the Autism Observation Scale for Infants (AOSI), reported no differences at 6 months between infants subsequently diagnosed with ASD and both high- and low-risk infants without ASD outcomes; significant group differences emerged at 12 months and increased over time. This pattern on the AOSI was later replicated by an independent research team (Gammer et al., 2015). Wan et al. (2013) found that infant-parent interaction quality at 6–10 months did not predict which children would be diagnosed with ASD at age 3, but by 12–15 months, such variables were significantly

associated with diagnostic outcome. Similar findings of lack of early group differences (or lack of early predictive ability), followed by later divergence from typically developing infants, have been reported by Landa and Garrett-Mayer (2006) using the Mullen Scales of Early Learning, Rozga et al. (2011) in joint attention, Bedford et al. (2012) on a gaze-following eye-tracking task, Elsabbagh et al. (2013) on a gap- overlap attention task, and Wolff et al. (2014) studying repetitive be- havior. In an incisive recent review that attempts to reconcile retro- spective and prospective studies of regression and explore how study design affects the likelihood of capturing regression, Pearson et al. (2018) conclude that, among infants who later develop ASD, “the ma- jority show declining fixation of eyes, gaze to faces, and social en- gagement, from typical levels in early infancy (2–6 months) to sig- nificantly reduced levels by 24–36 months (p. 14).”

2. Findings from the University of California Davis infant sibling study

In our laboratory, we have taken the analytic approach of growth curve modeling to examine directly the evidence of longitudinal de- velopmental change in the first three years of life. Between 2003 and 2015, the UC Davis Infant Sibling Study recruited three cohorts of later- born siblings, each composed of 50 low-risk and 100 high-risk infants. Participants were tested as early as 6 months of age and then seen every 3 to 6 months until their 3rd birthday (up to 7 in-person evaluations). They have since been followed into school age and tested at approxi- mately three-year intervals. The oldest children from Cohort 1 are now 16 years of age and the retention rate is over 80%. At each infant and preschool visit, a battery of age-appropriate standardized tests and experimental tasks was administered that measured language, cogni- tion, social, communication, motor, and many other domains. Approximately 20% of the high-risk infants were later diagnosed with ASD (Ozonoff et al., 2011b). Diagnoses of ASD were made at any point that a child met criteria (mean age 24.2 months) but a full diagnostic assessment was completed on all children, regardless of previous find- ings, at 36 months by examiners unaware of family risk or prior as- sessment results. In the following sections, we summarize several stu- dies from these cohorts that consistently demonstrate declining trajectories across a variety of different measures and developmental domains.

The phenomenon of regression is defined by loss or significant de- crease in already-acquired skills. Thus, a critical methodological issue in prospective studies that wish to examine onset patterns is selection of which behaviors to measure. They must be 1) developmentally appro- priate across the full age window of risk and 2) robustly present, at high frequency, in the first year of life. Such behaviors have the capacity to decrease and are therefore of highest relevance to the study of onset patterns. Measures that focus on socio-communicative behaviors that have not yet emerged in the first year of life, such as joint attention, imitation, and verbal communication, will be less useful for testing hypotheses about declining capacities. The behaviors our laboratory has focused on, including gaze to faces and eyes of others, shared affect, and social interest/engagement, are well developed in the first year of life (Inada et al., 2010) and therefore of highest relevance in the pro- spective measurement of regression.

2.1. Behavioral coding of social-communication rates

Our first exploration of longitudinal change in early social and communicative behaviors used video recordings of participants inter- acting with examiners during structured developmental testing (Ozonoff et al., 2010). Research assistants, unaware of family risk group or diagnostic outcome, were trained to 90% reliability to detect three behaviors: gaze to an adult’s face, smiles at an adult that were paired with eye contact, and vocalizations directed at an adult that were ac- companied by eye contact. Rate per minute of eye gaze, shared affect,

S. Ozonoff and A.-M. Iosif Neuroscience and Biobehavioral Reviews 100 (2019) 296–304

297

and directed vocalizations of the 25 children in Cohort 1 with outcomes of ASD were compared to those of 25 children who did not have ASD outcomes randomly selected from the low-risk group. The two groups behaved similarly at 6 months: frequencies of none of the three beha- viors differed between the groups and effect sizes were in the small range. Over time, the Low-Risk (LR) Non-ASD group had a significant increase in social smiles and directed vocalizations, while maintaining the same consistently high level of gaze to faces. In the ASD group, in contrast, the rates of all three behaviors dramatically decreased over time. Fig. 1A displays longitudinal trajectories of eye contact rate per minute, showing comparable values between groups at 6 months, fol- lowed by group differences that became statistically significant by 12 months and persisted and widened over time. These longitudinal de- creases in the rates of key social and communicative behaviors provided the first prospectively measured evidence consistent with a regressive onset pattern.

We have since replicated these findings (see Fig. 1B) using the same methods in an independent sample of 46 infants later diagnosed with ASD from Cohorts 2 and 3 of our longitudinal project. In this analysis (Gangi et al., in preparation), a third group, composed of high-risk (HR) infants who did not have ASD outcomes, was also included. This group was not different from the LR Non-ASD group in the frequency of gaze to adult faces at any age and did not show any evidence of decline in development, which was evident only in the participants developing ASD, replicating our 2010 study.

2.2. Observer ratings of social engagement

Coding behavior frequencies from video is time-consuming, labor- intensive, expensive, and not transferable to clinical contexts, so our research program has also sought to establish whether declining tra- jectories are evident using other methodological approaches. At the end of each visit, examiners rate the frequency of eye contact, shared affect, and overall social engagement (number of social initiations and social responses) made by the infant throughout the session, across all tasks, using a 3-point scale (1 = rare, 2 = occasional, 3 = frequent). These scores are summed to create a composite that ranges from 3 to 9. As reported in Ozonoff et al. (2010) and depicted in Fig. 2A, these ex- aminer ratings of social engagement showed similar longitudinal pat- terns to the social behaviors coded from video. There were no group differences in the 6-month examiner ratings; however, while the LR Non-ASD group showed a significant increase in social engagement ratings over time, reaching close to the maximum score by 36 months,

the children in the ASD outcome group had a strong decline in social engagement ratings over the same time period.

