Listening to and Learning from Disfluent Speech in Children with and without Developmental Language Disorder

PROJECT SUMMARY/ABSTRACT Developmental language disorder (DLD) impacts 7% of all children—both monolingual and bilingual— and has lifelong consequences for educational attainment and socioemotional wellbeing. DLD is characterized by deficits in structural language, but the mechanisms underlying these deficits remain under investigation. The Predictive Processing Hypothesis posits that prediction underlies human cognition and behavior, and it is increasingly being applied to language. Recent work has applied the Predictive Processing Hypothesis to DLD, investigating the possibility that deficits in predictive processing underlie language deficits. However, this hypothesis has only been tested in monolinguals, and only within the context of fluent, idealized language input. Fillers (e.g., uh and um), are a common disfluency in spontaneous language production and reliably precede difficult-to-retrieve words (e.g., words that are low-frequency). Neurotypical monolingual adults harness this association between fillers and production difficulty to anticipate upcoming novelty, but the mechanisms that underlie children’s comprehension of fillers as signals for upcoming difficult words are poorly understood. This is particularly so for children who have DLD and/or are exposed to more than one language (bilingual children). The overarching goal of this proposal is to examine both comprehension of fillers and consequences for learning in monolingual and bilingual children with and without DLD. A total of 80 children (40 English monolinguals and 40 Spanish-English bilinguals) ages 4-5, representing the full range of language ability (from DLD to typical language) will participate. The Specific Aims are: 1.) to determine the impact of disfluencies on language processing and learning in monolingual children with and without DLD and 2.) to determine the impact of disfluencies on language processing and learning in bilingual children with and without DLD. Two eye-tracking visual world tasks, in which children view novel and familiar objects side-by-side, will be used to assess predictive processing of fillers and word learning under fluent and disfluent conditions. Language ability (from typically developing to DLD) will be examined continuously under both aims, and bilingualism (proxied by years of second language exposure) will be examined continuously under Aim 2. Together, the findings across the two aims will test the innovative Predictive Processing Hypothesis within the novel context of disfluent speech and assess how consequences of disfluency on processing scale up to impact learning. Moreover, the findings will provide clinical insight into input characteristics that may enhance or disrupt learning and processing, with implications for structuring interventions. During the training period, the applicant will acquire new knowledge in the language profiles of bilingual and monolingual children with DLD as well as methodological considerations for working with this population; gain critical background knowledge in fluency and disorders; develop expertise in eye-tracking and advanced statistical analysis; and learn and practice principles of ethical science and effective mentorship.