Autism Spectrum Disorder in Infants and Toddlers An Introduction Fred R Volkmar Katarzyna Chawarska Ami Klin

CHAPTER I

Autism Spectrum Disorder

in Infants and Toddlers

An Introduction

Fred R Volkmar

Katarzyna Chawarska

Ami Klin

In his original report on the syndrome of earlu infantile autism, Leo Kanner (1943/1968) indicated that autism was a congenitial disorder. Although a minority of children seem to develop autism after some months of normal development, most of the subsequent work on autism has generally supported his contention (see Volkmar, Chawarska & Klin, 2005, for a review). Somewhat paradoxically, however, our knowlwdge of autism asit expressed in the first years of life is quite limited.

Fortunately, within in the last year decade ago various factors this situation has begun to change. A little more that a decade ago various factors to delay case detection and early diagnosis (Single, Pliner, Eschler, & Elliott, 1998), but now various programs specifically focused on early diagnosis of infants at  risk for autism have been developed. Growing   public awareness of the condition and an increasingly large body of work on the importance of early intervention and stability of early diagnosis (National Research Council, 2011) have increased interest in the stage of autism. A growing body of research work focused on this age group has begun to apper. In previous years most of this work was based on either parent report (Chawarska, Paul, Et al., 2007; Cohen, Volkmar, & Paul, 1986) or review of videotapes or movies (e,g., Osterling & Dawson, 1994; Werner, Dawson, Oesterling, & Dinno, 2000), with all the attendant problems associated with the lack of contemporaneous methods. The first prospective longitudinal studies of young children (Lord, 1995;Lord et al., 2006) and the recognition of the importance of early intervention have stimulated the National Institute of children with autism through early diagnosis and intervention. In this chapter we are concerned with issues of the clinical expression of autism in infants. Although our major focus is on infancy and early childhood, some of the work on preschool children is highly relevant and is touched upon as well. We  attempt to highlight areas critical for future research on this important topic.

AUTISM AS A DIAGNOSTIC CONCEPT

Kanner’s Original Report

Kanner’s (1943/1968) original report contrasted the lack of social interest (autism) with the normative marked predisposition to engage with others. In reciprocal interactions; he carefully frame his observation developmentally by citing the work of Gesell on the early emergence of social interest in first weeks of life. We now are aware that this interest is present from birth in the typically developing infant. Since Kanner’s first description, the diagnostic concept has undergone modification based on research and clinical work. At the same time, the diagnostic conceptualization retains important historical and conceptual continuities with Kanner’s first description. Kanner emphasized the centrality of the social difficulties, as well as the presence of sameness “or” resistence to change.”  These unusual behaviors included unusual movements and mannerisms as well as problems in dealing with change and novelty. Of the frist 11 patients described in his report, only one was below age 3 years when Kanner frist examined him, and three children were between the ages of  3 and 4.

            Although Kanner emphasized the uniqueness of the conditional and its apparent difference from schizophrenia, other clinicians tended to assume some from of continuity of the two conditions. This issue was clarified over the next several decades as longitudinal and other data made it clear that autism formed a distinct diagnostic category. As a result, however, Kanner’s early focus on “early infantile autism” was lot and most research focused on school-age or adolescent children.

DSM-I and DSM-II

Confusion with childhood schizophrenia

In the frist two editions of the Diagnostic and Statistical Manual of Mental Disorders (DSM) only the term childhood schizophrenia was officially available to describe autism. This situation was very unfortunate. Subsequently, the work of Kolvin (1971) and Rutter (1972) made clear that autism was distinctive and could not simply be considered an early form of schizophrenia (Volkmar & klin,2005). Furthermore, available research suggested that autism was a brain-based disorder and not a result of deviant parent-child interaction. In parallel with attempts to provide better definitions of adult psychiatric disorders for research ( Spitzer, Endicott, & Robins, 1978), similar attempts were made of childhood-onset disorders like autism. Among the investigators of this time, Rutter (1978) provided an important and influential synthesis of Kanner’s original report with subsequent research. Rutter suggested the importance of four essential features: (1) early onset, (2) distinctively impaired social development, (3) distinctively impaired communication, and (4) unusual behaviors of the type suggested in Kanner’s concept of “insistence on sameness” (resistance to change, idiosyncratic responses to the environment, motor mannerisms and stereotypies,etc.). Rutter was clear that the social and communication difficulties were not just a function of associated intellectual disability. These various issue were considered as autism was first included in the landmark, third  edition of DSM (DSM-III; American Psychiatric Association, 1980).

DSM-III and DSM-III-R

DSM-III (American Psychiatric Association, 1980) represented a marked change from its two predecessors. The taxonomy proposed was  based on research findings and emphasized the importance of an atheoretical and empirically based set of criteria. Autism was included in a newly designated class of childhood-onset disorders, Pervasive Developmental Disorders (PDD). A “subthreshold” condition was included as well, atypical PDD; this term had considerable (if unintended) overlap with earlier terms such as atypical personality development (Volkmar & Klin, 2005). The definition included in DSM-III was heavily dependent on Rutter’s earlier conceptualization and provided for a clear differentiation of autism from schizophrenia. Interestingly, the original DSM-III  approach lacked a developmental orientation and, if anything, the criteria proposed were much more oppropriate to very young children with autism, that is, consistent with the term infantile autism. Although the use of a multiaxial  approach was a clear benefit fir child psychiatry, some aspects of the organization of this system were confusing- for example, autism and related disorders were placed on a different axis than other developmental disorders. A much more developmental orientation  was introduced in DSM-III-R (American Psychiatric Association, 1987), which was greatly influenced by work of Lorna Wing (Wing & Gould, 1979). Although the now familiar three major areas of dysfunction were still included, the new criteria were much more detailed and include  a range of examples (with the goal of producing an approach applicable to the broad range of age and developmental levels.) The use of a polythetic apporoach was also adopted, and the requirement for early onset was dropped (although onset before or after age 3 could still be specified). The official name of condition was changed from infantile autism to Autistic Disoder in reflection of these changes. Although many aspects of the DSM-III-R approach were improvements, it quickly became apparent that the system tended to “ overdiagnose” autism, particularly in the  cases of more intellectually challenged children (Rutter & Schopler, 1992). This observation led to the potential for major difficulties in the comparison of studies using different diagnostic criteria and also posed problems for pending revision in the International Classification of Dieseases-tenth edition (ICD-10); World Health Organization, 1990). The ICD and DSM approaches  are fundamentally related and share many aspects of diagnostic coding, although there are also important differences.

ICD-10 and DSM-IV

Extensive revision of both the ICD and DSM systems was undertaken early in 1994. As part of the DSM-IV revision process (American Psyhiatric Association, 1994), attempts were made to indentify areas of both consensus and controversy such as clinical utility, reliability, and descriptive validity of categories and criteria. Coordination with the pending ICD revision was also a consideration. Literature reviews and data reanalyses were also undertaken for specific issues, such as those relative to the concept of Childhood Disintegrative Disorder- a concept included in previous version of ICD but not DSM. Data reanalyses suggested that the DSM-III-R approach was overbroad, and a decision was made to undertake a large multinational fieled trial (Volkmar et al., 1994). This fieled trial was conducted in coordination with the ICD-10 revision effort and included more than 100 raters working at more than 20 sites around the world. The final sample included information on nearly 1,000 cases seen by on (or sometimes more than one) rater. In the nearly 1,000 cases, more than 300 children were less than 5 years of age (although most were between ages 3 and 5 and no child younger than 2 was seen). A standard coding system was used to provide basic information on case and rater and on a number of diagnostic criteria.

            The overall result of the field trial (see Table 1.1) confirmed that DSM-III-R had a hinher sensitivity but lower specificity, whereas the ICD 10 draft definition, designed to be a research diagnostic system, had, as expected, higher specificity. A series of analyses were undertaken, includeing realiability of criteria and diagnosis, factor analyses, signal detection analysis, and so forth (Volkmar et al., 1994, klin, Lang, Cicchetti, & Volkmar, 2000). As expected, social criteria were, as individual diagnostic items, generally the most potent single diagnostic predictors, and a decision was made to weigh them more heavily in the final DSM-IV definition. Possible modification in the ICD-10 system were examined, the goal being to have convergent definitions in the DSM and ICD. The final diagnostic approach provided reasonable converage over the range of syndrome expression in autism as reflected in the filed trial sample and was applicable from early childhood (i.e.,at about age 3) through adulthood.

