EEG recordings from about 400 children suggest that face-processing differences in autism emerge across multiple stages of visual development, not simply as a delay in one familiar brain signal.
A large new study has found that autistic children show different patterns of brain activity when viewing human faces, adding to evidence that face perception develops along a different path in autism. Published in Nature Mental Health, the research used electroencephalography, or EEG, to examine how the brain responds across the full scalp rather than focusing on a small number of recording sites.
The work drew on data from about 400 autistic and neurotypical children in the Autism Biomarkers Consortium for Clinical Trials. Researchers showed participants images of faces and ordinary objects while recording electrical activity from 128 electrodes placed around the head. Using machine learning, the team found that face-related signals were more clearly distinguishable in neurotypical children than in autistic children. Jason Griffin, first author of the study and an assistant professor of psychology at the University of Houston, said the neural activity measured at the scalp was “just not as distinct” in autistic children.
The study also suggests that these differences are not simply a matter of timing around a single well-known brain response. Earlier work has often concentrated on the N170, a signal that typically appears about 170 milliseconds after a face is seen and has been reported as delayed in autism. By looking across a broader time window, the new research indicates that differences may emerge even earlier in visual processing. That matters because it suggests the brain’s response to faces in autism may diverge at multiple stages, not just at one late checkpoint.
A further finding was developmental. Among neurotypical children, the brain’s ability to distinguish faces became more specialised with age. That pattern was not seen in autistic children, pointing to a different developmental trajectory rather than a simple lag. James McPartland, who led the work at Yale, said the researchers were seeing the expected refinement in non-autistic children but a reduced developmental change in autistic children.
The study adds to a growing body of research showing that face processing in autism is not uniform and may involve differences in both brain structure and brain function. A separate multimodal study published in Nature in 2024 found that structural and functional differences in the fusiform gyrus, a region important for face processing, jointly affect social functioning in autistic people. Other EEG studies have also reported altered neural responses during face tasks, including differences in oscillatory activity and earlier evidence of atypical responses to facial emotion.
The practical implications remain limited for now, but the findings could help researchers develop more objective measures of autism and better understand which children may benefit from support. Clinicians have long used behavioural assessment rather than biological testing to diagnose autism, in part because no routine lab-based test exists. McPartland said the findings represent “a first step in introducing a new biomarker”, although much more work will be needed before such signals could be used in practice.
The study also sits within a broader shift away from treating eye contact as the goal of intervention. Researchers involved in the field have increasingly argued that forcing autistic children to look directly at faces is not an effective or appropriate response. Instead, the new findings may help explain why some autistic children find faces harder to decode and why support focused on understanding emotional expressions may be most useful when introduced earlier in development.
Disclaimer: This content is for informational purposes only and is not intended to be a substitute for professional medical judgment, advice, diagnosis, or treatment.