This finding was recently replicated in an independent group of 32 infants with ASD outcomes from Cohorts 2 and 3 of our sample (Ozonoff et al., 2018b). We used the same examiner rating variable (this time with an expanded 5-point scale) and compared the ASD group to both a low-risk and a high-risk group without ASD. Again, all three groups had comparable levels of social engagement based on examiner scores at 6 months of age. The ASD group then demonstrated a decrease in scores with age, while the HR Non-ASD group showed stable high scores over time and the LR Non-ASD group demonstrated increasing scores longitudinally. By 12 months, the two Non-ASD groups demon- strated significantly higher rates of social engagement, as judged by examiners, than the ASD group and these differences widened over time, as can be seen in Fig. 2B. Along with our 2010 paper, these findings demonstrate that the declines in the frequency of social and communication behaviors detected through more labor-intensive video coding methods are also detectable through much simpler methods that would be feasible for broader use, such as brief observational ratings of social engagement by clinical professionals.

2.3. Longitudinal parent ratings of social behavior

The question remained, however, whether such findings could be an artifact or byproduct of the assessment context with an unfamiliar ex- aminer. For both clinical use and future development of screening methods, it is critical to also establish whether parent ratings are sen- sitive to the developmental decline phenomenon we have reported. In a recent study (Ozonoff et al., 2018b), we examined parent prospective ratings of the same early-appearing socio-communicative behaviors measured in the video and examiner ratings. Parents in our study completed the Early Development Questionnaire (EDQ; Ozonoff et al., 2005) prior to each visit. The EDQ consists of 45 questions about the child’s current functioning in social, communication, and repetitive behavior domains. Each item is rated on a 4-point frequency scale (0=behavior never occurs, 3=behavior often occurs). Three items, comparable in content to the video codes and examiner ratings, were summed: item 1 (“my child looks at me during social interactions”), item 4 (“my child smiles back at me when I smile at him/her”), and item 13 (“when I call my child’s name, he/she looks at me right away”). In addition to being parallel to the behaviors rated by examiners, these items were selected because they represent early-appearing behaviors that are relevant and developmentally appropriate across all ages of the

Fig. 1. Declining trajectories of gaze to eyes in children developing ASD, coded from a videotaped interaction with an examiner. Panel A: Cohort 1, Panel B: Cohorts 2 and 3.

S. Ozonoff and A.-M. Iosif Neuroscience and Biobehavioral Reviews 100 (2019) 296–304

298

study (6–36 months). In contrast to other EDQ items that measure later- developing skills (e.g., joint attention, language), the items selected for the composite measure behaviors present in the first year of life (Inada et al., 2010). As with the behaviors we selected for coding and examiner observational ratings, it was critical that the behaviors rated by parents have the potential to demonstrate decreases over time as ASD signs emerge. The composite variable, quantifying parent report of the fre- quency of key early social behaviors, had a potential range of 0 – 9. On the parent-rated EDQ, there were again no group differences at 6 months. As with the other measures, the ASD group showed a decline in levels of social engagement with age, while both the high-risk and low- risk Non-ASD groups demonstrated gains in social engagement over time. The ASD group’s scores were significantly lower than both Non- ASD groups by 12 months and the differences increased with age, de- monstrating the same declining trajectory as evident in the coded be- havior and examiner ratings (see Fig. 3A).

Employing a similar approach, we replicated the ability of parent report to capture the decline in social and communication development using a standardized, normed measure (Parikh et al., 2018), the Infant- Toddler Checklist (ITC), a 24-item parent questionnaire from the Communication and Social Behavior Scales (CSBS; Wetherby and

Prizant, 2002). This instrument is normed from 6 to 24 months and includes questions that span this developmental range, from early-ap- pearing behaviors like social smiling to those that emerge at older ages, such as spoken language and pretend play. We created a composite of three items that represent behaviors typically present in the first year of life (item 2: “when your child plays with toys, does he/she look at you to see if you are watching?”; item 3: “does your child smile or laugh while looking at you?”; item 19: “when you call your child’s name, does he/she respond by looking or turning toward you?”). We then com- pared growth trajectories in infants subsequently diagnosed with ASD (n = 46) to the HR Non-ASD (n = 139) and LR Non-ASD groups (n = 96). There were no group differences on the 3-item ITC composite at 6 months of age; however, over time, the ASD group showed a de- cline in scores, while the two Non-ASD groups demonstrated gains (see Fig. 3B). This resulted in the ASD group having significantly lower scores at 24 months than both comparison groups.

These studies from our lab show that children with ASD, as a group, evidence declines in development from 6 to 36 months. Such declines are seen only in the ASD group and not in comparison samples, even those with elevated genetic risk or other developmental concerns. Findings of declining trajectories have since been replicated by other

Fig. 2. Declining trajectories of social engagement in children developing ASD, as rated by examiners unaware of risk group or outcome. Panel A: Cohort 1, Panel B: Cohorts 2 and 3.

Fig. 3. Declining trajectories of social engagement in children developing ASD, as rated by parents, Cohorts 2 and 3. Panel A: Early Development Questionnaire, Panel B: Infant Toddler Checklist.