            It must be emphasized that the DSM-IV and ICD-10 approach did consider development aspects of syndrome change, but not surprisingly at that time, the focus was not on infants and very young children; that is, it appeared that the approach derived worked satisfactorily starting at about age 3. Interestingly, examination of some of the DSM-IV field trial data ( children under age 5) reveals a few items with stronger developmental correlates. In general, such items were discarded because they would not be applicable to the entire range of syndrome expression. For example, attachment to unusual objects has low sensitivity (.50) but high specificity (.90), so that when it is observed, it has high predictive power for autism but only in this younger age group.

            Interest in the earliest development of children with problems included in the autism spectrum was also fueled by inclusion of additional.

#Tabel hal 5

Disorders within the revised PDD section of DSM-IV (e,g., Asperger’s Disorder, Rett’s Disorder and childhood Disintegrative Disorder). A need to differentiate these disorders highlighted the importance of understanding development history and early clinical presentations.

            At the time that DSM-IV appeared (1994), “these was little concern with the manifestation of autism in infants and very young children. For children by about age 3, the DSM system appeared to generally work well with reasonable stability of diagnosis (Lord & Risi, 2000). However, with the growing interest in genetic mechanism, screening of at-risk populations such as siblings, and the marked increase in research in the earliest manifestation of autism, there has been progressively more concern about autism as it is manifested in infancy. We consider these issue before returing to the problem of early diagnosis.

CLINICAL PHENOMENOLOGY

Onset of the condition

As noted, Kanner (1994/1968) emphasized the apparently congenital nature of autism in his original report. Direct evidence regarding the actual onset of the symptoms is still lacking, and a vast majority of the current reports rely on parental recollection regarding the age of onset and type of frist abnormalities. Although these report s have their obvioud limitations and the onset of parental concern is likely to follow the actual time when the symptoms of autism spectrum disorder (ASD) (equivalent to the term PDD) begin to manifest, they also offer some insight into the nature of the first concerns that likely to motivate parents to seek professional advice, which in turn may lead to an earlier initiation of treatment. Raising parental awareness of the first signs of various developmental disorders, including ASD, has become one of the priorities of a number of parent organizations, such as Autism Speaks (www.cdc.gov/ncbddd/autism/actearly/) , as one of the factors that are likely to contributr to early identification and treatment of infants with developmental disabilities.

            A number of studies have suggested that the vast majority of parents of children with ASD first notice abnormalities during the course of the first 2 years life (Baghdadli, Picot, Pascal, Pry, & Aussilloux, 2003; Chawarska, Paul, et al., 2007; De Giacome & Fambonne, 1998; Rongers & DiLalla, 1990; Tolbert, Brown, Fowler, & Parsons, 2001; Volkamar, Stier, & Cohen, 1995). The first concerns arise  on average, in  the second years, usually at about 14 month (Chawarska, Paul, et al., 2007), 17 months (Baghdidli et al.,2003), or 19 months (De Giacomo & Fambonne, 1998). These ages are likely to be sensitive to several factors, such as the time elapsing between the onset of parental concerns and the time when the information was collected. With a shorter lag, reports of earlier ages of onset are to be expected; otherwise a “Forward-telescoping” effect seems mate regarding the age when the child began manifesting first symptoms to later ages.

            The time when parents begin to notice the first abnormities varies, such that 30-50% of parents report concerns in the first year of the child’s life and 80-90% by the second birthday (Baghdadli et al., 2003;Chawarska, , Paul, et al., 2007; De Giacomo & Fambonne, 1998; Volkamar et al.,1985) there are relatively few studies reporting on the association between clinical outcome and the onset of parental concerns. Most of the studies were conducted retrospectively and produced very mixed results. A recent study examined prospectively the link between the onset of parental concerns, measured when the toddlers were between 18 and 36 month, and clinical diagnosis at the age 4 (Chawarska, Paul, et al., 2007). Children who were identified by their parents as having problems nosed with autism than with Prevasive Developmental Disorder-Not Otherwise Specified (PDD-NOS). However, those identified by parents as having difficulties between 11 and 18 month were equally likely to receive a diagnosis of autism or PDD-NOS at 4 years. Finally, all children in the group with concerns arising at or after 18 month received a diagnosis of tionship between the age of parental recognition (and presumably the onset of symptoms) and clinical diagnosis assigned 2-3 years later and raises a question of possible variants that manifest differently in the onset of symptoms.

            Among the most common and often first noted concerns are delays in speech and language development, followed by an abnormal social responsivity level, medical problems, and nonspecific difficulties related to sleeping, eating, and attention (Chawarska, Paul, et al.,2007; De Giacome & Fambonne, 1998). Notably, in young children, the apprearance of stereotyped behaviors, motor mannerisms, and unusual interests rarely trigger parental concerns, most likely because of their relatively mild manifestations in infancy or a later onset. Although concerns regarding the development of speech and the level of social engagement are frequent for toddlers with autism and PDD-NOS, the nonspecific concerns related of feeding, eating, and sleep apper to be more frequent for toddlers with PDD-NOS (Chawarska, Paul, et al.,2007).

            Although the presence of specific delays constitutes a strong basis for parental concerns, such concerns may also emerge in response to unusual variations in the rate of progress, such an apparent slowing of development (e.g.,if babbling is not followed by the emergence of the first words) or a los of previously acquired skills (regressions) (Siperstein & Volkamar, 2004). Regression is usually reported in 20-35% of cases (Chawarska, Paul, et al.,2007;Goldberg et al.2003;Luyster et al., 2005; Rapin & Katzman, 1998;Ronger,2004;Werner&Dowson,2005) and can involve  the loss of words, vocalizations, nonverbal communication skills (e.g.,eyecontact, gestures, social dyadic interaction skills, imitation, or pretend play (Davidovitch, Glick, Holtzman, Tirosh & Safir, 2000; Goldberg et al.,2003; Luyster et al.,2005). The perception of regression appears to be specific, though clearly not universal, to ASD (Luyster et al.,2005; Siperstein & Volkamar, 2004). Parental reports of regression do not necessary indicate normal development prior to the perceived loss of skills, nor do early abnormalities preclude regression (Lord, Shulman & DiLavore,2004; Siperstein & Volkamar, 2004; Werner & Dawson, 2005;Wilson, Djukie, Shinnar, Dharmani, & Rapin, 2003). In fact, unequivocal loss of skills following normal developmental milestones is relatively uncommon (Siperstein & Volkamar, 2004). However, it is clear that in some cases a marked regression does occur-such regression  has been documented in very young children with ASD through analysis Of video recording in the first years of life  (Werner & Dawson, 2005). Werner and Dowson (2005) used home videotapes of the first and second birthday parties of children with ASD and of typically developing controls. Reters  blind to diagnosis and history of regression confirmed regression, as defined by a decline in frequency of joint attention acts and word/babble use in a subset of the ASD sample.

             It is clears that skills loss after a prolonged period of normal development (e.g., to 3 or 4 years) is relatively uncommon. A specific diagnostic term, Childhood Disintegrative Disorder, exist for this category of cases, and the outcome appears to be worse than that in autism, with little or no recovery of previously exhibited abilities (Volkamar, Koening, & State, 2005). Given the complexities of understanding the role of regression in autism, it remains unclear as to what relationship exist between this less  common later-onset condition and reported early regression in autism.

            Among the factors that precipitate the onset of parental concerns are concurrent cognitive delays, delays in motor development, and the presence of medical problems (De Giacomo & Fombonne, 1998). The presence of perinatal complications and sensory deficits has also been associated with earlier recognition (Baghdadli et al., 2003). A more recent study suggest that in the first year, late onset of social smile, delays in responsivity to speech and language understanding, and late onset of independent walking are possible factors precipitating parental concerns (Chawaraska, Paul, et al., 2007). Factors that have not been found to influence the age genetic factors in autism and increased risk for ASD in younger siblings of the affected to earlier recognition of developmental problems (Klin et al., 2004; Zwaigenbaun et al.,2007).

