PMCCPMCCPMCC

Search tips
Search criteria 

Advanced

 
Logo of nihpaAbout Author manuscriptsSubmit a manuscriptHHS Public Access; Author Manuscript; Accepted for publication in peer reviewed journal;
 
Autism Res. Author manuscript; available in PMC 2010 September 27.
Published in final edited form as:
PMCID: PMC2946084
NIHMSID: NIHMS172986

Language Outcomes of Toddlers With Autism Spectrum Disorders: A Two Year Follow-Up

Abstract

Thirty-seven children 15–25 months of age received clinical diagnoses of autism spectrum disorder (ASD) and were re-evaluated two years later. All subjects were judged to have retained a diagnosis of ASD at the follow-up evaluation. Communication scores for the group as a whole during the first visit were significantly lower than nonverbal IQ. However, by the second visit, verbal and nonverbal scores were no longer significantly different. The group was divided into two subgroups, based on expressive language (EL) outcome at the second visit. The two groups were similar in the second year of life in terms of expressive communication skills and autistic symptoms, except for a trend toward more stereotypic and repetitive behavior in the worse outcome group. By the second visit, however, the groups differed significantly on all standard measures of expression and reception, as well as on autistic symptomotology and nonverbal IQ. When assessed during their second year, children who ended up in the better outcome group showed higher average nonverbal cognitive level, receptive language (RL) scores, number of sounds and words produced, use of symbolic play schemes, and response to joint attention bids. Regression analysis revealed that the variables for which significant differences between the two outcome groups in their second year of life were found provided significant prediction of EL outcome at age four. Stepwise regression identified RL and presence of stereotypic and repetitive at the first visit as significantly associated with EL outcome. Implications of these findings for early identification and intervention are discussed.

Keywords: autism, language, communication, toddlers

Introduction

Although autism spectrum disorders (ASDs) are characteristically life-long conditions, they show great variability in outcome. Many young children diagnosed with ASD go on to show significant deficits in academic achievement and independent living [Howlin, 2005], but others can achieve age-appropriate levels of function [Kelley, Paul, & Fein, 2006; Nordin & Gilberg, 1998; Sigman, 1999; Sutera et al., 2007; Toth, Munson, Meltzoff, & Dawson, 2006]. One of the most reliable predictors of positive outcome in ASD is the acquisition of spoken language by the end of the preschool period [DeMyer, Hingtgen, & Jackson, 1981; Gillberg, 1991; Howlin, 2005; Paul & Cohen, 1984].

Until recently children with ASDs were rarely diagnosed before the age of 3–4 years, [Chakrabarti & Fombonne, 2001; Charman & Baird, 2002; Fombonne, 2005; Filipek et al., 1999]. But a major thrust of current research has been to lower the age of identification, due in part to evidence supporting the effectiveness of early intervention [Rogers, 2006; Stahmer, Ingersoll, & Koegel, 2004]. Recent research suggests that clinical diagnosis of autism can be reliably assigned in the second year, and shows at least short-term stability [Matson, Wilkins, & Gonzales, 2008] when conferred by an team of experienced clinicians [Charman et al., 2005; Chawarska et al., 2007; Cox et al., 1999a, 1999b; Eaves & Ho, 2004; Lord, 1995; Lord, Risi, & DiLavore, 2006, Stone et al., 1999a, 1999b; Turner, Stone, Pozdol, & Coonrod, 2006; Wetherby et al., 2004].

Most studies of symptoms of ASD in children under the age of two are based on parental retrospective reports [Klin, Volkmar, & Sparrow, 1992; Volkmar, Stier, & Cohen, 1985], and retrospective analyses of videotaped materials [Baranek, 1999; Maestro et al., 2002; Osterling, Dawson, & Munson, 2002], with only a few studies relying on direct observations [e.g., Charman et al., 1997; Lord et al., 2000; Watt et al., 2006; Wetherby et al., 2007]. These studies converge on several findings. Symptoms are seen primarily in the expression of preverbal social communication, including difficulties in the use of gaze and conventional gestures for communicative purposes [e.g., Baron-Cohen, Cox, & Baird, 1996; Charman et al., 1997; Lord, 1995; Mundy, Sigman, Ungerer, & Sherman, 1987; Swettenham et al., 1998] and reduced expression of communicative intentions—especially joint attention [e.g., Dawson et al., 2004; Wetherby et al., 2004]. Other communication impairments include limited responsivity to speech, delayed emergence of spoken language, and the use of other’s body as a tool [Lord, 1995]. Differences are also seen in play and imitation [Charman, Swettenham, & Baron-Cohen, 2000; Charman et al., 1997; Cox et al., 1999a, 1999b; Dawson et al., 2004; Sigman and Ruskin, 1999; Sigman and Ungerer, 1984; Stone et al., 1999a, 1999b]. Stereotypic behaviors such as hand and finger mannerisms and unusual sensory behaviors may be present, but are often manifest only intermittently or with low intensity [Cox et al., 1999a, 1999b; Lord, 1995; Stone et al., 1999a, 1999b].

Because of the known connection between language development and prognosis in ASD [DeMyer et al., 1981; Howlin, 2005; Rutter, Greenfield, & Lockyer, 1967], several studies have looked for social-communicative behaviors in early-identified children that are related to language outcome in the preschool period. Several studies have examined relations between the acquisition of vocabulary and earlier-appearing behaviors. McDuffie, Yoder, and Stone [2005] reported that preverbal commenting and motor imitation in 2–3-year-old children with ASD were the strongest predictors of vocabulary development 6 months later. Yoder [2006] reported that frequency of communication and diversity of object play predicted lexical growth over a 12-month period in children with ASD. Smith, Mirenda, and Zaidman-Zait [2007] reported on rate of vocabulary growth, as measured by parent report, in children with ASD and small vocabularies followed from age 3 to 5. The most rapid vocabulary growth was seen in children with more words, high verbal imitation, pretend play, and gestural initiation of joint attention at baseline.