S. Ozonoff and A.-M. Iosif Neuroscience and Biobehavioral Reviews 100 (2019) 296–304

299

independent research teams. Landa et al. (2013) examined growth trajectories in infants later diagnosed with ASD and Non-ASD partici- pants. Approximately half of the children with ASD, labeled the Early- ASD group, demonstrated differences from the Non-ASD cases at 14 months but the other half (Later-ASD group) did not diverge from ty- pical infants until 24 months. The Later-ASD group demonstrated a steep decline in shared positive affect, as measured by the CSBS (Wetherby and Prizant, 2002), between 14 and 24 months. Jones and Klin (2013) conducted a prospective eye-tracking study with high- and low-risk infants to assess attention to eyes. The authors reported that very early in development (i.e., first two months of life), both low-risk and high-risk infants displayed high levels of attention to eyes, with no group differences. However, high-risk infants who were later diagnosed with ASD began to demonstrate a steady decline in looking at eyes at four months, reaching a level that was approximately half that of low- risk infants by 24 months. What was most predictive of a later ASD outcome was not the amount of visual fixation on eyes displayed at any particular age, but the overall declining trajectory over time. This study found that the majority of infants developing ASD demonstrated this declining pattern.

2.4. Growth curve modeling approaches to determining onset classifications

In aggregate, the studies reviewed up to this point converge on the conclusion that longitudinal decreases in key social behaviors are a signature of the early emergence of ASD. But these data do not clarify how widespread such phenomena are within ASD and whether the group-level findings are driven by extreme outliers or characterize a majority of young children developing ASD. In our lab, we have ap- proached this issue analytically using multivariate Latent Class Analysis (LCA; Muthen, 2004), permitting us to identify distinct subgroups of children based on their longitudinal patterns on multiple measures of social communication. This technique does not rely on preconceived notions or poorly defined definitions of onset phenomena, but instead uses statistical modeling to empirically derive the optimum number of classes described by the patterns of performance demonstrated in the measures.

Data from a recent paper (Ozonoff et al., 2018b) address the ques- tion of how widespread the declining trajectories pattern is within a group of 32 infants subsequently diagnosed with ASD. We employed latent class growth models to examine potential within-group variation in onset patterns, using both examiner ratings of social engagement and parent ratings from the EDQ (see Sections 2.2 and 2.3 for instrument descriptions). Best model fit for examiner ratings was a two-group so- lution. Using their highest posterior group probability, the 32 partici- pants were classified into two trajectories (see Fig. 4A, which also presents the Low- and High-Risk Non-ASD groups as contrasts). Only a small proportion of the ASD cases (n = 4; 13%) were assigned to an Early Onset/No Regression group by the latent class analyses, based on examiners prospectively reporting low levels of social behavior at all ages. The vast majority (n = 28; 88%) were classified by these analyses into a Regression group, in which examiners prospectively rated in- itially high levels of social engagement that dropped significantly over time.

The best fit for the parent EDQ 3-item composite in the latent class models was a three-group solution: Group 1, an Early Onset trajectory, Group 2, a Declining trajectory, and Group 3, an Improving trajectory (see Fig. 4B). Parents prospectively reported low levels of social en- gagement at all ages for Group 1, which again made up a small minority of the sample (n = 4; 13% of the sample). The majority of the sample (n = 22; 69%) was classified in Group 2; these children were pro- spectively reported by parents to show high rates of social engagement early in life, which significantly declined over time. Parents of children in Group 3 (n = 6; 19%) prospectively reported low levels of skills at early ages that then significantly increased over time.

2.5. Concordance between retrospective and prospective onset classifications

In several of our studies, we have examined the correspondence between prospectively- and retrospectively-defined onset patterns and in each case have found them to be quite poor. For example, in our initial paper (Ozonoff et al., 2010), we compared onset classifications based upon coded frequencies of social and communicative behaviors to onset classifications employing retrospective parent report on the Autism Diagnostic Interview-Revised (ADI-R; Le Couteur et al., 2003). Using prospective observational data, 86% of the ASD sample showed decreasing rates of eye contact, social smiles, and vocalizations over time, but by parental recall using the ADI-R, only 17% of the children were classified as having regressive onset. In a more recent study (Ozonoff et al., 2018b), 69% of parents rated their child in a manner consistent with regression on a prospective questionnaire (the EDQ, described in section 2.3), but only 29% rated that same child as losing skills using a retrospective measure (the ADI-R). Parents were able to implicitly identify the changes in their child’s development over time when making ratings of the frequencies of current behaviors, but often did not explicitly label these changes as skill loss or regression when asked in a more categorical way. These results were particularly striking since both relied upon parental observations of the child. Si- milar findings of low concordance between retrospective and pro- spective methods of defining onset were reported by Landa et al. (2013). And a recent large general population study (Havdahl et al., 2018) found similar under-reporting of losses based on a retrospective parent interview. Of parents who prospectively reported a loss, defined as rating certain social behaviors as present at 18 months but absent at 36 months, only a striking 2% of them recalled such a decline, or la- beled it as a loss, when asked at age 3.

3. Conclusions and theoretical implications

A number of conclusions can be drawn from the collective body of work reviewed in this paper.

3.1. Onset involves declining social development

ASD emerges over the first two years of life and is not present “from the beginning of life” as stated by Kanner (1943, p. 242) in his seminal paper. For many years, it was presumed that ASD signs were present, but were just challenging to measure, from birth. Diagnostic criteria for ASD were developed at a time when children with autism were rarely, if ever, identified in infancy and thus many symptoms in the DSM and ICD criteria, such as delays or deficits in gestures, language, imitation, and pretense, are less relevant to the first year of life. As we have empha- sized in this review, one key to understanding early symptom emer- gence is to focus on very early-appearing social behaviors, those that are robustly present in early infancy, such as social interest, shared affect, gaze to faces and eyes, and response to name. When such a methodological approach is taken, there is clear evidence, across mul- tiple methods, replicated by independent research teams, of declining social behavior over time, after a period of relatively typical develop- ment. There is convergence across studies of a lack of group differences from comparison samples without ASD before 9 months of age, fol- lowed by statistically significant differences starting at 12 months that widen over time. Logically, if certain skills are evident at typical rates at an early point in development and then those same skills, defined and measured the same way later in development, have substantially di- minished, resulting in statistically significant differences from typical infants, a loss or regression of some magnitude must have occurred.