Clinical Presentation in the First Years of Life

Kanner’s original report emphasized the central role of social difficulties in autism. It is tribute to his powers of observation that most subsequent research has supports this observation, albeit with considerable refinement (Carter, Davis, Klin & Volkamar;2005;see also Chawarska & Volkamar, 2005, for a review). Although early repots on symptoms of autism in the first year of life relied heavily on parental report (e.g., Dahlgren &Gilllberg, 1989; Klin, Volkamar & Sparrow, 1992) and single case studies (Dawson, Oesterling, Meltzoff & Kuhl, 2000), these reports were later supplemented by analytic studies of home video recording depicting, for instance, a first birthday party or other family events (e.g., Baranek,1999; Maestro et al., 2001; Oesterling, Dawson & Munson, 2002; Werner et al.,2000). Studies based on these approaches have contributed greatly to raising awareness, regardless of possible early sympotoms of ASD. However, they suffer a number of important methodological limitations related, for instance, to parental ability to detect and report on the more subtle and contextualized symptoms of ASD (Chawarska, Klin, Paul, Volkmar, 2007;S Stone, Hoffman, Lewis, & Ausley, 1994) as well as to the sensitivity and specificity of the deficits to ASD owing to issues with control groups or the representativenesss of the source material (i.e., vidiotapes) (see also Zwaigenbaum et al.,2007 for a review).

            More recently, however, the finding of increased genetic liability for ASD in younger children enable researchers to study ASD in statu nascendi by following prospectively large cohorts of younger siblings at risk for developing the disorders (Zwiegenbaum et al.,2007). The sibling recurrence rate of autism has been estimated between 3 and 8% (Bailey et al.,1995;Bailey, Philips & Rutter, 1996; Ritvo et al., 1989). These numbers may underestimate the true recurrence rate for several reasons, including (1) increased prevalence rates related to the employment of more inclusive diagnostic criteria for autism and PDD-NOS since the advent of DSM-IV, and (2) the stoppage phenomenon exemplified by a high number of families avoiding futher pregnancies once of offspring is diagnosed with autism (Jones & Szatmari, 1988;Slager, Faround, Haghighi, Spence & Hodge, 2001). Increased rates for nonautistic PDD in siblings (Asperger Syndrome, PDD-NOS) have also been reported in 15-45% of family members (Bailey, Palferman, Heavey & Le Couter 1998; Folstein et al.,1999),  with higher rates of both narrow and board autistic phnotype in male rather than female relatives of individuals with autism (Bolton et al., 1994; Pickles et al.,1995; Piven, Palmer, Jacobi, Childress & Arndt, 1997). Preliminary findings from ongoing studies on high-risk siblings suggest that 20-25% of younger siblings of children with autism may exhibit developmental impairments in the first or second year of life (Zwaigenbaum et al., 2005), thaough studies examining the developmental trajectories of younger siblings are clearly needed and are slowly emerging (Lnda % Garrett-Mayer, 2006; Yirmiya et al., 2006).

            Analysis of videotapes suggests that as compared with typical controls, infants who were later diagnosed with ASD were less likely to look at and seek other people, and they were less likely to smile and vocalize at others in the first year, infants later diagnosed with ASD might show difficulties in responding when their names were called and look at others less frequently, as compared with typically developing children oe infants with developmental delays (Baranek, 1999; Oesterling et al.,2002;Werner et al., 2000). However, as recent prospective study of high-risk infants suggests, limited response to their names at 12 month, although quite specific to infants with ASD as well as high-risk siblings with developmental delays, is by no means universally parents in all infants who are later diagnosed with the disorder (Nadig et al.,2007). Thus, failure to respond to his or he by name may be an indicator that a 12-month-old child would benefit from futher evaluation, but passing the “name-calling” test does not mean tnat the child is not at risk of developing ASD. Studies of the presence of unusual sensory behaviors and motor stereotypies in sample of children with ASD, as compared with children with developmental delays, yield mixed result. Although some suggest the presence of excessive mouthing and possibly aversion to social touch (Barenek, 1999; Loh et al., 2007; Oesterling et al., 2002), others fail to detect similar effects. Furthermore, motor stereotypies have been reported in some sample (Loh et al., 2007; Oestreling et al., 2002) but not in others (Baranek, 1999; Werner & Dawson, 2005).

            Presently the vast majority of prospective baby sibling studies report on the expression of the broader autism phenotype that can be detected in infant siblings who are not affected with a full-blown ASD, rather than in children who were actually diagnosed with ASD later on (e.g., Toth, Dawson, meltzoff, Greenson & Fien, 2007; Cassel at al., 2007; Merin, Young, Ozonoff & Rongers 2007; Gamliel, Yirmiya & Sigman, 2007). This current trend is related to the fact that owing to a relatively low recurrence rate among sibilings, very large longitudinal samples need to accumulate for certain research questions to be addressed. Nonetheless, the first experimental studies reporting on the presentation of infants with ASD  in the first years of life are beginning to emerge. Prospective studies of infant siblings, followed from 6 to 24 or 36 months and identified as having some from of ASD, suggest that robust behavioral features of ASD that could be captured through standard assessment instruments such the autism Observation Scale for Infants (AOSI; Bryson, Zwaigenbaum, McDermott, Rombough & Brian, 2007)  and the Mullen Scales of Early Learning (MSEL; Mullem, 1995) may not emerge until some time after 6 months and before 12 months, with further intensification of their expression occurring between 12 and 24 months (Bryson, Zwaigenbaum, Brian, et al., 2007; Landa & Garrett-Mayer, 2006; Zwaigenbaum et al., 2005). Zwaigenbaum and colleagues (2005) identified several features at 12 month that are likely to differentiate siblings with ASD from those without social disability. Among the features were poor eye contact, limited social interest and smiling, limited use of gestures, poor response to name, poor imitation, and delays in receptive and expressive language. These infants also exhibited temperamental abnormalities, including initial passivity in early development followed by the emergence of a tendency for extreme distress reactions by 12 months. Difficulties in disengagement of visual attention were also noted. Studies such as these constitute the first step toward establishing clear diagnostic criteria for ASD in the first years of life, although extensive studies are needed to establish both sensitivity to and specificity of the identified abnormalities.

            A complementary approach to identifying behavior markers of ASD in infancy involves the employment of experimental designs targeting basic perceptual and cognitive processes involved in development of social interaction and communication. Among these are eye-tracking studies of perception of social and nonsocial stimuli (e.g.,Chawarska & Shic, 2007; Klin & Jones, in press; Merin et al., 2007) and speech perception (Nadig et al., 2007). These studies are discussed detail by Klin, Saulnier, Chawarska, and Volkamar (Chapter 6, this volume).

Symptoms of ASD in the Second and Third Years of Life

Several factors have contributed to much large body of data on autism as it manifest after the first birthday and before age 3. Recent advances in clinical research suggest that in 2 and 3 years- old, symptoms of autism center on areas of social interaction and communication and are often accompanied by delays in multiple areas of functioning, including motor and nonverbal cognitive development (see Chawarska &Volkamar, 2005 for a review; see also Bishop, luyster, Richler, & Lord, Chapter 2 Chawarska &Bearss, Chapter 3; and Paul, Chapter 4 for this volume). In the social domain, the most frequently reported symptoms are diminished aye contact, limited inters in social games and trun-taking exchanges, low frequency of looking referentially at parents, and preference for being olone (Cox et al., 1999; Lord, 1995; Stone, Lee, et al., 1999). Vocal and motor imitation and symbolic play skills appears delays as compared with the children’s overall developmental  levels (Baron-Cohen, Cox, Baired, Sweettenham & Nightingale, 1996; Cox et al., 1999). Young children with  autism direct their visual attention more frequently toward objects than toward people (Dawson et al., 2004; Swettenham et al.,1998). A limited range of facial expressions and infrequent instances of sharing affect (e.g., by smiling and looking at others) have been reported as well (Cox et al., 1999; Lord, 1995; Stone, Lee, et al., 1999). In the areas communicative exchange through nonverbal (e,g., use of gestures or gaze to communication interest or joint attention) and vocal or verbal means. The child’s responsivity  to speech in general, and  to his or her name particular, continues limited (Baron-Cohen et al., 1996; Cox et al., 1999;DiLavore, Lord & Rutter, 1995; Klin 1991; Lord & Pickles, 1996; Paul, Chawarska, Klin & Volkamar 2007). Vocalizations may take on an abnormal quality (Sheinkopf, Mundy, Oller & steffens, 2000; Wetherby, Yonclas, & Bryan, 1989). Stereotypic and repetitive behaviors reach a clinical threshold in the second years in some children (Chawarska, Klin et al., 2007), and in a vast manjority of children by the age of 4 (e.g., Lord, 1995). Adaptive skills are usually delayed beyond what would be expected based on the developmental level (Klin et al., 1992; Stone, Ousley, Hepburn, Hogan, & Brown, 1999).