Other prospective research has looked more broadly at language and communication profiles and outcomes. Significant relationships have been found between early joint attention and later language development in preschool children with ASD [e.g., Charman et al., 2003; Dawson et al., 2004; Mundy, Sigman, & Kasari, 1990; Sigman et al., 1999], for example. Rogers, Hepburn, Stackhouse, and Wehner [2003] found relations between imitation and later language. Toth et al., [2006] reported on children with ASD seen first at an average age of 42 months and followed to five years of age. They found that joint attention, symbolic play, and imitation were all related to the rate of language development over this period. Thrum, Lord, Lee, and Newschaffer [2007] followed children from age 2 to 5 years. They reported nonverbal cognitive ability at age 2 was the strongest predictor of language at age 5, and that early joint attention, vocal and motor imitation skills were more impaired in children who did not develop language by age 5. Wetherby et al. [2007] found that the best predictors of age 3 language ability in children with ASD first seen at 18–24 months were the expression of communication for requests, understanding of words, and inventory of consonants. Thus, it appears that a range of communicative behaviors in children with ASD are related to rate of language acquisition.

This study used both direct observation and concurrent parental report of a well-defined sample of very young toddlers with autistic symptoms seen for evaluation between the ages of 15 and 25 months, at the earliest stage at which ASD can currently be diagnosed, in order to address the following aims:

  1. To replicate findings of previous studies [e.g., Wetherby et al., 2004, 2007] that describe early communicative behaviors of children with ASD under the age of 2, using direct, concurrent measures.
  2. To follow this well-defined cohort of young toddlers to the early preschool period, in order to provide prospective description of their communicative development in the preschool period.
  3. To identify variables in the second year of life that predict expressive language (EL) outcome in the preschool period, and may therefore be related to long-term outcome.

Methods

Participants

Thirty-seven children aged 15–25 months were selected for their young age, from all children evaluated for a differential diagnosis of ASD at a specialized clinic over a 2.5-year period. Given the limitations of current diagnostic instruments for very young children [Lord & Corsello, 2005], best estimate diagnosis was assigned based on history, clinical observation, and review of test results by two highly experienced clinicians at both Visit 1 (15–25 mo; mean 22 mo) and, independently, at Visit 2 (36–58 mo; mean: 47 mo). The reliability of expert clinical judgment has been shown to be high and is, at present, the “gold standard” for diagnosis in the age group [Klin, Lang, Cicchetti, & Volkmar, 2000]. At Visit 1, all these children were given a diagnosis of ASD (27 autism and 10 PDD-NOS) using this approach. Following the initial assessment, an intensive and comprehensive intervention program targeting specific areas of social communication deficits was recommended for each child.

At Visit 2, which took place at least one year after Visit 1, when children were between 3 and 5 years of age (mean 47 mo), all children diagnosed initially with ASD remained “on the spectrum” by an assessment independent of the first, in which at least one clinician was seeing the child for the first time. In eight cases, the diagnosis was changed from autism to PDD-NOS at Visit 2. All children diagnosed initially with PDD-NOS retained the diagnosis at follow-up. Informed consent was obtained from all parents and the study was conducted in accordance with the Human Investigation Committee of the Yale University School of Medicine.

Procedures

Visit 1

At the initial assessment, each child was seen for an intensive set of assessments that were conducted over the period of two successive days in a clinical setting. Direct assessment was conducted through testing and structured observations by licensed psychologists and speech-language pathologists, with the involvement of other professionals, including social workers, child psychiatrists, and pediatricians participating in the collection of parent interview data. Each child was assessed with the Mullen Scales of Early Learning [Mullen, 1995], the Communication and Symbolic Behavior Scales—Developmental Profile Behavioral Sample [CSBS-DP; Wetherby & Prizant, 2003], and the Autism Diagnostic Observation Scale—General; Module 1 [ADOS; Lord et al., 2000]. Each child’s parents were interviewed to collect data for the Vineland Adaptive Behavior Scales—Survey version [Sparrow, Balla, & Cicchetti, 1984] and the Autism Diagnsostic Interview—Revised [ADI-R; Lord, Rutter, Couture, 1994a; Lord, Rutter, Le Couteur, 1994b]. They also filled out the MacArthur Communicative Development Inventory—Words & Gestures [CDI; Fensen et al., 2002].

Visit 2

The follow-up visit took place at least one year following Visit 1. A similar assessment protocol was followed. The Mullen, Vineland, ADI-R, and ADOS Module 1 were re-administered. The CDI was not collected at Visit 2. At Visit 2, parents also completed a questionnaire on the community-based interventions received by their children. All children involved in the study, with two exceptions, received community-based interventions between Visits 1 and 2. Parental report of these interventions is summarized in Table I, which displays the proportion of children receiving various types of community-based interventions, and the average number of hours/week of intervention of each type received between Visits 1 and 2.

Table I
Intervention History: Treatments Received Between Visits 1 and 2

Results

Aim 1: Description of Early Communicative Behaviors

Figure 1 displays the average scaled scores, based on the CSBS-DP normative database, for this cohort on each of the CSBS-DP Behavior Sample Scales, which have means of 10 and standard deviation of 3 (normal range between 7 and 13). This display allows a comparison of performance across the scales within our cohort, in order to identify areas of relative strengths and weakness. It should be noted that the scores for the Sounds scale were not computed in the standard way, so scores from this scale are eliminated from the graphic display and data are reported for the Words subscale only. However, examination of consonant inventories produced during the CSBS-DP behavior samples, independent of referential intent, demonstrated inventories similar in size and content to those seen in typical age-mates [Paul, Chawarska, Klin, & Volkmar, 2007]. Data from the semi-structured CSBS-DP behavior sample, then, present a profile of young toddlers with ASD showing relative strengths in the ability to produce speech sounds and words. Lower levels of performance are seen in the abilities to respond to others’ speech, to use gestures and manage affect, gaze and reciprocity. The most profound deficits, however, are in the areas of expressing communicative functions, and engaging in age-appropriate play with objects. These findings are in accordance with those of earlier studies of ASD, both for this age group, as well as for children in the third year of life [Chawarska & Volkmar, 2005; Wetherby et al., 2004, 2007].