This paper advocates for taking a dimensional approach and using trajectories to identify patterns of onset. We are not arguing, however, for a fundamental reconsideration of the use of the word “regression.” The Merriam-Webster definition of the word regression is “a trend or

S. Ozonoff and A.-M. Iosif Neuroscience and Biobehavioral Reviews 100 (2019) 296–304

300

shift toward a lower or less perfect state, such as (a) progressive decline of a manifestation of disease or (b) gradual loss of acquired skills.” This aptly describes the loss of established skills (e.g., eye contact, response to name, social interest) that occurs during the declines in social de- velopment described in this paper.

3.2. Regression in ASD is the Rule, not the exception

The data summarized in this review suggest that a regressive pattern of onset is much more common than previously thought, the rule rather than the exception. While retrospective studies yielded regression es- timates of 20–30%, prospective data put them much higher, in some studies well over 80%. We and others (Jones et al., 2014; Ozonoff et al., 2018b; Pearson et al., 2018) have suggested that the regressions re- ported by parents retrospectively on measures like the ADI-R represent just “the tip of the iceberg,” while prospective studies are able to cap- ture earlier, more gradual, subtle changes that may be less noticeable in real-time observation. A hypothesis deriving from this supposition is that concordance between retrospective and prospective methods should be most frequent when the regression occurs later, is more drastic or severe, and involves loss of clearly defined skills like lan- guage. No published studies have yet examined this question and it would be a fruitful avenue for future investigation.

We propose that the way ASD starts, for all children, is through declines in early social and communication abilities. This presents a testable hypothesis: that all infants developing ASD lose some skills, but at different ages, some of which may be harder to detect with current measurement approaches than others. It may be difficult for parents to perceive and describe changing patterns of development that occur over many months during infancy, particularly when the period of normalcy is fairly brief. Our team (Ozonoff et al., 2010, 2011a) and others (Pearson et al., 2018; Rogers, 2009; Szatmari et al., 2016; Thurm et al., 2014) have suggested that onset is better thought of dimensionally, as a continuum of age when social and communication behaviors begin to diverge or decline, rather than a dichotomy (regression v. early onset). In a dimensional conceptualization of onset, at one end of the con- tinuum lie children who display declines so early that they are difficult to measure and symptoms appear to have always been present. At the other end of the continuum are children who experience losses so late, when more skills have been acquired and thus there are more skills to lose, that the regression appears quite overt and dramatic. We propose that variable timing of these processes across children leads to symp- toms exceeding the threshold for diagnosis at different points in the first

3 years of life, resulting in a distributed curve of onset timing.

3.3. Simplex v. multiplex samples

An important question to consider is whether regression in infant sibling samples is representative of regression in children with ASD who are the first in their families to be diagnosed with the condition. If symptoms emerge differently in multiplex and simplex families, then the insights about onset afforded by prospective research may not be applicable to the general population of children with ASD. For example, perhaps children in multiplex families are more likely to experience a regression than children from simplex families, accounting for the higher rates of decline apparent in prospective studies. We have no reason to believe this is the case. In fact, the rate of retrospectively- reported regression in multiplex families has been reported to be similar or lower than in simplex families (Boterberg et al., 2019b; Parr et al., 2011), failing to account for the high rates apparent in infant sibling studies, whose participants are, by definition, from multiplex families. A related issue is that parents participating in infant sibling investiga- tions have an older child with ASD. It is possible that these parents may be different reporters than other parents, given their previous experi- ence with ASD. This may make them more astute observers of devel- opment than parents in the general population and therefore more likely to recognize skill decline. This hypothesis, however, is not sup- ported by the data presented earlier in this review in which parents in multiplex families also under-report skill loss (Landa et al., 2013; Ozonoff et al., 2010, 2018a, 2018b). Nevertheless, it is important to keep these cautions in mind in interpreting the extant data on onset patterns. Validating these results in different kinds of samples, such as community-based epidemiological cohorts or other high-risk groups like very preterm infants, will be critical.

3.4. Improving the measurement of onset

Collectively, the studies reviewed in this paper present significant concerns about the accuracy of the most widely used methods of measuring regression, that is, retrospective parent report, and argue against their widespread use. The challenge currently faced by the field is that there are no practical alternative strategies to parent report for characterizing onset status. The time-intensive process and cost of home videotape analysis is prohibitive for large samples. Future studies will continue to rely on retrospective data, of necessity, since inclusion criteria for most samples require a confirmed ASD diagnosis (i.e., not

Fig. 4. Latent classes of social engagement, demonstrating declining trajectories in the majority of children developing ASD, Cohorts 2 and 3. Panel A: Examiner ratings, Panel B: Parent ratings.

S. Ozonoff and A.-M. Iosif Neuroscience and Biobehavioral Reviews 100 (2019) 296–304

301

prospective data). Several strategies have been proposed to improve reporting (Ayhan

and Isiksal, 2005). To minimize comprehension or interpretation pro- blems, it is recommended that further specific information about the behavior in question be provided. ASD screening instruments have begun to incorporate video to improve accuracy (Marrus et al., 2018; Smith et al., 2017) and this strategy could be adapted to improve re- porting of onset patterns. For example, longitudinal video of a child experiencing skill loss could be shown to parents to illustrate the kinds of changes in behavior that define regression. To minimize recall pro- blems, the simplest approach, and the one shown to have the best va- lidity, is to ask respondents to consult relevant records prior to com- pleting the interview (Ayhan and Isiksal, 2005). Parents could, for example, review entries in baby books or journals or watch home video of the child prior to the interview. Another approach is to link reporting to key events in the respondent’s life by creating a detailed timeline and context that assist recall of specific details (Loftus and Marburger, 1983). This method has already been used by Werner et al. (2005) to improve recall of early development in ASD and it could be further adapted for reporting about onset patterns. Whether these methods will enhance the validity of parent report of onset remains to be seen and would be a fruitful area of future study. For further discussion, see also Boterberg et al. (2019a).