            The relative mild expression of the unusual repetitive behaviors (stereotyped movements and mannerisms) and the general category of “resistance to change” behaviors in this age group is of some interest (e.g., Chawarska, Klin et al., 2007; Loh et al., 2007;Lord, 1995). The absence of clears- cut behaviors in this general category is one of the more general conceptual problems in the application of categorical (DSM-IV or ICD-10) diagnostic criteria. In Lord’s longitudinal study the absence of such behaviors before age 3 was a frequent reason that a diagnosis of autism could not be made (Lord, 1995; Lord et al., 2006). Although clear precursors of such behaviors may potentially be used as alternatives for this age group, relatively few attempts have been made to identify such precursors (Loh et al., 2007) and assess their specificity in visually repetitive phenomena (e,g., ceiling fans), and overattention to the nonsocial environment (focusing on alphabet latters on blocks or small details or play materials) are potential candidates. Furthermore, an increase in the second years, rather than the expected decrease, of some of the repentitive movements observed in the first year (Thelen, 1979) may be a sign of abnormal development in this areas (Loh et al., 2007).

IMPLICATIONS FOR DIAGNOSIS AND SCREENING

Issue of diagnosis and screening are discussed in detail by Bishop et al. (chapter 2 this volume) and are only briefly touched upon here. Clearly by about age 3 (and often even before) the current DSM-IV/ICD-10 categorical approach can be used with little difficulty. Available work does highlight some limitations of their criteria for very young children (Stone, Lee, et al., 1999). An alternative categorical classification (National Center for Clinical Infant Programs (NCCIP), 1994) has been proposed, but its utilization in the clinical community has been limited, probably because its nosological research in autism. Thus, there is little information on its concurrent validity with DSM-IV and related literature. because  the history of this system precedes the current wave of nosological effeorts related to children under the age of 3 years, it would be critical for the NCCIP (now Zero to Three) system to be properly researched and its clinical and concurrent validity (relative to other systems), reliability, and other psychometric properties to be adequately assessed.

            More generally, well-documented diagnostic instruments may work well after age 3-4 years or past a certain development level (often around 18 months), but their use is not clearly established for the first years of life. Dimensional assessment instruments have a number of potential advantages-for example, in their approach to developmental change and/ or developmental level-and may be of particular use, given the greater potential for change in this age group. Similarly, screening approaches (see Bishop et al., Chapter 2, this volume) are particularly important in terms of identification of children in need of services but present their own issue in terms of design and evalution. Unfortunately, what is critically needed, but not yet available, are methods that rely on biological markers or some other robust, readily measured indicator of risk. Given the lack of such makers, clinician-assigned diagnosis, as provided by experienced clinicians, remains the “gold standart” for diagnosis in infancy (Chawarska, Klin, et al., 2007; Cox et al., 1999; Gillberg et al., 1990; Lord, 1995; Stone. Lee, et al., 1999).

SUMMARY AND CONCLUSIONS:

THE SIGNIFICANCE OF EARLY CASE DETECTION

The growing body of work on autism in infants is important for several reason. Available data suggest that earlier case detection the out come of autism is gradually improving; for example, more and more individuals are able to live indenpendently and fawer are likely to remain mute and to exhibit comorbid intellectual disability (Howalin, 2005). The recent National Research Council (2001) review of evidence on early treatment notes that, despite various limitations, a considerable body of work on the important for treatment and log-term outcome, early detection is also important in clarifying the earliest developmental processes, which may be disrupted in autism.

            Prospective research is critically needed to help us to more fully understand basic mechanism of psychopathology  and to clarify how early difficulties become entrained in subsequent development. Somewhat paradoxically, those who work with both higher-functioning older indivuals with autism and very young infants are impressed not only by the potential for significant development change, but also by the severty and continuity of difficulties across time and developmental- for example, in modulation of the human voice in prosody and in the use of eye contact to mediate social interaction (Paul, Augustyn, Klin, & Volkmar 2005). The ability to observe these early processes without the accompanying overaly of subsequent development will be particularly important. Study of the range of early developmental skills in this population may also result in some clinical surprises; for example, there is now a suggestion that for a subgroup may be the more striking initial signs of autism rather than disturbances in social interaction (Bryson, Zwaigenbaum, Brian, et al., 2007).

            Consistent with kanner’s (1943/1968) orginal description, it ap[ears that in many cases infants are born with autism. It is also clear that in a variably reportand, apparently small number of cases, the child develops reasonably normally for a time before autism appears. Although much work remains to be done, it is possible  even now to begin to understand how some of the early manifestations of autism become antrained in subsequent development. Data from this age group may shed important light on perplexing clinical question-for instance, the well established differences in gender ratio and severity may be apparent before age 3 years (Certer et al., 2007). Careful follow-up studies also emphasize the potential difficulties of early diagnosis (Sutera et al., 2007), futher underscoring the importance of biological markers and the study of specific biological and neuropsychological processes for batter early diagnosis. To this end, the study of very specific social processes under highly controlled conditions may be particularly important (e,g., Chawarska, Klin & Volkmar, 2003; Chawarska & Volkmar, 2005; Chawarska & shic, 2007; Klin, 1992; Klin & Jones, in press; Klin, Jones, Schultz, Volkmar & Cohen, 2002; Merin et al., 2007; Presmanes et al., 2007). As such processes are identified, siblings can also be studied to address potential contributions of these processes to the broader autism phenotype (Cassel et al.,2007; Presmenes, Walden, Stone & Yoder, 2007; Toth et al., 2007; Gamliel et al., 2007).

ACKNOWLEDGMENTS

Preparation of this chapter was supported in part by grants from the National Alliances for Autism Research/Autism, Speaks and the National Institute of Mental Health (Grant No. U54 MH676494) to Fred R. Volkar, Katarzyna Chawarska, and Ami Klin.

CHAPTER 2

Diagnostic Assessment

SOMER L. BISHOP

RHIANNON LUYSTER

JENNIFER RICHLER

CATHERINE LORD

Autism is a neurodevelopmental disorder characterized by deficits in social reciprocity and communication and by the presence of restriced and repentitive behaviors and/or interests. According to the criteria out lined in the Diagnostic and Statistical Manual of Mental Disorders (DSM-IV; American Psychiatric Association, 1994) and the International Classification of Diseses (ICD-10;World Health Organization, 1992), in order to receive a diagnosis of autism, a child must have shown abnormalities, or symbolic/imaginative paly before the age of 3 years. If a child does not meat all of the above criteria for autism, her or she may given a diagnosis of Asperger Syndrome (AS) or Pervasive Developmental Disorder-Not Otherwise specified (PDD-NOS). we refer  to these three diagnoses together as autism spectrum disorder (ASD).

            Because there is not yet a biological marker for ASD, a diagnosis of ASD is made on the basis of a behavioral profile, which is characterized by both the absence of typical behaviors as well as the presence of atypical behaviors. Recently, researchers and clinicians have sought to identify ASD earlier and earlier, owing to finding that early intervation is associated with improved outcomes (Harris & Handleman, 2000). This is somewhat problematic, however, because whereas the behaviors features of ASD are well established for children in the preschool years and beyond, less in know about symptom presentation in the first 2 years of life (Zwaigenbaum et al., 2005; Mitchell et al., 2006). Indeed, DSM-IV criteria were established based on the profile exhibited in early and middle childhood and do not necessarily apply to children under the age of 3. Therefore, professionals should exercise caution when making diagnoses in very young children. Furthermore, any assessmention of other disorders of  early childhood. Because ASD is a developmental disorder and different symptoms are diagnostic at different points in development, understanding what is developmentally appropriate for children under 3 is an important first step in early identification of the disorder.

            This chapter addresses issues in the assessment and diagnosis of ASD in onfants and toddlers. The first section provides a brief summary of the development of social, communication, and play behaviors in typically developing young children. Next, we provide guidelines for assessment and differential diagnosis of children with ASD, including the importance of considering social, communication, and play behaviors in the context of a child’s overall developmental functioning. Finally, we review the currently available screening instruments for identifiying ASD in infants and toddlers, with special attention given to their appropriateness and liniatations for use with children under  3 years of age.