Figure 1
CSBS-DP* Scaled Scores (mean = 10, s.d. = 3): Comparison Across Scales for All Subjects at Visit 1 (20 mo).

Table II displays the mean scores and standard deviations obtained from these toddlers with ASD at their first clinic visit, when they were 15–25 months of age, for standard T-scores on the Mullen Visual Reception (VR), Fine Motor (FM), Receptive Language (RL), and Expressive Language (EL) scales; scale scores from Summary scales of the CSBS-DP; and age-equivalent (AE) scores on Vineland Adaptive Behavior Scales EL and RL, as well as on the CDI Expressive Vocabulary (EV), Receptive Vocabulary (RV), Play and Gesture (P&G) and Understanding Phrases (Ph) scores (n.b. Vineland subdomains and CDI scales do not provide standard scores; only AEs). Mean scores for the measures that were repeated at Visit 2 are also presented.

Table II
Means (and Standard Deviations) of Scores at Visits 1 and 2: All Subjects (n = 37)

One-way analysis of variance among three of the Mullen standard scores (VR, EL, and RL) suggests, in accordance with previous literature [Tager-Flusberg & Joseph, 2003; Thrum et al., 2007; Wetherby et al., 2004], that in comparison to relatively preserved skills in nonverbal areas (Mullen VR), these toddlers showed significant deficits in both EL and receptive language (RL) (overall F (2, 25) = 13.2; P < 0.005; Cohen’s d1 (VR vs. EL) = 1.0 [large]; Cohen’s d1 (VR vs. RL) = 0.92 [large]). Expressive scores obtained through parental report on the Vineland and CDI yield results comparable to those obtained via direct observation on the Mullen. Although both the former measures provide only AE and not standard scores, these AEs, at 10–13 months, are significantly below the children’s chronological age levels (Vineland: t = 13.5, P < 0.0001, Cohen’s d1 = 2.3 [very large]; CDI: t = 16.1, P < 0.0001, Cohen’s d1 = 2.4 [very large]). Average receptive AEs on the Vineland and CDI range from 11 to 14 months, and are also significantly below chronological age (Vineland: t = 7.2, P < 0.0001, Cohen’s d1 = 1.4 [large]; CDI: t = 18.7, P < 0.0001, Cohen’s d1 = 2.5 [very large]). Unlike toddlers with more specific language delays [e.g., Paul, Spangle-Looney, & Dahm, 1991; Rescorla, Roberts, & Dahlsgaard, 1997; Thal, 1991], who show relatively preserved comprehension abilities, these toddlers have deficits in both modes of language, whether measured in terms of parental checklists (CDI), standardized direct assessment (Mullen; CSBS), or parent report of adaptive use of communication via structured interview (Vineland).

Aim 2: Prospective Description of Communicative Development in the Preschool Period

Table II demonstrates marked growth in language skills across the two time periods studied, which were, on average two years apart. Average T-scores on the Mullen in both the receptive and expressive domains at Visit 2 were 41–44, in the low normal range. These scores on the Mullen show significant growth, i.e., at least one standard deviation (i.e., 10 points or more), in both receptive and expressive areas between Visits 1 and 2. Vineland AE scores corroborate this result, with significant increases in AE scores in both Expressive (t = 10.6, P < 0.0001, Cohen’s d1 = 2.2 [very large]) and Receptive (t = 8.5; P < 0.0001 Cohen’s d1 = 1.8 [large]) domains between Visits 1 and 2. Increases in AE scores were generally as large or larger than the number of months elapsed between Visits 1 and 2. Mullen VR T-scores remained stable (changing by 0.3 s.d.s) in the low average range, so language appears to be “catching up” to nonverbal abilities. ADOS-1 Communication and Socialization scores for the group as a whole also showed significant improvement between Visits 1 and 2 (Communication: t = 4.3; P < 0.0001 Cohen’s d1 = 0.81 [large]; Social: 7.9; P < 0.0001 Cohen’s d1 = 1.3 [very large]), suggesting a decrease in the severity of autistic symptoms in both areas, although subjects continued to receive scores within the autism spectrum at both visits.

In order to investigate the growth of expressive communication skills in another way, these data can also be used to explore the differences between groups with relatively better vs. worse preschool outcome in terms of EL. We can, for the purpose of the present discussion, define “good” preschool outcome as an AE score above 30 months on the Vineland Adaptive Behavior Expressive Communication Scale. This benchmark was for the following reasons:

  1. An AE of 30 months represents the score at least two standard deviations below the mean score for the average age of the cohort at Visit 2 (44 mo). Therefore, scores below 30 months would represent a significant delay.
  2. The Vineland Expressive scale measures habitual, everyday use of language and communication as reported by parents through detailed structured interviews, and is thought to reflect expressive communication skills more broadly, across a wider range of contexts, than checklists or observations in short, structured settings. For this reason, it has been used widely in the literature on ASD to document various aspects of adaptive development in this population (e.g., Carter et al., 1998a, 1998b; Liss et al., 2001a, 2001b; Loveland & Kelley, 1991a, 1991b; Rodrigue, Morgan, & Gefken, 1991; Schatz & Hamdan-Allen, 1995].

