SCSB Colloquium with Dr. Jesse Snedeker: The use of discourse context for predictive language comprehension in autism: evidence from a naturalistic EEG paradigm

Date: Wednesday, September 23, 2026 Location: 46-3002 (Singleton Auditorium) Time: 4:00pm-5:00pm ET, followed by reception Speaker: Jesse Snedeker, Ph.D. Affiliation: Professor, Department of Psychology, Harvard University Host: Dr. Pawan Sinha Talk title: The use of discourse context for predictive language comprehension in autism: evidence from a naturalistic EEG paradigm Abstract: An increasing proportion of the people diagnosed with autism have strong language skills, as measured by both standardized tests and laboratory tasks. Nevertheless, difficulties in social communication and interaction remain a core diagnostic feature of autism. There are two broad ways to reconcile these observations: 1) our assessments could be failing the capture higher-level language skills that are impaired (or the temporal dynamics of these processes); 2) the social communication difficulties could arise for reasons largely independent from linguistic processing. To explore the first possibility my lab has focused on predictive lexical processing which appears to be the central bottleneck in language comprehension. We first look at explicit lexical prediction using a sentence cloze task and discover that explicit lexical prediction in both adults and children depends more on structural language abilities than on diagnostic status. This data does not rule out deficits in lexical prediction due to the limited context provided in such tasks and their insensitivity to differences in the timing of prediction. To explore moment-to-moment lexical prediction in a richer discourse, we have developed a naturalistic EEG task that we call the Storytime Paradigm. I’ll present recent data which demonstrates that lexical processing this task is sensitive to both development change and individual differences, with the effects of prediction growing with age and linguistic knowledge. Then I will reveal our findings from a group of children with autism diagnoses (6-12 years) and a matched control group. I will end by discussing the need for parallel research on other well-defined subgroups of autistic folks.