EARLY TYPICAL DEVELOPMENT

The impairments that result from ASD are defined in relation to typical development. Reciprocal interaction and communication difficulties involve dificits in behaviors that emerge in typically developing children without explicit teaching and that are adaptive their social context. In the area of restricted and repentitive behaviors and interest (RRBs), impairement refers to the presence of unusual, sometimes maladaptive behaviors that are, at least according to commom wisdom, not usually seen in typically developing children. Thus, in order to determine if a child is showing signs of ASD, it is crucial to have a clear understanding of what constitutes typical behavior in a child of the same developmental level. As more and more children are being referred for a diagnosis of ASD at very young ages, it has becomes particularly important to have a comprehensive picture of social and communicative behaviors in typically developing infants and toddlers. This understanding can help clinicians and researchers avoild overdiagnosing autism as well as wrongly dismissing real, appropriate concerns about behaviors associated with ASD.

            A large body of evidence suggest that children come into the world already socially oriented and that their social understanding becomes richer and more sophisticated in relatively short period of time. Newborns prefer looking at faces over nonface patterns (Valenza, Simon, Cassia, & Umilta, 1996) and prefer listening to speech over nonspeech sound (Vouloumanos &  Werker, 2004). Meltzoff and Moore (1989) have shown that newborns can imitate simple human gestures, such as tongue protrusions and head movements, and by just 6 weeks ago of age they can engage in deferrend imitation, emulating others facial movements after a 24-hours delay (Meltzoff and Moore, 1994). Children as young as 6 months can distinguish between purposeful and nonpurposeful action  (Woodwards, 1999) and by 9 months of ag they are able to follow and direct the attention of adults to outside, a capacity know as joint attention (Tomasello, 1995). At approximately 12 months of age, infants begin to angage in social referencing byy using the emotional reactions of others to determine how to behave (Walden & Ogan, 1988). At approximately 18 months, toddlers can infer an adult’s intended action by watching failed attempts (Meltzoff, 1995). By 24 months age, children adjust the language they use in conversation based on their understanding of what the listener knows (Tomasello, Farrar & Dines, 1984).

            This progressions shows an increasing understanding of others’s intentions in the first 2 years of life. Tomasello (1995) has proposed a developmental trajectory for the understanding of other’s intentions, whereby children progress from following and directing the attention of others without understanding their intentions, to understanding other’s as international agents, to learning that other’s intentions may not always match the situation. This work has recently been expanded to suggest that these early developments culminate in the understanding of shared intentions with another individual, which is believed to be a defining feature of human social interaction (Tomasello, Carpenter, Call, Behne & Moll, 2005). Understanding of other’s intentions is also thought to underline the ability to learn the referent of novel word, he or she must  infer the referential intent of the speaker, using subtle cues such as the direction of the speaker’s gaze and other contextual clues (Baldwin, 1993).

            One of the most remarkable aspects of early development is it rapid pace. In a relatively short period of time, children’s understanding of the social world becomes quite sophisticated. Yet it is important to remember that there is a great deal of variability in early trajectories of social and communication development. Fenson et al. (1994) emphasize the importance of going beyond descriptions of the “Modal Child” in order to understand the range of variability that can be expected in typically developing children.

            For example, there is a great deal of variability in both early receptive language and expressive vocabulary development. As children get older, this variability increases, because children whose initial language is more advanced also shoe a higher rate of word acquisition. When considering variability in early social and communication development, then, it may be useful to examine not only differences in children’s abilities at a given point in development, but also differences in the developmental trajectories of these abilities over time.

            Marked individual differences have been found in other areas of communication, such as use of gestures, as well as in the development of social cognition. In a comprehensive study of children’s social and communication development from 9 to 15 months of age, Carpenter, Nangell, Tomasello, Butterworth, and Moore (1998) found considerable variability in attention and gaze following, imitation, gesture production, and joint angegement, among other skills. Although the manority of the children in their sample displayed these skills by the age of 12 months, some children acquired these skills earlier than others (e.g., as young as 9 months) and some children had not acquired certain skills 15 months.  

            Some researchers have argued that differences in child temperament might explain some of the variability in early social and communication development. Dixon and smith (2000) found that temperament exhibited in early development was related to subsequent langue skill, both receptive and expressive. In their sample, children who showed greater adaptability, more positive mood, and greater persistence at 13 months tended to have more productive language at 20 months, and children who had long durations of orientation, smiled and laughed frequently, or were easily soothed at 7 months tended to have advanced comprehension at 7 and 10 months of age.

            Other studies have found similar relationships between children’s early temperament and later language (Slomkowski, Nelson, Dunn, & Plomin, 2992). Dixon and Smith (2000) suggest that the relationship between temperament and language may be mediated by amount of joint engagement. This is, parents and others may be less likely to enter into a social exchange with a child who shows negative affect and poor adaptability than they would with a child who shows positive affect. The reduce amount of social interaction may in turn adcersely affect the chhild’s understanding and production language. This model partly supported by a finding in the study by Carpenter et al., (1998) that the amount of time mother-infant dyads spent in joint engagement was related to the child’s early verbal non verbal communication skills.

            The fact that there is a wide range of social and communication skills among typically developing young children presents a challenge to those trying to identify “Markers” of ASD in children of this age. How does one decide if a toddler has a true puts him at risk for a diagnosis of ASD, or and/ or communication that puts him at risk for a continuum of typical whether he simply falls on the lower and of the continuum of typically development in these areas? Is a toodler who does not smile at others very often simply showing less positive affect than the “average” typically child of the same age because of her temperament, or does she have a more fundamental difficulty in interacting with others? It is also important to consider the role that culture differences play in a child’s social and communication behaviors (see babad et al., 1983). The child’s social environment, including culturally based parenting practices, is likely to influence some of the aspects of infants social communication behaviors.

            Additional insight into this issue may come from considering “constellation” of deficit, rather than individual impairments. ASD is commonly  though of as involving deficits in several different areas (Siegel, Pliner, Eschler & Elliott, 1988). It may be that, order to be considered “at risk” for ASD, a child should be showing deficits in more than one of these areas. Therefore, when considering a particular social or communicative behavior in young child, it may important to consider whether the child is “below average” or “impaired” in a specific behavior, but also whether difficulties in that behavior occur in the context of other impairments. For example, a child who shows delays in using sounds and words but who shows positive affect, good eye contact, and use of early gestures would likely not elicit much concern as a child who, in addition to having delayed expressive language, shows impairments in others areas. Practitioners may also want to consider these issue when conducting evaluations of slightly older children. Despit the minimum onset requirement presented by DSM-IV-delay or abnormality in social interaction, or in language as used in social communication, or in imaginative play to age 3-practitioners may want to require that types of early atypicalities occur in conjuction with one another in order to establish the onset of ASD.

            Even with these destinctions in mind, it can be difficult for clinicians and researchers to determine whether a child who shows two or three developmental difficulties falls somewhere on the spectrum, is developmentally delayed but not on the spectrum, or is simply behind relative to the “average” typical  child, but still within the range of typical development. PDD-NOS is usually diagnosed in young children who show some impairments is fewer than that required for a diagnosis of autism, the impairments do not occur across all there areas specified (i.e., social, communication, RRBs), or the impairments are not as severe. This could explain why studies have found that the majority of children diagnosed with PDD-NOS at age 2 remained on the autism spectrum at age 9. Nevertheless, more than 100% of the children with PDD-NOS diagnoses at 2 years moved to a nonspectrume classifications by 5 or 9 years (Lord et al., 2006). Given these criteria, it is possible to see how some young children thaough to have “ mild autism” might, at older ages, more clearly appear to have nonspectrume development delays or fall at the lower end of the continuum of typical development.

            RRBs differ from most social and communication deficits required for a diagnosis of ASD, because they involve the presence of “atypical” behaviors rather than the absence of typical ones.  Yet it is important to remember that some RRBs are actually seen in young children with typical development. Thelen (1979) reported motor stereotypies, such as kicking, waving, bangin, rocking and bouncing in normal infants in their first year of life, especially between the ages of 24 and 42 weeks. As toddlers and preschoolers, many typical young children display compulsive like behaviors, such as insistence on sameness in their routines and/or environment, strong likes and dislikes, a rigid idea of how thighs should transition, such as at bedtime. Evants et al., (1997) found evidence for two kinds of compulsive-like behaviors in a substantial portion of young children: “Just right” behaviors (e.,g.m Lining objects up or insistence on sameness (e.g.,Preferring to have the same schedule every day). It has also been argued that some repetition in object use or exploration (i.e., Piagetian secondary circular reactions) is important for developing cognitive skills, such as problem solving.