By this standard, 54% of the sample would be considered to have achieved a “good” language outcome by Visit 2, when they were, on average, approximately four years of age. Table III shows, in addition, that this proportion of the children scored broadly within the normal range on expressive communication by their second visit, as indexed by their standard scores on the Mullen EL. These scores show a mean of 51 with a standard deviation of 9.9; a distribution very close to that in the normative data for the test. It is important to note, however, that “typical” performance on the Mullen EL does not mean that communication skills are normal. Even these children with good outcomes retain significant deficits in the ability to use the language skills they have to achieve social interaction, as evidenced by a significant difference between their AE scores on the Vineland EL and their chronological age (t[1,19] + 2.7; P < 0.02; d1 = 0.78 [medium–large]) and the continued presence of autistic communication deficits seen in the ADOS Communication Algorithm scores (see Table III). When the cohort was divided this way, there was no significant difference in age between the two outcome groups at Visit 1 (t [35] = 1.5) or at Visit 2 (t [35] = 1.2). There was also no difference in the exposure to any of the types of intervention in Table II between the two groups. However, there was a significant difference between the two groups, favoring the “good” outcome group, in Mullen VR (nonverbal cognitive score) at Visit 1 (F (1, 35) = 7.0; P < 0.001; Cohen’s [1988] d1 = 0.60; effect size: medium). For this reason, Visit 1 Mullen VR was entered as a covariate in the subsequent analyses of differences between the two groups. Scores of the two groups at Visit 1 were compared, using a Multivariate Analysis of Co-Variance (SPSS 14.0), controlling for Mullen VR at Visit 1. Table IV shows that at Visit 1, a significant difference was found between the two groups in their second year of life. A greater ability in the “good” outcome group was seen in their ability to demonstrate understanding of words in a standard test format (Mullen RL), with a large effect size. Likewise, scores on the CSBS-DP behavior sample Understanding Words scale also distinguished the groups. In addition, CSBS-DP Behavior Sample scores on the size of consonant inventory produced, regardless of the presence of communicative intent. Scores on the use of words, understanding of words and use of symbolic play schemes were also ADOS-Receptive Joint Attention significantly different, with medium to large effect sizes. Finally, an exploratory Multivariate Analysis of Co-Variance was used to compare performance on individual subscales of the ADOS-1. The only scale on which the two groups were found to differ significantly in this analysis was on the measure of Responsive Joint Attention (RJA,) with a large effect size. All differences were in the direction of better performance in the group with “good” outcome. An additional Multivariate Analysis of Co-variance, again controlling for Mullen VR score at Visit 1, was performed on scores that were collected at Visit 2 (Mullen VR, EL, RL, Vineland EL and RL, ADOS-1 Communication, Socialization, Play/Imagination, and Stereotyped Behaviors and Restricted Interests [SBRI]). These analyses, also shown in Table IV, revealed significant differences between children with and without “good” outcome on all measures taken at Visit 2. Moreover, the effect sizes of all these differences were large to very large [Cohen, 1988].

Table III
Means (and Standard Deviations) of Scores at Visits 1 and 2 for Subjects with and without “Good” Language Outcome at Visit 2
Table IV
Results of Analysis of Co-Variance of Differences at Visits 1 and 2 Between Subjects With and Without Functional Language at Visit 2, Controlling for Mullen Visual Reception at Visit 1, With Effect Sizes

Aim 3: Predicting Preschool EL Outcome

An additional aim of the study was to use the data to determine how performance on the measures collected during the second year of life could predict preschool language outcome. To achieve this aim, a composite spoken language outcome measure was computed by converting scores on the three measures of expressive communication collected at Visit 2, Mullen EL T-score, Vineland EL AE score, and ADOS-1 Communication Algorithm score, to z scores (for the ADOS-1 Communication Algorithm z score, signs were reversed, since higher scores on this measure indicate worse performance, i.e., higher levels of autistic communication symptoms). The three z scores were then averaged to obtain a pooled Expressive Communication outcome measure. This pooled measure served as the outcome variable for linear regression.

To determine the most likely predictors, we examined the measures for which differences between the better and worse outcome groups, as indexed by Vineland EL scores at Visit 2, were found. (n.b. To use the pooled z score measure to divide the group into two outcome categories makes less clinical sense than to use Vineland EL outcome, since the derived scores compared each participant only with the rest of the group, and not to any standard of typical development.) The significant differences found between the two groups during their second year of life (see Table IV), all of which showed medium to large effect size, could be classified within five major constructs:

  • Nonverbal developmental level.
  • Production of sounds and words.
  • Receptive language.
  • Symbolic play.
  • Response to joint attention.

Additionally, the difference between groups on the ADOS-1 SRBI approached significance, and showed a moderate effect size in this analysis. These six constructs, then, were chosen as predictors in the regression analysis. Using six predictor variables with a sample size of 37 subjects falls within the standard guidelines for regression analysis [Bentler, 1985; Fletcher, Rice, & Ray, 1978; Francis, 1988]. The following variables were chosen to index each construct:

  • Nonverbal IQ: Mullen VR.
  • EL: Vineland EL AE.
  • Receptive language: Vineland RL AE.
  • Symbolic play: CDI play and gestures.
  • Response to joint attention: ADOS-1 RJA.
  • Stereotypic behaviors: ADOS-1 SRBI.

Regression was run to assess the impact of these variables on the pooled Expressive Communication z score at Visit 2. Results revealed the six predictors entered together yielded an R2 value of 0.47 (F [6, 24] = 3.6; P < 0.01). When the variables were entered in a stepwise fashion, only two were found to make significant contributions to prediction: Vineland RL (R2 = 0.41; F [2,28] = 9.7, P < 0.001), and ADOS-1 SRBI (R2 = 0.31; F [1,29] = 12.8, P < 0.001).

