People with autism spectrum disorder (ASD) often face challenges understanding spoken language, especially in noisy or everyday environments. Until now, most research has relied on simplified models that examined isolated syllables or words, leaving key questions unanswered.

In a study published online on July 30 in NeuroImage, Théo Vanneau, Ph.D., Sophie Molholm, Ph.D., and colleagues used advanced neuromodeling techniques to better understand how people with ASD process multisensory speech in a more natural, continuous context. Using high-density EEG and innovative linear modeling, the team studied adolescents and young adults with ASD as they listened to speech in audio-only, visual-only, and combined audiovisual conditions, with different levels of background noise.

The researchers found that while individuals with ASD can process the sound of speech itself, they struggle to combine it with visual cues—such as lip and facial movements—that typically make spoken communication easier to follow. These insights could pave the way for new therapies designed to strengthen sensitivity to visual speech cues, ultimately helping people with ASD better navigate the challenges of everyday conversation.

Dr. Molholm is professor of pediatrics, in the Dominick P. Purpura Department of Neuroscience, and of psychiatry and behavioral sciences, Muriel and Harold Block Faculty Scholar in Mental Illness, and director of the Rose F. Kennedy Intellectual and Developmental Research Center at Einstein. Dr. Vanneau, co-first author of the study, is a research fellow in Dr. Molholm’s lab.