3.5. Validity of previous studies of regression

The studies reviewed in this paper call into serious question the validity of previous studies of regression, which have, of necessity and the lack of alternatives, relied upon retrospective measures. Refining methods of studying the onset of ASD has the potential to transform research programs on etiological factors that contribute to the devel- opment of ASD by providing more precise and accurate measurements of an important phenotype (Barbaresi, 2016; Thurm et al., 2018). A better understanding of the inflection points at which development diverges from a typical trajectory to an autism trajectory could be highly informative to the search for risk factors. Better measures of onset are urgently needed for etiologic studies, which have been hin- dered already by the tremendous heterogeneity of the autism pheno- type (Constantino and Charman, 2016). Many recent studies have ex- amined whether onset types are associated with potential etiologic factors and biological correlates, such as brain growth (Nordahl et al., 2011; Valvo et al., 2016), seizures (Barger et al., 2017), vaccinations (Goin-Kochel et al., 2016), gastrointestinal problems (Downs et al., 2014; Richler et al., 2006), immunological function (Scott et al., 2017; Wasilewska et al., 2012), and genetic and genomic variations (Goin- Kochel et al., 2017; Gupta et al., 2017; Parr et al., 2011), including mitochondrial and MeCP2 mutations (Shoffner et al., 2010; Veeraragavan et al., 2016; Xi et al., 2007). So far, none of these factors has been firmly associated with onset patterns. This may be due to the errors that are likely to have occurred in the classifications of onset done in these studies. Clearly, examining the biological underpinnings of an imprecise measure is problematic.

In a review of autism genetics, one of the major priorities identified for future research is the characterization of ASD subtypes to relate to genetic variations (Geschwind, 2011). As more and more risk genes for ASD are identified, the common molecular pathways that these genes share are becoming understood, with some expressed early in neuro- biological development and others later (Konopka et al., 2012). A twin study (Hallmayer et al., 2011) suggested a greater role for environ- mental factors in ASD than previously appreciated. A more precise timing of first symptom emergence would enhance identification of etiological factors and when they might operate, with potential im- plications for intervention and prevention.

4. Clinical implications

Finally, the studies reviewed here provide hope and promise for improvements in screening, early diagnosis, and treatment. Many pro- spective studies (e.g., Bosl et al., 2018; Jones and Klin, 2013; Ozonoff et al., 2010) used measures, such as behavioral coding, eye tracking, and EEG, that are expensive, labor intensive, and not practical for routine use. Studies reviewed in this paper, however, have demon- strated that prospective parent report can identify declining trajectories of development (Ozonoff et al., 2018b; Parikh et al., 2018), as long as the instruments focus on early-appearing social behaviors, present in the first year of life, that have the potential to demonstrate decreases over time as ASD signs emerge. Brief rating scales of this type, ad- ministered longitudinally at regular well-child health care visits, could provide a clinically feasible and cost-effective screening tool capable of detecting declines over time. We hypothesize that dynamic screening, which utilizes longitudinal screenings over time and comparison of scores across ages to identify declining trajectories, will improve identification over static, cross-sectional screenings examining whether a single score at a single age exceeds a cutoff. This approach has been successfully used in identifying Rett syndrome (RTT), where head cir- cumference is normal at birth, followed by deceleration of head growth between 5 months and 4 years (Hagberg et al., 2001; Tarquinio et al., 2012). Through the development of RTT-specific growth references throughout early childhood, based on mapping head circumference trajectories, diagnosis of RTT has been possible at earlier ages (Schultz et al., 1993; Tarquinio et al., 2012). We (Ozonoff et al., 2010, 2018b) and others (Landa et al., 2013; Pearson et al., 2018; Thomas et al., 2009) have suggested that this kind of dimensional, trajectory-based methodological approach, percentiling social and communication milestones as we percentile other growth parameters, could be applied to detect ASD early.

Prospective studies have repeatedly demonstrated that develop- mental declines follow a period in the first year of life when socio- communicative skills are largely intact. Such early intact skills can be capitalized upon in treatment, presenting opportunities for preventive intervention when the brain is rapidly developing and maximally malleable. For many years, the holy grail has been finding a marker present prior to symptom emergence, thus affording the possibility of earlier, possibly preventative, treatment during the prodromal period. If the prospective methods described in this paper can be harnessed to identify infants at risk for ASD, during the decline of skills, rather than after the decline was over, it might be possible to disrupt these trajec- tories prior to the full onset of symptoms (Dawson, 2011). Children could be provided immediate access to infant interventions (Fein et al., 2016; Rogers et al., 2014), capitalizing on still-preserved skills and harnessing the brain plasticity of early infancy to improve outcomes, lessen disability, and perhaps, prevent the full disorder from devel- oping.

Conflict of interest

The authors have no competing interests to declare.

Acknowledgments

This work was supported by NIH grants R01 MH068398 (Ozonoff) and R01 MH099046 (Ozonoff). Thank you to Sofie Boterberg for her reading of an earlier version of this manuscript. We are deeply grateful to all the children and parents who participated in and showed sus- tained commitment to our longitudinal program of research.

References

Andrews, N., Miller, E., Taylor, B., Lingam, R., Simmons, A., Stowe, J., et al., 2002. Recall bias, MMR, and autism. Arch. Diseases Childhood 87, 493–494.

S. Ozonoff and A.-M. Iosif Neuroscience and Biobehavioral Reviews 100 (2019) 296–304

302

Ayhan, H.O., Isiksal, S., 2005. Memory recall errors in retrospective surveys: a reverse record check study. Qual. Quant. 38, 475–493.

Barbaresi, W.J., 2016. Commentary: the meaning of “regression” in children with autism spectrum disorder: why does it matter? J. Dev. Behav. Pediatr. 37, 506–507.

Barger, B.D., Campbell, J.M., McDonough, J.D., 2013. Prevalence and onset of regression within autism spectrum disorders: a meta-analytic review. J. Autism Dev. Disord. 43, 817–828.