Discussion

In addressing Aim 1, this study corroborates and extends the profile of communicative skills observed in older toddlers with ASD [e.g., Lord, 1995; Lord, Risi, & DiLavore, 2006; Mundy et al., 1987; Sigman & Ungerer, 1984; Stone, Ousley, Yoder, Hogan, & Hepburn, 1997; Wetherby, Prizant, & Hutchinson, 1998; Wetherby et al., 2004], and reported in the largely retrospective studies of children in the second year of life [Baranek, 1999; Klin et al., 1992; Maestro et al., 2002; Osterling et al., 2002; Volkmar et al., 1985; Wetherby et al., 2007]. That is, children who meet diagnostic criteria for ASD late in their second year of life show below average levels of performance in the abilities to respond to others’ speech, to use gestures and manage affect, gaze and reciprocity; to express communicative functions; and to employ symbolic play schemes. Relative strengths are seen in their performance on nonverbal cognitive measures, as well as in the production of sounds and words in a semi-structured play context. This latter finding may appear somewhat surprising, since these children are significantly delayed in terms of word use in confrontation naming (Mullen EL), do not use as many different words as peers (on CDI parent report), or use communication effectively in everyday settings (on parent report in the Vineland EL; Communication scale on the CSBS-DP). What these findings suggest is that young toddlers with ASD, on average, do produce some sounds of speech and even word approximations (only 3–4 word productions are needed to score within the normal range on this scale of the CSBS-DP at this age level) in unstructured settings, even when more functional uses of language—to respond to others’ requests and initiate social interaction—are at low levels. This finding can be interpreted to suggest that at least some portion of very young toddlers with ASD acquire the basic speech production capacities necessary for development of spoken language, and may benefit from intensive intervention aimed at elaborating sound production and shaping early productions toward communicative speech.

In describing profiles of expressive communication to address Aim 2, we examined outcomes differentially for those children who made strong gains during the two-year follow-up period, and contrasted these with children who showed less progress. This strategy was adopted because of the known association between the acquisition of EL and better long-term outcomes in this population, so that data concerning the language development of children with better vs. worse outcomes during the preschool period could improve understanding of the course of communication acquisition in this population. This analysis demonstrated that approximately half the group achieved expressive communicative status on the Vineland that placed them within two standard deviations of the mean for a typical population. All participants, in both outcome groups, showed significant growth in language skills, on all measures, over the two-year period, with a closing of the gap apparent at Visit 1 between verbal and nonverbal skills, as well as significant decreases in autistic symptomotology. However, average language scores for the poorer outcome group were two years below age level at preschool age; whereas those for the better outcome group were broadly within the normal range by age four.

To examine in more detail how skills in the second year of life were related to these outcomes, we looked at differences between Visit 1 scores of children who ended up in the “good” and “poorer” outcome groups. Here we saw that there were initial differences in nonverbal cognitive performance, favoring the good outcome group. Controlling for nonverbal scores and comparing the other Visit 1 measures revealed that children with “good” language outcomes at age four scored higher on two direct measures of RL (but not by parent report), on production of sounds and words in the CSBS-DP behavior sample, use of play schemes, response to joint attention; and lower on appearance of stereotyped and repetitive behaviors.

Entering indices of the above constructs into a regression equation to address Aim 3 of the study showed that, collectively, these variables accounted for a significant amount (47%) of variance in EL outcome. Receptive language and level of stereotyped/repetitive behaviors were found to make significant contributions in stepwise regression analysis.

These findings are in accordance with those of earlier studies that identified nonverbal cognition [Thrum et al., 2007], play [Rogers et al., 2003; Yoder, 2006], RL, and consonant production [Wetherby et al., 2007] as predictors of later language outcome. These findings suggest the importance of including treatments targeted at the full range of these domains be incorporated into early intervention programs for children with ASD. However, the latter two findings merit special attention, especially for those concerned with improving communication development. Although all our young toddler participants with ASD showed low levels of expression of communicative intention, those who ended up in the poorer outcome group also tended to show lower performance on RL tasks and to produce fewer speech-like sounds or word approximations during their second year. These findings suggest that attention not only to increasing the production of nonverbal communicative acts, but also increasing responses to the speech of others and working toward increasing and shaping speech sound productions are likely to contribute to improved preschool language outcomes.

Another commonly found predictor, expression of joint attentional acts of communication [see for a review, Mundy, & Burnette, 2005], did not appear either to differentiate our two outcome groups in MANOVA, whether measured in the CSBS-DP Behavior Sample, or on the ADOS. Like Wetherby et al. [2007], we found that on both the CSBS-DP measure of joint attention (Scale 6) and on the ADOS initiation of joint attention item (B-10) average frequency of production for both outcome groups in the second year of life was very low. There were simply too few of these acts present in the second year of life to provide any degree of variability in relation to outcome. Thus, very low rates of initiation of joint attention, although useful for identifying children with autism at this age level, are not helpful in predicting communication outcome. We do, however, observe a difference between outcome groups on the ADOS item testing the ability to respond to bids for joint attention. In the second year of life, the ability to respond to joint attention may be a more reliable prognostic indicator than the ability to initiate joint attentional episodes, and may help identity those children with a higher risk for very delayed language development at this early age.

One finding not previously reported concerns the presence of the ADOS SRBI as a significant predictor. Earlier reports [Cox et al., 1999a, 1999b; Lord, 1995; Stone et al., 1999a, 1999b] have shown that stereotypic behaviors are not usually prominent at this age level. However, our findings suggest that when they do appear, they can have important consequences. Very young toddlers who display more of these behaviors tend to have poorer language outcomes. Thus, early-identified children with significant repetitive and stereotypic behaviors should be seen as at especially high risk for language disabilities.

In sum, this study adds to the growing body of evidence substantiating a profile of relative strengths and deficits in very young toddlers who meet diagnostic criteria for ASD. Data like these will enhance the capacity for earlier diagnoses and, consequently, the earlier provision of treatment to these children. In the critical area of language development, findings of connections between early profiles and later outcomes will help to focus educational treatments on areas likely to be pivotal in optimizing the opportunities of children with ASD for attainment of these crucial communicative skills.

Acknowledgments

Preparation of this article was supported by a MidCareer Development grant to Dr. Paul, K24 HD045576 funded by NIDCD. We also thank Carolyn Gosse, Kelly Cardona, Kate Elliot, and Elizabeth Schoen for their assistance in collecting data and preparing this manuscript, as well as the families who participated in our research.