            It is interesting to note the similarities between these behaviors and those considered to be “restricted and repetitive behaviors and interests” (American Psychiatric Association, 1994) in children with ASD.  Many of the behaviors that constitute the category of RRBs are similar to those described in studies of typical children. Factor analyses of RRBs scales in ASD have also found evidence for the two “subtypes” of RRBs described above (Cuccaro et.al., 2003;Bishop,Richler, & Lord, 2006). Given these similarities, it is important to ask what is defferent about these behaviors in typically developing children as opposed to children on the autism spectrum, and why their presence at a young age is not necessarily and indicator of later impairment.

            Part of the answer may lie in the developmental trajectories of these behaviors in typically developing children as compared to those of children with ASD. In the study by Evans et al. (1997) , children   between the ages of 24 and 36 months were found to exhibit the highest frequency and intensity of compulsive-like behaviors; after 26 months, score tended to decrease steadily and then more steeply, so that mean score between the that most typically developing should be made with caution, given that the data in this study were cross-sectional. In contrast, longitudinal studies of RRBs in children with ASD have found that many of these behaviors tend to increase in prevalence and severity with time, at least up until the age of 5 and possible until older ages (Moore & Goodson, 2003; Charman et al.,2005). One of the main differences in these behavior typically developing children versus those with ASD could be that, for most typical children, the behaviors tend to be common only within a relatively narrow window of development, in contrast to children, with ASD, who exhibit these behaviors for longer periods of time (Thelen, 1979).

This again highlights the importance of considering the differences in constellations of behaviors in young children with typical development as opposed to those with ASD. In the study by Evans et al. (1997), mean scores of typical children on the Childhood Roputines Invetory (CRI) were very low relative to the maximum achievable score. Most children exhibited one or two of these behaviors, or if they did exhibit a few, the behaviors were relatively mild. In contrast, studies of RRBs in children with ASD indicate that the majority of children, even those at young ages, tent to exhibit more than  two repetitive behaviors, and that these behaviors often interfere with the functioning of the child of the family (Richler, Bishop, Kleinke & Lord, 2007). These findings suggest that it is important to consider whether a child who shows on particular compulsive-like behavior (e.g.,lining up toy cars) also shows other similarly it is important to consider whether the interaction communication.

            Part of the reason that repetitive behaviors tend not to be as severe in typically developing children may be that these children do not have the added component of impairment in social interaction and communication to contend with. As a result, they are likely to spend more of their time interacting and communicating with others than engaging in repetitive activities. Even when they do engage in repetitive activities, they are likely to involve others in these activities, which makes the activities, which makes the activities more social and flexible. In contrast, children with ASD often prefer to participate in repetitive activities rather than interact with others, which further deprives them of social stimulation (Rogers & Ozonoff, 2005).

            Considering the early pictures of typical development and typical development in social interaction communication side by side, it is interesting to consider why the trajectories of typical development begin to diverge from those of the development of children with ASD. Some have argued that most children are born “hard-wired” to be oriented to turn, receiving more input (see Johnson et al., 2005). It has also been suggested that the early plasticity of the brain may provide an opportunity for experience to shape synaptic connections, eliminating those that are not needed and strengthening those that are crucial for higher-order functions, such as social cognition (Courchesne, Carper, & Akshoomoff, 2003). In contrast, some children may be born without the some predisposition to prioritize social input over nonsocial input (Dawson, Meltzoff, Osterling, Rinaldi & Brown, 1998; Dowson et al.,2004) or may experience change in trajectories of social and communication development, such as reaching a developmental plateau (Siperstein & Volkmar, 2004) or experiencing an actual worsening or regression in social and communication skills (Ozonoff, Williams & Landa, 2005).  Thus, for a number of reason, children with ASD may not receive the same social input from the environment as typically developing children during this critical period of brain development (Mudy & Nael, 2001). Consequwntly, the parts of the brain normally involved in social cognition may not be selectively shaped for this role (Johnson et al., 2005). As typically developing children become more socially sophisticated in the first few years of life, the impairments of children with ASD may build on each other and become more  apparent.

EARLY ASSESSMENT OF ASD

 Because of the cumulative effect of early appearing deficits, such as those describe above, detection of the symptoms of ASD tends to become easier as children get older. However, as we have come to understand more about early development of ASD, it has become increasingly possible to differentiate children with ASD from typically developing young children. Furthermore, whereas professionals have traditionally been hesitant to make diagnoses of ASD in children under the age of 3, recent  literature suggests that when made by experienced clinicians, diagnoses of toddlers are relatively stable over time. Because different methods (i.e., clinical observation and parent report) provide different types of information, diagnoses are most accurate and stable when based on information obtained from multiple source (e.g., Lord at al., 2006;Chawarska, Klin et al., 2007).

            Even for experienced clinicians, diagnosis can be difficult when trying to distinguish between ASD and other early childhood disorders. Psychological diagnoses, such as intellectual disability, expressive and receptive language disorders, anxiety disorder, and Attention-Deficit/Hyperactivity Disorder (ADHD), as well as genetic disorder, such as fragile-X syndrome, share many features with ASD. Making diagnostic distinctions in very young children is therefore a difficult process.

            Diagnosis of ASD is further complicated when a researcher or clinician is trying to determine whether a child meets the criteria for autism versus another spectrum condition (e.g., PDD-NOS,AS). Childhood Disintegrative Disorder (GDD) and Rett syndrome are quite rere and are less likely to be confused with autism, especially if a through  medical history is obtained. On the other hand, clinicians often find themselves trying to distinguish between autism and AS or PDD-NOS, in part because of poor agreement about the diagnostic criteria of these disorders (Ozonoff, South & Miller, 2000).DSM-IV provides guidelines for making diagnostic distinctions between these disorder, but these guidelines were written on the basis of studies of substantially older children (Volkamer et.al.,1994; Chawarska & Volkmar 2005).

            Both GDD abd Rett syndrome are characterized by period of apparently normal development followed by a substantial regression. The onset of GDD must occur after 2 years of age, and the child must exhibit loss of skills, adaptive behaviors, play, toileting, or  motor skills (American Psychiatric Association, 1994). After the regression, children with GDD must also exhibit impairment in at least two of the three domains in autism (i.e., social interaction communications, restricted and repetitive behaviors/interests). The regression in Rett syndrome, which is genetic disorder that occurs primarily in grils, occurs between between the ages of 5 and 48 months and is characterized by decelerated head growth; loss of previously acquired purposeful hand movements, such as holding utensils or picking up  the objects;  development of stereotyped midline hand movements, such as hand wringing; loss of social engagement; appearance of poorly coordinated gait or trunk movements; and severaly impaired language development (Amercan Psyhiatric ASssociation,1994).

            Although approximately 20% of children with autism experience a regression in language or social behaviors (Lord, Shulman & DiLavore 2004; Volkamar, Chawarska & Klin, 2005), the regressions in GDD and Rett syndrome are qualitatively distinct from the most common forms of regression in autism. First, whereas the regressions in GDD and Rett syndrome follow a period of typical development, some abnormality in children with autism is most often recognized, in hindsight, in the first years life (Oesterling & dawson, 1994), period to the onset of regression. Second, the regression in autism is characterize by a loss of language and/or social behaviors, without a loss of adaptive or motor skills (Volkamar & Rutter, 1995; Luyster et al., 2005), which are both typically with autism is almost always before the age of 24 months (Lord et al., 2004;Luster et al., 2005; Ozonoff et al., 2005; Chawarska, Paul, et al., 2007).

            In addition to early regression, there are other behaviors markers that have established as indicators of ASD in the first few years of  life. Retrospective analyses of videotapes of ASD in the first years of life have indicated that those who later receive a diagnosis to name, lack of socially directed looking, excessive mouthing of objects, and aversion to social touch, relative to comparison groups of typically developing children and children with non ASD development delays (Baranek, 1999; Oesteling Dawson & Munson 2002).