Barger, B.D., Campbell, J., Simmons, C., 2017. The relationship between regression in autism spectrum disorder, epilepsy, and atypical epileptiform EEGs: a meta-analytic review. J. Intellect. Dev. Disabil. 42, 45–60.

Bedford, R., Elsabbagh, M., Gliga, T., Pickles, A., Senju, A., et al., 2012. Precursors to social and communication difficulties in infants at-risk for autism: gaze following and attentional engagement. J. Autism Dev. Disord. 42, 2208–2218.

Bölte, S., Marschik, P.B., Falck-Ytter, T., Charman, T., Roeyers, H., et al., 2013. Infants at risk for autism: a European perspective on current status, challenges, and opportu- nities. Eur. Child Adolesc. Psychiatry 22, 341–348.

Bosl, W.J., Tager-Flusberg, H., Nelson, C.A., 2018. EEG analytics for early detection of autism spectrum disorder: a data-driven approach. Sci. Rep. 8, 6828.

Boterberg, S., Charman, T., Marschik, P., Bolte, S., Roeyers, H., 2019a. Regression in Autism Spectrum Disorder: Characteristics, Etiology, Early Development, and Outcomes – a Review of Retrospective Studies (this issue). .

Boterberg, S., Van Coster, R., Roeyers, H., 2019b. The Clinical Contribution of Parent Reported Regression in Autism Spectrum Disorder: Characteristics, Early Development, and Later Outcomes (under review). .

Brignell, A., Williams, K., Prior, M., Donath, S., Reilly, S., et al., 2017. Parent-reported patterns of loss and gain in communication in 1- to 2-year-old children are not unique to autism spectrum disorder. Autism 21, 344–356.

Centers for Disease Control and Prevention, 2018. Prevalence of autism spectrum disorder among children aged 8 years – autism and developmental disabilities monitoring network, 11 sites, United States, 2014. MMWR Surv. Summary 67, 1–23.

Constantino, J.N., Charman, T., 2016. Diagnosis of autism spectrum disorder: reconciling the syndrome, its diverse origins, and variation in expression. Lancet Neurol. 15, 279–291.

Constantino, J.N., Zhang, Y., Frazier, T., Abbacchi, A.M., Law, P., 2010. Sibling recur- rence and the genetic epidemiology of autism. Am. J. Psychiatry 167, 1349–1356.

Dawson, G., 2011. Editorial: coming closer to describing the variable onset patterns in autism. J. Am. Acad. Child Adolesc. Psychiatry 50, 744–746.

Downs, R., Perna, J., Vitelli, A., et al., 2014. Model-based hypothesis of gut microbe populations and gut-brain barrier permeability in the development of regressive autism. Med. Hypotheses 83, 649–655.

Elsabbagh, M., Fernandes, J., Webb, S.J., Dawson, G., Charman, T., et al., 2013. Disengagement of visual attention in infancy is associated with emerging autism in toddlerhood. Biol. Psychiatry 74, 189–194.

Fein, D., Helt, M., Brennan, L., Barton, M., 2016. The Activity Kit for Babies and Toddlers at Risk: How to Use Everyday Routines to Build Social and Communication Skills. Guilford Press, New York, NY.

Finney, H.C., 1981. Improving the reliability of retrospective survey measures: results of a longitudinal field survey. Eval. Rev. 5, 207–229.

Gammer, I., Bedford, R., Elsabbagh, M., Garwood, H., Pasco, G., Tucker, L., et al., 2015. Behavioral markers for autism in infancy: Scores on the Autism Observational Scale for Infants in a prospective study of at-risk siblings. Infant Behav. Dev. 38, 107–115.

Geschwind, D.H., 2011. Genetics of autism spectrum disorders. Trends Cogn. Sci. (Regul. Ed.) 15, 409–416.

Goin-Kochel, R.P., Mire, S.S., Dempsey, A.G., et al., 2016. Parental report of vaccine re- ceipt in children with autism spectrum disorder: do rates differ by pattern of ASD onset? Vaccine 34, 1335–1342.

Goin-Kochel, R.P., Trinh, S., Barber, S., Bernier, R., 2017. Gene disrupting mutations associated with regression in autism spectrum disorder. J. Autism Dev. Disord. 47, 3600–3607.

Goldberg, W.A., Thorsen, K.L., Osann, K., Spence, M.A., 2008. Use of home videotapes to confirm parental reports of regression in autism. J. Autism Dev. Disord. 38, 1136–1146.

Gupta, A.R., Westphal, A., Yang, D.Y., Sullivan, C.A., Eilbott, J., et al., 2017. Neurogenetic analysis of childhood disintegrative disorder. Mol. Autism 8, 19.

Hagberg, G., Stenbom, Y., Engerstrom, I.W., 2001. Head grown in Rett syndrome. Brain Dev. 23 (Supplement), S227–S229.

Hallmayer, J., Cleveland, S., Torres, A., Phillips, J., Cohen, B., et al., 2011. Genetic heritability and shared environmental factors among twin pairs with autism. Arch. Gen. Psychiatry 68, 1095–1102.

Hansen, R.L., Ozonoff, S., Krakowiak, P., Angkustsiri, K., Jones, C., et al., 2008. Regression in autism: prevalence and associated factors in the CHARGE study. Ambul. Pediatr. 8, 25–31.

Havdahl, A., Bishop, S., Farmer, C., Schjolberg, S., Bresnahan, M., et al., 2018. Loss of Social-communication Skills and Outcomes During Childhood in a Large General Population Cohort. Paper presented at the International Society for Autism Research meeting, Rotterdam.

Henry, B., Moffitt, T.E., Caspi, A., Langley, J., Silva, P.A., 1994. On the “remembrance of things past.” A longitudinal evaluation of the retrospective method. Psychol. Assess. 6, 92–101.

Hus, V., Taylor, A., Lord, C., 2011. Telescoping of caregiver report on the autism diag- nostic interview-revised. J. Child Psychol. Psychiatry 52, 753–760.