Grant Sponsors: National Institute of Mental Health (NIMH)P01-03008, National Institute on Deafness and Other Communication Disorders (NIDCD)U54 MH66494, The National Institute of Environmental Health Sciences (NIEHS), The National Institute of Child Health and Human Development (NICHD), The National Institute of Neurological Disorders and Stroke (NINDS), NIDCDK24 HD045576, The National Alliance for Autism Research.

References

  • American Psychiatric Association. Diagnostic and Statistical Manual of Mental Diseases—IIIR. Washington, DC: American Psychiatric Association; 1987.
  • Baranek G. 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. 1999;29:213–224. [PubMed]
  • Baron-Cohen S, Cox A, Baird G. Psychological markers in the detection of autism in infancy in a large population. Journal of Psychiatry. 1996;168:158–163. [PubMed]
  • Bentler P. Theory and implementation of EQS: A structural equation program. Los Angeles: BMDP Statistical Software; 1985.
  • Carter A, Volkmar F, Sparrow S, Wang J, Lord C, et al. The Vineland Adaptive Behavior Scales: Supplementary norms for individuals with autism. Journal of Autism and Developmental Disorders. 1998a;28:287–302. [PubMed]
  • Carter A, Volkmar F, Sparrow S, Wang J, Lord C, et al. The Vineland Adaptive Behavior Scales: Supplementary norms for individuals with autism. Journal of Autism and Developmental Disorders. 1998b;28:287–302. [PubMed]
  • Chakrabarti S, Fombonne E. Pervasive developmental disorders in preschool children. Journal of the American Medical Association. 2001;285:3093–3099. [PubMed]
  • Charman T, Baird G. Practitioner review: Diagnosis of autism spectrum disorders in 2- and 3-year-old children. Journal of Child Psychology and Psychiatry and Allied Disciplines. 2002;43:289–305. [PubMed]
  • Charman T, Baron-Cohen S, Swettenham J, Baird G, Drew A, et al. Predicting language outcome in infants with autism and pervasive developmental disorders. International Journal of Language and Communication Disorders. 2003;38:265–285. [PubMed]
  • Charman T, Swettenham J, Baron-Cohen S. An experimental investigation of social-cognitive abilities of infants with autism: Clinical implications. In: Muir D, Slater A, editors. Infant development: The essential readings. Malden, MA, USA: Blackwell Publishing; 2000. pp. 343–363.
  • Charman T, Swettenham J, Baron-Cohen S, Cox A, Baird G, et al. Infants with autism: An investigation of empathy, pretend play, joint attention, and imitation. Developmental Psychology. 1997;33:781–789. [PubMed]
  • Charman T, Taylor E, Drew A, Brown CJ, et al. 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 & Psychiatry. 2005;46:500–513. [PubMed]
  • Chawarska K, Klin A, Paul R, Volkmar F. Autism spectrum disorder in the second year: stability and change in syndrome expression. Journal of Child Psychology and Psychiatry. 2007;48:128–138. [PubMed]
  • Chawarska K, Volkmar F. Autism in infancy and early childhood. In: Volkmar F, Paul R, Klin A, Cohen D, editors. Handbook of autism and pervasive developmental disorders. 3rd ed. New York: Wiley; 2005. pp. 223–246.
  • Cohen J. Statistical power analysis for the behavioral sciences. 2nd ed. Hillsdale, NJ: Lawrence Erlbaum Associates; 1988.
  • Cox A, Klein K, Charman T, Baird G, Baron-Cohen S, et al. Autism spectrum disorders at 20 and 42 months of age: Stability of clinical and ADI-R diagnosis. Journal of Child Psychology and Psychiatry and Allied Disciplines. 1999a;40:719–732. [PubMed]
  • Cox A, Klein K, Charman T, Baird G, Baron-Cohen S, et al. Autism spectrum disorders at 20 and 42 months of age: stability of clinical and ADI-R diagnosis. Journal of Child Psychology and Psychiatry. 1999b;40:719–732. [PubMed]
  • Dawson G, Toth K, Abbott R, Osterling J, Munson J, et al. Early social attention impairments in autism: Social orienting, joint attention, and attention to distress. Developmental Psychology. 2004;40:271–283. [PubMed]
  • DeMyer M, Hingtgen J, Jackson R. Infantile autism reviewed: A decade of research. Schizophrenia Bulletin. 1981;7:388–451. [PubMed]
  • Eaves L, Ho H. The very early identification of autism: Outcome to 4 1/2–5. Journal of Autism and Developmental Disorders. 2004;34:367–378. [PubMed]
  • Fensen L, Dale P, Reznick S, Thal D, Bates E, et al. MacArthur communicative developmental inventories. Baltimore: Paul H. Brookes; 1993.
  • Fensen L, Dale P, Reznick S, Thal D, Bates E, et al. MacArthur communicative developmental inventories. 2nd ed. Baltimore: Paul H. Brookes; 2002.
  • Filipek P, Accardo P, Baranek G, Cook E, Dawson G, et al. The screening and diagnosis of autistic spectrum disorders. Journal of Autism and Developmental Disorders. 1999;29:439–484. [PubMed]
  • Fletcher J, Rice W, Ray R. Linear discriminant function analysis in neurophychological research: Some uses and abuses. Cortex. 1978;14:564–577. [PubMed]
  • Fombonne E. Epidemiological studies of pervasive developmental disorders. In: Volkmar F, Paul R, Klin A, editors. Handbook of autism and pervasive developmental disorders. 3rd ed. Vol. 1. New York: Wiley; 2005. pp. 42–69.
  • Francis D. An introduction to structural equation models. Journal of Clinical and Experimental Neuropsychology. 1988;10:623–639. [PubMed]
  • Gillberg C. Outcome in autism and autistic-like conditions. Journal of the American Academy of Child and Adolescent Psychiatry. 1991;30:375–382. [PubMed]