            More  recently, prospective studies of infants siblings of children with ASD, a population of children a high risk for developing ASD, have suggested a number of early features that are associated with a later diagnosis of ASD. Using the autism Observation Scale for Infants (AOSI) (see below), Zwaigenbaum et al. (2005) found that at 12 months of age, children who were later diagnosed with ASD showed evidence of language delay, as well as several behaviors abnormalities, such as difficulties with eye contact, visual tracking and attention, social smiling, imitation, social interest and affect. These infants also tended to demonstrate decreased positive affect and were more likely to exhibit extreme distress reactions and to fixate on objects. Using a parent-report measure of early communication skills, the MacArthur-Bates Communication Development Inventory-Infants Form  (Fenson et al., 1993), Mitchell et all. (2006) found that infant siblings who later met criteria for ASD reportedly understood fewer phrases and demonstrated significantly fewer gestures at 12 months of age than typically developing controls or siblings who at 12 months of age than typically developing controls or siblings who did not receive ASD diagnoses. These delays in understanding and use of single words.

            Studies comparing very young children with ASD to those with other types of development delays have made it increasingly possible for clinicians and researchers to differentiate between ASD and other non spectrum disorders. However, obtaining agrrement between diagnoses of autism, AS, and PDD-NOS has been much more difficult, particularly in young children, several different definitions for these disorder exits, which has complicated communication between professionals in the field (Ozonoff et al., 2000;Klin, Pauls, Schultz & Volkmar 2005). According to DSM-IV, AS is characterized by both qualitative impairments in social interaction and the presence of RRBs that are identicthose seen in  autism. However, unlike autism there can be no delay in language, cognitive development, or adaptive behaviors (except social skills) in a diagnosis of AS (American Psychiatric Association, 1994). A diagnosis of PDD-NOS is intended for children who exhibit significant impairments in reciprocal social interaction, as well as difficulties in either communication or the presences of RRBs (or subthreshold difficulties in both areas), who do not meet criteria for another ASD.

            Evidence suggest that diagnoses of PDD-NOS in early preschool are less stable than autism diagnoses (e.g., Stone et al., 1999; Lord et al., 2006). Chinicians are more reliable when making distinctions in 2-year older between ASD and nonspectrume diagnoses that between specific diagnoses on the spectrum, and it is not uncommon for children to heve a change in diagnosis within the spectrum (e.g., from a diagnosis od PDD-NOS to autism) (Stone et al., 1999; Lord et al., 2006). What may be important for very young children, therefore , is making a distinction between a spectrum and nonspectrume diagnosis, because differential diagnoses within the spectrum tend to be less stable (Lord et al., 2006). Thus, the intervention that children receive should be based more on their individual profiles of strengths and weakness, rather than on their specific diagnostic classifications.

SCREENING FOR ASD

General Issues

Our increasing knowledge of early development in children later diagnosed with ASD has facilitated the creation of a number of screeners targeted identifying young children with ASD. A major challenge as sociated with the development of these instruments is being able to discriminate children with significant developmental delays. In large with less pervasive and often temporary developmental delays. In large part, the ability to do this depends on our understanding of what verbal and nonverbal skills cluster together in the frst few years of typical development, and the degree to which these skills are impaired in children for whom ASD is a concern

Clarifying these early profiles of development and using them to screen for ASD in young children has both theoretical and practical implications. First, identifying children with ASD in the first few years of life allows for the collection of data about early profiles and trajectories of development. Such information efforts to improve the accuracy of the screening instrument themselves. In addition, earlier research has indicated that intervention is more effective if provided earlier (Harris & Handlemen, 2000).

Different approaches have been taken in designing screeners, some using caregivers as informants and others using professional. There are also two levels of screeners, one designed for population-based screening (i.e.,lavel 1 screeners), and the other designed for more targeted screening of children suspected of having a developmental disorder (i.e., level 2 screeners). In general, screening is distinct from diagnostic assessment in children with unrecognized or ambiguous symptoms of developmental disabilities, whereas the latter is most apporopriate for children for whom there is already some clear evidence of developmental abnormality.

Level 1 screeners typically employ caregiver reports as a means of gathering information. The primary advantage of this approach is that parents and caregivers are most familiar with the skills of the child across a variety of situations, and they may be more accurate than professionals in reporting low-frequency behaviors (such as using another person’s hand as a tool). However, caregivers may have less experience with children and a less refined understanding of the questions on the screener, which could potentially result in either over- or underestimating their child’s skills. There is also a risk of biased reporting (if caregivers already have beliefs about the diagnostic status of their child). Finally, creating a scale that caregivers will interpret as intended can be quite difficult.

            Level 2 screeners that use the reports of professional (such as health care worker or psychologist) have a different set of advantages and disadvantages. Professional may be more highly trained in observing and identifying certain diagnostically meaningful behaviors, and completion of the screeners can be standardized across reporters. However, they spend much less time with the child. As a result, they generally do not have the opportunity to evaluate the child across context and are musch less likely than caregivers to note low-frequency behaviors.

            In evaluating the effectiveness of level 1 and 2 screeners, it is useful to consider the constructs “rule in” all individuals with the targeted trait, and specificity refers to its ability to accurately “rule out” all individuals without the target trait. In the context of screening for ASD, sensitivity can maximize the detection of children who are showing a behaviors profile suggestive of ASD. Regardless of whether ASD is their final diagnosis or not, these children are likely to be “at risk” for one from of disability or another and will benefit from identifaction and result in more false positive, which can be expensive and potentially problematic, it is better to identify these children with developmental complicaers and diagnostic measure, specificity is a higher priority, because it is phenotypically similar conditions, such as language delay or intellectual disability.

Screening Instruments

Current measures for screening and diagnosis are considered here in turn, with reference to current research on their advantages and limitation for children under age 3. Population (level 1) and focused (level 2) screening are addressed (see Fingure 2.1)

            The Checklist for autism in Toddlers (CHAT;a level 1 screener) was initially introduced in the United Kingdom as a population screening measure for ASD (Baron-Cohen, Allen& Gilberg, 1992). The CHAT emphasized joint attention and imagination and was administered to children by health nurse, who routinely visit 18-month-olds in their homes in the Unites Kingdom. During the visit, the parents were also asked a series of question about their child’s development. Results indicated that most children classified by the CHAT as having autism were, in fact, later diagnosed with the disorder. However, it later became clear that two-thirds of the children who eventually received an ASD diagnosis were missed by the CHAT (Baird et al., 2000). Moreover, because children with suspected developmental disabilities were eliminated even before the screening, the CHAT’s effectiveness in distinguishing between ASD and other developmental disabilities was unclear.

            The Modified Checklist for Autism in Toddlers (M-CHAT), a modified version of the CHAT (Robins, Fein, Barton & Green, 2001), was created to address some of these concerns. The M-CHAT was administered.

Level 1 Screeners  CHAT M-CHAT ESAT CSBS Checklist PDDST-Stage 1 Physician Referrals Public Awareness

 

Level 2 Screeners CSBS Behaviors Sample/SORF PDDST-Stage 2 SCQ STAT CARS GARS

Diagnosis ADI-R ADOS Clinician’s best estimate

 

Figures 2.1. Level of screening and diagnosis for children with ASD at age 3 or younger.

To parents of 24-month-old children who were recruited from pediatric practices and special education program in the United State. In contrast to the CHAT, the M-CHAT is not administered to the child and instead relies on parent report. Like its predecessor, the M-CHAT successfully identifies children with autism at age 2. The M-CHAT was tested on two groups of children, a population sample and sample from special education program. More than 90% of the children identified as having autism were already in special education  program (Robins et al.., 2001), so the effictiveness of the M-CHAT for use in the general population is not yet clear. Initial report of sensitivity and specificity were very high (.87 and .99, respectively), but the authors coution that absolute psychometrics for this measure cannot be determined until follow-up evaluations are completed (Robins et al., 2001; Robins & Dumont Mathieu, 2006). A large study of the M-CHAT in a more representative sample is now under way, and the result of this study will be important in evaluating the effectiveneed of the M-CHAT.