Inada, N., Kamio, Y., Koyama, T., 2010. Developmental chronology of preverbal social behaviors in infancy using the M-CHAT: baseline for early detection of atypical social development. Res. Autism Spectr. Disord. 4, 605–611.

Jones, W., Klin, A., 2013. Attention to eyes is present but in decline in 2 to 6 month old infants later diagnosed with autism. Nature 504, 427–431.

Jones, E.J.H., Gliga, T., Bedford, R., Charman, T., Johnson, M.H., 2014. Developmental pathways to autism: a review of prospective studies of infants at risk. Neurosci. Biobehav. Rev. 39, 1–33.

Kanner, L., 1943. Autistic disturbances of affective contact. Nerv. Child 2, 217–250. Kern, J.K., Geier, D.A., Geier, M.R., 2015. Evaluation of regression in autism spectrum

disorder based on parental reports. N. Am. J. Med. Sci. 6, 41–47. Konopka, G., Wexler, E., Rosen, E., Mukamel, Z., Osborn, G.E., et al., 2012. Modeling the

functional genomics of autism using human neurons. Mol. Psychiatry 17, 202–214. Landa, R., Garrett-Mayer, E., 2006. Development in infants with autism spectrum dis-

orders: a prospective study. J. Child Psychol. Psychiatry 47, 629–638. Landa, R.J., Stuart, E.A., Gross, A.L., Faherty, A., 2013. Developmental trajectories in

children with and without autism spectrum disorders: the first 3 years. Child Dev. 84, 429–442.

Le Couteur, A., Lord, C., Rutter, M., 2003. Autism Diagnostic Interview-Revised (ADI-R). Western Psychological Services, Los Angeles.

Loftus, E.F., Marburger, W., 1983. Since the eruption of Mt. St. Helens, has anyone beaten you up? Improving the accuracy of retrospective reports with landmark events. Mem. Cognit. 11, 114–120.

Lord, C., Shulman, C., DiLavore, P., 2004. Regression and word loss in autistic spectrum disorders. J. Child Psychol. Psychiatry 45, 936–955.

Marrus, N., Kennon-McGill, S., Harris, B., Zhang, Y., Glowinski, A.L., et al., 2018. Use of a video scoring anchor for rapid serial assessment of social communication in toddlers. J. Vis. Exp. 133, 57041. https://doi.org/10.3791/57041.

Muthen, B., 2004. Latent variable analysis: growth mixture modeling and related tech- niques for longitudinal data. In: Kaplan, D. (Ed.), Handbook of Quantitative Methodology for the Social Sciences. Sage Publications, Thousand Oaks, CA, pp. 345–368.

Nordahl, C.W., Lange, N., Li, D.D., et al., 2011. Brain enlargement is associated with regression in preschool-age boys with autism spectrum disorders. Proc. Natl. Acad. Sci. 108, 20195–20200.

Ozonoff, S., Williams, B.J., Landa, R., 2005. Parental report of the early development of children with regressive autism: the “delays-plus-regression” phenotype. Autism 9, 495–520.

Ozonoff, S., Iosif, A., Baguio, F., Cook, I.C., Hill, M.M., et al., 2010. A prospective study of the emergence of early behavioral signs of autism. J. Am. Acad. Child Adolesc. Psychiatry 49, 258–268.

Ozonoff, S., Iosif, A., Young, G.S., Hepburn, S., Thompson, M., et al., 2011a. Onset pat- terns in autism: correspondence between home video and parent report. J. Am. Acad. Child Adolesc. Psychiatry 50, 796–806.

Ozonoff, S., Young, G.S., Carter, A., Messinger, D., Yirmiya, N., et al., 2011b. Recurrence risk for autism spectrum disorders: a Baby Siblings Research Consortium study. Pediatrics 128, e488–e495.

Ozonoff, S., Li, D., Deprey, L., Hanzel, E.P., Iosif, A., 2018a. Reliability of parent recall of ASD symptom onset and timing. Autism 22, 891–896.

Ozonoff, S., Gangi, D., Hanzel, E.P., Hill, A., Hill, M.M., et al., 2018b. Onset patterns in autism: variation across informants, methods, and timing. Autism Res. 11, 788–797.

Palomo, R., Belinchon, M., Ozonoff, S., 2006. Autism and family home movies: a com- prehensive review. J. Dev. Behav. Pediatr. 27 (Supplement), S59–S68.

Parikh, C., Iosif, A., Ozonoff, S., 2018. A longitudinal examination of onset patterns and developmental trajectories among infant siblings of children with autism spectrum disorder. Paper Presented at the Annual Gatlinburg Conference.

Parr, J.R., LeCouteur, A., Baird, G., Rutter, M., Pickles, A., the International Molecular Genetic Study of Autism Consortium, et al., 2011. Early developmental regression in ASD: evidence from an international multiplex sample. J. Autism Dev. Disord. 41, 332–340.

Pearson, N., Charman, T., Happe, F., Bolton, P.F., McEwen, F.S., 2018. Regression in autism spectrum disorder: reconciling findings from retrospective and prospective research. Autism Res. 11, 1602–1620.

Pickles, A., Pickering, K., Taylor, C., 1996. Reconciling recalled dates of developmental milestones, events and transitions: a mixed generalized linear model with random mean and variance functions. J. R. Stat. Soc. Ser. A Stat. Soc. 159 (Part 2), 225–234.

Pickles, A., Simonoff, E., Conti-Ramsden, G., Falcaro, M., Simkin, Z., et al., 2009. Loss of language in early development of autism and specific language impairment. J. Child Psychol. Psychiatry 50, 843–852.

Richler, J., Luyster, R., Risi, S., Hsu, W., Dawson, G., et al., 2006. Is there a ‘regressive phenotype’ of autism spectrum disorder associated with the measles-mumps-rubella vaccine? A CPEA study. J. Autism Dev. Disord. 36, 299–316.