  • Howlin P. Outcomes in autism spectrum disorders. In: Volkmar F, Paul R, Klin A, Cohen D, editors. Handbook of autism and pervasive developmental disorders. 3rd ed. Vol. 1. New York: Wiley; 2005. pp. 201–222.
  • Kelley E, Paul J, Fein D. Residual language deficits in optimal outcome children with a history of autism. Journal of Autism and Developmental Disorders. 2006;36:807–828. [PubMed]
  • Klin A, Lang J, Cicchetti DV, Volkmar FR. Brief report: Interrater reliability of clinical diagnosis and DSM-IV criteria for autistic disorder: results of the DSM-IV autism field trial. Journal of Autism and Developmental Disorders. 2000;30:163–167. [PubMed]
  • Klin A, Volkmar F, Sparrow S. Autistic social dysfunction: Some limitations of the theory of mind hypothesis. Journal of Child Psychology and Psychiatry. 1992;33:861–876. [PubMed]
  • Liss M, Harel B, Fein D, Allen D, Dunn M, et al. Predictors and correlates of adaptive functioning in children with developmental disorders. Journal of Autism and Developmental Disorders. 2001a;31:219–230. [PubMed]
  • Liss M, Harel B, Fein D, Allen D, Dunn M, et al. Predictors and correlates of adaptive functioning in children with developmental disorders. Journal of Autism and Developmental Disorders. 2001b;31:219–230. [PubMed]
  • Lord C. Follow-up of two-year-olds referred for possible autism. Journal of Child Psychology and Psychiatry and Allied Disciplines. 1995;36:1365–1382. [PubMed]
  • Lord C, Corsello C. Diagnostic instruments in autism spectrum. In: Volkmar F, Paul R, Klin A, Cohen D, editors. Handbook of autism and pervasive developmental disorders. 3rd ed. Vol. 2. New York: Wiley; 2005. pp. 730–771.
  • Lord C, Risi S, DiLavore P. Autism from 2 to 9 years of age. Archives of General Psychiatry. 2006;63:694–701. [PubMed]
  • Lord C, Risi S, Lambrecht L, Cook EH, Leventhal BL, et al. The Autism Diagnostic Observation Schedule—Generic: A standard measure of social and communication deficits associated with the spectrum of autism. Journal of Autism and Developmental Disorders. 2000;30:205–223. [PubMed]
  • Lord C, Rutter M, Conture A. Autism Diagnostic Interview—Revised: A revised version of a diagnostic interview for caregivers of individuals with possible pervasive developmental disorders. Journal of Autism and Developmental Disorders. 1994a;24:659–685. [PubMed]
  • Lord C, Rutter M, Le Couteur A. Autism Diagnostic Interview—Revised: A revised version of a diagnostic interview for caregivers of individuals with possible pervasive developmental disorders. Journal of Autism and Developmental Disorders. 1994b;24:659–685. [PubMed]
  • Loveland K, Kelley M. Development of adaptive behavior in preschoolers with autism or Down syndrome. American Journal of Mental Retardation. 1991a;96:13–20. [PubMed]
  • Loveland K, Kelley M. Development of adaptive behavior in preschoolers with autism or Down syndrome. American journal of Mental Retardation. 1991b;96:13–20. [PubMed]
  • Maestro S, Muratori SM, Pei F, Stern D, Golse B, et al. Attentional skills during the first 6 months of age in autism spectrum disorder. Journal of the American Academy of Child and Adolescent Psychiatry. 2002;41:1239–1245. [PubMed]
  • Matson J, Wilkins J, Gonzalez M. Early identification and diagnosis in autism spectrum disorders in young children and infants: How early is too early? Research in Autism Spectrum Disorders. 2008;2:75–84.
  • McDuffie A, Yoder P, Stone W. Prelinguistic predictors of vocabulary in young children with autism spectrum disorders. Journal of Speech, Language, and Hearing Research. 2005;48:1080–1097. [PubMed]
  • Mullen E. Mullen scales of early learning: AGS Edition. Circle Pines, MN: AGS Publishing; 1995.
  • Mundy P, Burnette C. Joint attention and neurodevelopmental models of autism. In: Volkmar F, Paul R, Klin A, Cohen D, editors. Handbook of autism and pervasive developmental disorders. 3rd ed. Vol. 2. New York: Wiley; 2005. pp. 650–681.
  • Mundy P, Sigman M, Kasari C. A longitudinal study of joint attention and language development in autistic children. Journal of Autism and Developmental Disorders. 1990;20:115–129. [PubMed]
  • Mundy P, Sigman M, Ungerer J, Sherman T. Nonverbal communication and play correlates of language development in autistic children. Journal of Autism and Developmental Disorders. 1987;17:349–364. [PubMed]
  • Nordin V, Gillberg C. The long-term course of autistic disorders: Update on follow-up studies. Acta Psychiatrica Scandinavica. 1998;97:99–108. [PubMed]
  • Osterling J, Dawson G, Munson J. Early recognition of 1-year-old infants with autism spectrum disorder versus mental retardation. Developmental Psychopathology. 2002;14:239–251. [PubMed]
  • Paul R, Chawarska K, Klin A, Volkmar F. Dissociations in development of early communication in ASD. In: Paul R, editor. Language disorders from a developmental perspective: Essays in honor of Robin Chapman. Hillsdale, NJ: Erlbaum; 2007. pp. 163–194.
  • Paul R, Cohen DJ. Outcomes of severe disorders of language acquisition. Journal of Autism and Developmental Disorders. 1984;14:405–422. [PubMed]
  • Paul R, Spangle-Looney S, Dahm P. Communication and socialization skills at ages 2 and 3 in “late-talking” young children. Journal of Speech and Hearing Research. 1991;34:858–865. [PubMed]
  • Rescorla L, Roberts J, Dahlsgaard K. Late talkers at 2: Outcome at age 3. Journal of Speech and Hearing Research. 1997;40:556–566. [PubMed]
