Early Developmental Trajectories of Social Communication in Infants at Risk for ASD

Background: Siblings of children with autism spectrum disorders (ASD) are at increased risk for developing an ASD. By prospectively studying the younger siblings of children with an ASD (high risk for ASD, or HRA), as well as children without an ASD (low risk controls, or LRC), we may identify endophenotypes of ASD risk and predictors of ASD outcomes; such advances could enable earlier intervention and diagnosis – and improved outcomes – for affected individuals. In this study we seek to identify endophenotypes and diagnostic predictors of autism, by modeling the longitudinal trajectories of early social communication behaviors in HRA and LRC infants. 

Objectives: In this study, we investigate the trajectories along which social communication behaviors develop in a sample of HRA and LRC infants; we examine the extent to which ASD risk predicts the trajectories of these behaviors; and we determine whether and which early social communication trajectories predict later ASD diagnoses. 

Methods: Parents of 127 HRA and 119 LRC infants filled out the Communication and Symbolic Behavior Scales (CSBS) and the MacArthur-Bates Communicative Development Inventory (MCDI) when their children were 6, 9, 12 and 18 months old. Raw scores from CSBS social and symbolic subscales and MCDI gestures were modeled as a social communication factor using confirmatory factor analysis. Growth mixture modeling was then used to divide individuals into latent classes based on their social communication trajectories. Risk status was related to class membership, and the proportion of ASD diagnoses within each class was estimated for the subset of children who had confirmed diagnostic outcomes (N(typically developing) = 122; N(ASD) = 21; N(non-ASD disorder) = 3). 

Results: A 3-class solution was optimal. Class 1 comprised children with high social communication scores and accelerated social communication development between 6 and 12 months (28.0% of the sample); Class 2 comprised children with intermediate social communication scores and a roughly linear developmental trajectory from 6 to 18 months (58.3% of the sample). Class 3 comprised children with low social communication scores, and decelerated social communication development between 6 and 12 months (13.8% of the sample). HRA infants were significantly more likely than LRC to be in class 3 than either class 1 (odds-ratio = 7.19, p <0.01) or 2 (odds-ratio = 5.17, p = 0.02). Class membership was not predicted by the infants’ sex. ASD diagnoses were significantly predicted by class membership (χ²(2) = 7.46, p = 0.02) such that only 6% of the individuals with outcomes in Class 1, but 12% of those in Class 2 and 38% of those in class 3 developed ASD.

Conclusions: On parent-reported measures, infants at high risk for ASD were more likely to have lower social communication abilities and decelerated early social development compared to children without an ASD sibling. Infants belonging to the class exhibiting the lowest scores and slowest early development of social communication abilities were also significantly more likely to develop ASD. Given the relatively small sample examined herein, these results will be confirmed and extended using Bayesian estimation.