          The Early Screening for Autistic Traits (ESAT;Swinkels et al.,2006) in level 1 screener questionnaire with a greater emphasis on play and less on joint attention that the previous instrument. Children who earned high score on the instrument were likely to have developmental problems. However, for children younger than 24 months of age, the EAST did successfully distinguish children with ASD from those with non-ASD conditions. In addition, like the CHAT and M-CHAT, it also failed to identify many children who were later diagnosed with ASD (see, e.g., Buitelaat et al., 2000). Despite  the measure’s problems with poor provided easy access to referrals. As a result of these realted benefits, early identification increased.

          The communication and Symbolic Behaviors Scales-Developmental profile (SCBS-DP;Wetherby & Prizant, 2002) is a brief caregiver questionnaire intended to identify children with communication disorder (not specifically ASD) between the ages of 6 and 24 months. If a child screens positively on the questionnaire besed on his or her caregiver’s responsed then a direct assessment (the behaviors Sample) and an additional caregiver questionnaire are administered. Although the initial questionnaire is a level 1 screener, a level 2 screener- the Scale of Red Flags (SORF;wetherby &Woods, 2002) for autism –was developed for use in scoring vidiotapes of the Behaviors Sample. With the SORF, researchers were able successfully identify most children with language delays as having or not having autism. However, because most of the children observed had screened positively on the CBS caregiver questionnaire, there was no way to identify missed cases and determine the measures’s sensitivity.

          The pervasive Developmental Disorder Screening Test (PDDST;Siegel, 1996) also offers a level 1 and a level screener and is intended for children over age 18 months and under the age of 6 years. It is a parent report questionnaire and is designed to screen specifically for ASD children  with ASD, such as nonverbal communication, temperament, play, language, and social engagement. Stage 1 of the PDDST (intended for use in primary care settings) was reported to have a sensitivity of .85 and a  specificity of .71 in a clinic-based sample. In a sample of children with ASD and children with other developmental disorders, sensitivity and specificity of .71 of the PDDST-stage 2   (intended for use in the developmental disorders clinics) varied according to the cutoff used, ranging from .69 to .88 and.25 to .63, respectively (siegel, 1996;  siegel & Hayer, 1999). Research on the PDDST is ongoing to provide further details about its pyschomeytric properties and usefulness in different populations.

          The Screening Test for Autisn in Two-Year-Olds (STAT; Stone, Coonrod, & Ousley, 2000) involves a direct assessment and , as a level 2 screener, is intended for children already suspected of having ASD. However, unlike the diagnostic tests described below, it is relatively brief. In addition, it is more straightforward to administer and score; consequently, it does not require extensive training on the part of the examiner. In a validation sample of 12 children with autism and 21 children with non-spectrum developmental disorders, the STAT correctly identified 10 (83%) of the children with autism  and 18 (86%) of the children with other developmental disorders (stone et al., 2000).

          The Social Communication Questionnaire (SCQ) Rutter, Bailey, Lord,& Berument, 2003) is a level 2 caregiver questionnaire designed to identify participants with ASD for research purposes. Although the meansure was normed on older children and adults, research has indicated that if the cutoff is modified so that fewer endorsed items are required, the SCQ works well for children as young as 3 years old (Corsello et al., 2007).  However, because the children had already been referred for services, it is unclear how appropriate the SCQ is for use in the general population.

          There are two other well-known scales primarily intended for level 2 screening but which may be mistaken for diagnostic instruments: the Childhood Autism Rating Scale (CARS) and the Gilliam Autism Rating Scale (GARS). The CARS (Schopler,Reichler,& Renner, 1988) is most useful with children beyond the 2-year-old level and up to 4 or 5 years in developmental skills, and it has been shown to have high sensitivity in older children and adults ( Sevin, Matson, Coe, Fee, & Sevin, 1991; Eaves & Milner, 1993). Studies have yielded mixed results with regard to the utility of the CARS for use with very young children. Lord (1995) reported that the CARS overidentified autism in 2-year-old with cognitive impairments, whereas Stone and colleagues (1999)reported good agreement with clinical diagnosis at age 2 (82% agreement). Agreement of ther CARS and clinical impression is better by age (Lord, 1995; Stone et al ., 1999), and specificity can be improved by raisig the CARS cutoff by 2 points (Lord, 1995).

          The GARS( Gilliam, 1995) is a behavioral checklist that was developed to screen for autism. However, the measure was not designed for or normed on children under 3 years of age, and thus its usefulness for a young population is unknown. One study (South et al., 2002) employed the GARS  underdiagnosed autism, failing to accurately classify more then half of the sample. Until revisions are made, the appropriateness of the GARS for children under 3 is limited.

                                               

                                                EARLY DIAGNOSIS OF ASD

                                                                                                                                         

Once a child has been identified as being at risk for ASD , he or she should be referred to a psychologist, psychiatrist , or developmental  pediatrician who specializes in early diagnosis of developmental disabilities. To aid clinicians in making accurate diagnoses, it is essential that a diagnostic assessment be multidimensional and multidisciplinary ( see Figure 2.2).

          This includes gathering information from different sources and assessing  a child’s behavior across a variety of contexts. Research has indicated that diagnoses  of 2-years-olds were significantly more stable when confirmed across two or three sources (I,e., standardized parent interview, direct child observation, and clinician’s best estimate diagnosis) as opposed to just one (Lord et al., 2006).

          Several instruments have been designed to aid professionals in gathering information needed to make a diagnosis of ASD. ( For a more comprehensive discussion of practice parameters and diagnostic instruments, see Filipek , Accardo, & Ashwal, 2000; Klinger & Renner, 2000; Lord & Corsello, 2005; and Bishop & Lord , 2006,) Standardized parent interviews and questionnaires can be useful in eliciting information from parents about their child’s behavior. In contrast to the traditional open-ended interview, semistructured interviews allow for a more comprehensive assessment of communication, social, and play behaviors associated with ASD and other developmental disorders. The most widely used and well established  semistructured interview that is designed to diagnose ASD is the Autism Diagnostic Interview-Revised (ADI-R; Lord, Rutter , & Le Couteur , 1994). The ADI-R provides quantifiable scores related to severity of symptoms in the areas of communication, reciprocal social interaction, verbal and nonverbal children. In order to meet criteria for a diagnosis of autism, the child must meet cutoffs in communication, Reciprocal Social Interaction, Restricted and Repetitive Behaviors and Interests, and Age of Onset. In children over the age of 3, these cutoffs have been found to clearly differentiate between children with autism and those with other disorders (Lord et al., 1994).

                                                            Diagnostic Assessment

Medical Examination •Rule out sensory impairment (chek hearing and vision) •Conduct genetic testing if indicated based on dysmorphology or family history •Conduct neurological exam

         

Parent Interview •Obtain thorough developmental history (attainment and/or loss motor, speech, self-help milestones). •Administer semistructured interview to gather information  about social and communication development, play, restricted and repetitive behaviors, and adaptive skills.

Child Observation •Create context in which to observe child’s social communication behaviors, play, and repetitive behaviors (with both parent and examiner). •Consult parents and teachers about whether behaviors observed during assessment were consistent with child’s behavior in other settings.

         

Developmental and Language Testing •Assess verbal (expressive and receptive) and nonverbal abilities. •Gather information about receptive and expressive language abilities. •Evaliate gross and fine motor skills.

          The validity of this instrument for children under the age of 3 has not been established. Therefore, a “toddler” version of the ADI-R is undergoing development and being used in some investigations. It includes 32 additional questions and codes specifically relevant to onset of difficulties in the early (C-Lord, personal communication, August,2006). Because this modified instrument is not yet available for general use, professionals may decide to use the published version of the ADI-R the age of 3. In particular, some studies have reported low sensitivity of the ADI-R  for populations of young children because many children do not meet cutoffs in the Restricted, Repetitive and Stereotyped Patterns of Behavior Domain (Ventola et al ., 2006; Chawarska, Klin , et al., 2007). Although the majority of 2-years-olds with ASD exhibit RRBs (Richler et al., 2007), some of the RRBs items that are currently include in the ADI-R algorithm (e.g.,compulsions on rituals) may be less prevalent in very young children with ASD.

          The diagnostic Interview for Social and Communication Disorder (DISCO;Wings, Leekeam, Libby, Gouls & Larcombe, 2002) is another semistructured interview designed to aid in the diagnosis of ASD Whereas the ADI-R is a diagnostic measure of ASD, the DISCO includes question about a wider range of difficulties and can be used to compile infor,ation necessary to diagnose other developmental and psychiatric disorders. The development, Diagnostic and Dimensonal and psychiatric disorders.