Examining Unique Predictors of Language Growth in Infants at High and Low Risk for Autism Spectrum Disorder

Background: Language ability is a powerful predictor of outcome for individuals with ASD (Anderson et al., 2013; Venter et al., 1992), highlighting the importance of understanding how early language emerges. Previous studies exploring predictors of language development in ASD often examined one predictor in isolation, preventing the identification of predictors that contribute most to language outcome. To address this, recent work examined several predictors within a single model, identifying three that uniquely account for language growth in preschool-aged children with ASD: intentional communication (IC), response to joint attention (RJA), and parent verbal responsiveness (PVR; Yoder et al., 2015). The current study evaluates these value-added predictors of language growth in infants at high risk (HR) and low risk (LR) for ASD to capture the developmental period in which language is rapidly developing.

Objectives: (1) To examine expressive language (EL) and receptive language (RL) growth trajectories for HR and LR infants. (2) To examine the unique contribution of predictors (IC, RJA, and PVR) on EL and RL growth trajectories.

Methods: Data were collected at 12, 15 and 18 months as part of a prospective, longitudinal, multi-site study of HR (n=45) and LR (n=32) infant siblings. IC and RJA were measured using the Early Social Communication Scales (Mundy et al., 2003). PVR was coded from videos of parent-child free play (data coding ongoing). EL and RL were measured using the MacArthur-Bates Communicative Development Inventory (MCDI; Fenson et al., 2007). Hierarchical linear modeling was used to examine the contributions of risk group and hypothesized predictors at 12-months on EL and RL growth between 12-18 months.

Results: Preliminary analyses tested growth in EL and RL over time for HR and LR infants with time modeled as infant age. For EL, growth was best modeled by fixed-intercept, random quadratic time, b=1.09, p<.001. There was a significant quadratic time by group interaction, b=-.83, p=.027, suggesting that LR infants’ growth in EL accelerated more quickly than that of HR infants. For RL, the best fitting growth model included a random intercept and random linear slope, b=14.49, p<.001. Although RL increased on average across age, LR infants showed consistently higher RL compared to HR infants across time as indicated by a main effect of risk group, b=-29.67, p=.005, and a lack of linear time by group interaction. IC and RJA were not predictive of level or change in EL or RL. Upon completion of PVR video coding, analyses will examine PVR in relation to these trajectories.

Conclusions: Risk group differentially predicted trajectories for both EL and RL. Preliminary analyses indicate that some hypothesized 12-month behavioral predictors (i.e., IC and RJA) did not account for these differences in growth, potentially because joint attention, an important aspect of both constructs, may still be emerging at 12 months (Ibanez et al., 2013; Mundy et al., 2007). Further analyses will examine the contribution of PVR, and whether changes in IC and RJA as they emerge predict language growth.