Language use in ASD is extremely heterogeneous, ranging from age-appropriate to nonverbal. The PI has tracked language development in three cohorts of children with ASD (total N = 42) for the past 8-12 years, from initial diagnosis at approximately 2.5 years to follow-up at 5-8 years, with typically developing (TD) controls (total N = 45) language-matched at study onset (the Longitudinal Study of Early Language; LSEL). We demonstrated the utility of fine-grained measures of production and comprehension in characterizing individual variation in both groups. These children are now in middle childhood to adolescence, allowing us to study language use appropriate to this age, its relationship to `individual' early child and parent language measures from the LSEL and `interactive' measures drawn from parent-child conversational dynamics. This project investigates three questions about the language of school age children with ASD or TD: How does variability in children's grammatical, semantic, and pragmatic language usage during school age demonstrate the inter-dependence vs. distinctiveness of these language areas? We hypothesize that composite measures of grammar and semantics collected in the current project will manifest distinctiveness, with little overlap in variability. Composite measures of semantics and pragmatics will manifest stronger inter-dependence. (2) Which individual early child and parent measures predict children's school age language use? We hypothesize that early child measures of grammar, semantics, and social cognition will predict school age measures in the same domains (e.g., early grammatical processing will predict grammatical usage at school age). We also hypothesize that parents who used more diverse words and grammatical constructions during earlier time periods will have children with more advanced categorization and grammatical usage at school age, and that parents who used more decontextualized language and provided more narrative scaffolding during earlier time periods will have children who use more sophisticated narratives at school age. We expect that early parent measures will exert relatively more influence on later language in the TD vs. ASD group. (3) Which early and concurrent measures of interactive conversational behavior will vary by group and predict children's school age language use? We examine the relationships between interactive dynamics during play sessions gathered when the children were preschoolers, and their school age language. We hypothesize that turn-taking and backchanneling will be more prevalent in TD dyadic interactions than in interactions with children with ASD while alignment measures will vary less by group. Moreover, conversational dynamics measures will predict aspects of children's subsequent speech and comprehension levels, as well as their narratives during school age. Finally, conversational dynamics measures will mediate the predictive power of individual measures, possibly in interaction with diagnosis.

Public Health Relevance

? This project investigates the individual and interactional determinants of language variation in school age and adolescent children with Autism Spectrum Disorder (ASD) or typical development (TD). We utilize UCONN's Longitudinal Study of Early Language in Autism dataset for richly detailed individual early language and social measures, derived from the children and their parents, and the parent-child interactions. We assess the same children, now 8-17 years of age, on detailed structural, conceptual, and pragmatic components of language, both individually and during adult-child interactions. We conduct sophisticated statistical analyses including Bayesian multi-level growth curve models to capture group, developmental, and nonlinear effects.

National Institute of Health (NIH)
National Institute on Deafness and Other Communication Disorders (NIDCD)
Research Project (R01)
Project #
Application #
Study Section
Language and Communication Study Section (LCOM)
Program Officer
Cooper, Judith
Project Start
Project End
Budget Start
Budget End
Support Year
Fiscal Year
Total Cost
Indirect Cost
University of Connecticut
Schools of Arts and Sciences
United States
Zip Code