  • Rodrigue J, Morgan S, Gefken G. A comparative evaluation of adaptive behavior in children and adolescents with autism Down syndrome and normal development. Journal of Autism and Developmental Disorders. 1991;21:187–198. [PubMed]
  • Rogers S. Evidence-based intervention for language development in young children with autism. In: Charman T, Stone W, editors. Social and communication development in autism spectrum disorders: Early identification, diagnosis, and intervention. New York: Guilford; 2006.
  • Rogers S, Hepburn S, Stackhouse T, Wehner E. Imitation performance in toddlers with autism and those with other developmental disorders. Journal of Child Psychology and Psychiatry. 2003;44:763–781. [PubMed]
  • Rutter M, Greenfield D, Lockyer L. A five- to fifteen-year follow-up study of infantile psychosis: II. Social and behavioral outcome. British Journal of Psychiatry. 1967;113:1183–1199. [PubMed]
  • Schatz J, Hamdan-Allen G. Effects of age and IQ on adaptive behavior domains for children with autism. Journal of Autism and Developmental Disorders. 1995;25:51–60. [PubMed]
  • Sigman M. Change and continuity in the development of children with autism. Journal of Child Psychology and Psychiatry, and Allied disciplines. 1999;39:817–827. [PubMed]
  • Sigman M, Ruskin E. Continuity and change in the social competence of children with autism Down syndrome and developmental delays. Monographs of the Society for Research in Child Development. 1999;64:v113–v114. [PubMed]
  • Sigman M, Ruskin E, Arbeile S, Corona R, Dissanayake C, et al. Continuity and change in the social competence of children with autism. Journal of Autism and Developmental Disorders. 1999;33:15–23.
  • Sigman M, Ungerer J. Cognitive and language skills in autistic mentally retarded and normal children. Developmental Psychology. 1984;20:293–302.
  • Smith V, Mirenda P, Zaidman-Zait A. Predictors of expressive vocabulary growth in children with autism. Journal of Speech, Language, and Hearing Research. 2007;50:149–160. [PubMed]
  • Sparrow S, Balla D, Cicchetti DV. The Vineland Adapative Behavior Scales (Survey Form) Circle Pines, MN: American Guidance Service; 1984.
  • Stahmer A, Intersoll B, Koegel R. Inclusive programming for toddlers with autism spectrum disorders: Outcomes from a children’s toddler school. Journal of Positive Behavior Interventions. 2005;6:67–82.
  • Stone W, Lee E, Ashford L, Brissie J, Hepburn S, et al. Can autism be diagnosed accurately in children under 3 years? Journal of Child Psychology and Psychiatry. 1999a;40:219–226. [PubMed]
  • Stone W, Lee E, Ashford L, Brissie J, Hepburn S, et al. Can autism he diagnosed accurately in children under 3 years? Journal of Child Psychology and Psychiatry. 1999b;40:219–226. [PubMed]
  • Stone WL, Ousley OY, Yoder PJ, Hogan KL, Hepburn SL. Nonverbal communication in two- and three-year-old children with autism. Journal of Autism and Developmental Disorders. 1997;27:677–696. [PubMed]
  • Sutera S, Pandey J, Esser E, Rosenthal M, Wilson L, et al. Predictors of optimal outcome in toddlers diagnosed with autism spectrum disorders. Journal of Autism and Developmental Disorders. 2007;37:98–107. [PubMed]
  • Swettenham J, Baron-Cohen S, Charman T, Cox A, Baird G, et al. The frequency 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. 1998;39:747–753. [PubMed]
  • Tager-Flusberg H, Joseph R. Identifying neurocognitive phenotypes in autism. Philosophical Transactions of the Royal Society Series, B. 2003;358:303–314. [PMC free article] [PubMed]
  • Thal D. Language and cognition in normal and late-talking toddlers. Topics in Language Disorders. 1991;11:33–42.
  • Thrum A, Lord C, Lee L, Newschaffer C. Predictors of language acquisition in preschool children with autism spectrum disorders. Journal of Autism and Developmental Disorders. 2007;37:1721–1734. [PubMed]
  • Toth K, Munson J, Meltzoff A, Dawson G. Early predictors of communication development in young children with autism spectrum disorder: Joint attention, imitation, and toy play. Journal of Autism and Developmental Disorders. 2006;36:993–1005. [PMC free article] [PubMed]
  • Turner L, Stone W, Pozdol S, Coonrod E. Follow-up of children with autism spectrum disorders from age 2 to age 9. Autism. 2006;10:243–265. [PubMed]
  • Volkmar FR, Stier DM, Cohen DJ. Age of recognition of pervasive developmental disorder. American Journal of Psychiatry. 1985;142:1450–1452. [PubMed]
  • Watt N, Wetherby A, Shumway A. Prelinguistic predictors of language outcome at three years of age. Journal of Speech, Language, and Hearing Research. 2006;49:1224–1237. [PubMed]
  • Wetherby A, Prizant B. Communication and Symbolic Behavior Scales-Developmental Profile. Baltimore: Paul H. Brookes; 2003.
  • Wetherby A, Prizant B, Hutchinson T. Communication, social/affective, and symbolic profiles of young children with autism and pervasive developmental disorders. American Journal of Speech-Language Pathology. 1998;7:79–91.
  • Wetherby A, Watt N, Morgan L, Shumway S. Social communication profiles of children with autism spectrum disorders late in the second year of life. Journal of Autism and Developmental Disorders. 2007;37:960–975. [PubMed]
  • Wetherby A, Woods J, Allen L, Cleary J, Dickinson H, et al. Early indicators of autism spectrum disorders in the second year of life. Journal of Autism and Developmental Disorders. 2004;34:473–493. [PubMed]
  • Yoder P. Predicting lexical density growth rate in young children with autism spectrum disorders. American Journal of Speech-Language Pathology. 2006;5:378–388. [PubMed]