Rogers, S.J., 2009. What are infant siblings teaching us about autism in infancy? Autism Res. 2, 125–137.

Rogers, S.J., Vismara, L., Wagner, A.L., McCormick, C., Young, G., Ozonoff, S., 2014. Autism treatment in the first year of life: a pilot study of Infant Start, a parent-im- plemented intervention for symptomatic infants. J. Autism Dev. Disord. 44, 2981–2995.

Rozga, A., Hutman, T., Young, G.S., Rogers, S.J., Ozonoff, S., Dapretto, M., Sigman, M., 2011. Behavioral profiles of affected and unaffected siblings of children with autism in the first year of life: contributions of measures of mother-infant interaction and triadic communication. J. Autism Dev. Disord. 41, 287–301.

Schultz, R.J., Glaze, D.G., Motil, K.J., Armstrong, D.D., del Junco, D.J., et al., 1993. The pattern of growth failure in Rett syndrome. Am. J. Dis. Child. 147, 633–637.

Scott, O., Shi, D., Andriashek, D., Clark, B., Goez, H.R., 2017. Clinical clues for auto- immunity and neuroinflammation in patients with autistic regression. Dev. Med. Child Neurol. 59, 947–951.

Shoffner, J., Hyams, L., Langley, G.N., Cossette, S., Mylacraine, L., et al., 2010. Fever plus mitochondrial disease could be risk factors for autistic regression. J. Child Neurol. 25, 429–434.

Shumway, S., Thurm, A., Swedo, S.E., Deprey, L., Barnett, L.A., Amaral, D.G., Rogers, S.J.,

S. Ozonoff and A.-M. Iosif Neuroscience and Biobehavioral Reviews 100 (2019) 296–304

303

Ozonoff, S., 2011. Symptom onset patterns and functional outcomes in young chil- dren with autism spectrum disorders. J. Autism Dev. Disord. 41, 1727–1732.

Smith, C.J., Rozga, A., Matthews, N., Oberleitner, R., Nazneen, N., et al., 2017. Investigating the accuracy of a novel telehealth diagnostic approach for autism spectrum disorder. Psychol. Assess. 29, 245–252.

Szatmari, P., Chawarska, K., Dawson, G., Georgiades, S., Landa, R., et al., 2016. Prospective longitudinal studies of infant siblings of children with autism: lessons learned and future directions. J. Am. Acad. Child Adolesc. Psychiatry 55, 179–187.

Tarquinio, D.C., Motil, K.J., Hou, W., Lee, H.S., Glaze, D.G., et al., 2012. Growth failure and outcome in Rett syndrome: specific growth references. Neurology 79, 1653–1661.

Thomas, M.S.C., Annaz, D., Ansari, D., Scerif, G., Jarrold, C., Karmiloff-Smith, A., 2009. Using developmental trajectories to understand developmental disorders. J. Speech Lang. Hear. Res. 52, 336–358.

Thurm, A., Manwaring, S.S., Luckenbaugh, D.A., Lord, C., Swedo, S.E., 2014. Patterns of skill attainment and loss in young children with autism. Dev. Psychopathol. 26, 203–214.

Thurm, A., Powell, E.M., Neul, J.L., Wagner, A., Zwaigenbaum, L., 2018. Loss of skills and onset patterns in neurodevelopmental disorders: understanding the neurobiological mechanisms. Autism Res. 11, 212–222.

Valvo, G., Baldini, S., Retico, A., et al., 2016. Temporal lobe connects regression and macrocephaly to autism spectrum disorders. Eur. Child Adolesc. Psychiatry 25, 421–429.

Veeraragavan, S., Wan, Y.W., Connolly, D.R., Hamilton, S.M., Ward, C.S., et al., 2016. Loss of MeCP2 in the rat models regression, impaired sociability and transcriptional

deficits of Rett syndrome. Hum. Mol. Genet. 25, 3284–3302. Wan, M.W., Green, J., Elsabbagh, M., Johnson, M., Charman, T., et al., 2013. Quality of

interaction between at-risk infants and caregivers at 12-15 months is associated with 3-year autism outcome. J. Child Psychol. Psychiatry 54, 763–771.

Wasilewska, J., Kaczmarski, M., Stasiak-Barmuta, A., Tobolczyk, J., Kowalewska, E., 2012. Low serum IgA and increased expression of CD23 on B lymphocytes in per- ipheral blood in children with regressive autism aged 3-6 years old. Arch. Med. Sci. 8, 324–331.

Werner, E., Dawson, G., 2005. Validation of the phenomenon of autistic regression using home videotapes. Arch. Gen. Psychiatry 62, 889–895.

Werner, E., Dawson, G., Munson, J., Osterling, J., 2005. Variation in early developmental course in autism and relation with behavioral outcome at 3-4 years of age. J. Autism Dev. Disord. 35, 337–350.

Wetherby, A., Prizant, B., 2002. Communication and Symbolic Behavior Scales Developmental Profile–First Normed Edition. Paul H. Brookes, Baltimore.

Wolff, J.J., Botteron, K.N., Dager, S.R., Elison, J.T., Estes, A.M., et al., 2014. Longitudinal patterns of repetitive behavior in toddlers with autism. J. Child Psychol. Psychiatry 55, 945–953.

Xi, C.Y., Ma, H.W., Lu, Y., Zhao, Y.J., Hua, T.Y., Zhao, Y., Ji, Y.H., 2007. MeCP2 gene mutation analysis in autistic boys with developmental regression. Psychiatr. Genet. 17, 113–116.

Zwaigenbaum, L., Bryson, S., Rogers, T., Roberts, W., Brian, J., Szatmari, P., 2005. Behavioral manifestations of autism in the first year of life. Int. J. Dev. Neurosci. 23, 143–152.

S. Ozonoff and A.-M. Iosif Neuroscience and Biobehavioral Reviews 100 (2019) 296–304

304