Predicting Late Talkers in Infants who are at Elevated Familial Likelihood for Autism

University Of Minnesota, Minneapolis MN

Investigators

Abstract

ABSTRACT Predicting Late Talkers in Infants who are at Elevated Familial Likelihood for Autism Autistic children are late to say their first word, and these early challenges persist with over 70% of autistic preschoolers having language impairment. Siblings of autistic children unaffected by autism themselves are at a 4-5-fold increased risk of developing language challenges. The field lacks screening tools with strong predictive power to identify late talking autistic children during the birth-to-three early intervention period. The proposed project makes significant steps towards early identification of language impairment and understanding the developmental sequelae of language in autistic toddlers by leveraging data from the Baby Sibling Research Consortium database, the largest collection of language data of infants who develop autism. Aim 1 of the proposed study is create a normative model sensitive to the heterogeneity of autistic language development using summary-level data from the MacArthur-Bates Communicative Development Inventories (CDI). This model will then be used to predict late talkers. The normative modeling framework quantifies individual differences in scores, flagging individuals for further follow up. This framework moves away from the limitations of simple “case- control” designs by allowing for the heterogeneity that is inherent to developmental disorders. We will make our normative model publicly available, to be used by investigators interested in samples that tend to be smaller in size, increasing the significance and