A new study of hundreds of children reports that autistic brains show less distinct electrical signals for faces than for objects during vision tasks, shifting how scientists think face processing develops in autism.
Story Highlights
- Researchers recorded brain activity across the whole scalp while children viewed faces and objects.
- Autistic participants showed less clearly separable signals for faces than neurotypical peers.
- Face-related brain responses became more specialized with age in neurotypical children but not in autistic children, in this sample.
- Machine-learning tools identified whether a child viewed a face or object from brain patterns, with differences by diagnosis.
What the researchers measured and why it matters
Yale School of Medicine reported results from a large pediatric sample that included about 400 autistic children. Researchers recorded electrical activity from 128 sensors across each child’s scalp while they viewed images of faces and objects. This whole-scalp approach aimed to capture a fuller picture than older studies that used only a few sensors. The team then studied how clearly the brain’s response to faces differed from its response to objects, a key sign of visual specialization in development.
ScienceDaily’s report summarized the core result: brain activity linked to face perception was less clearly distinguishable in autistic participants than in neurotypical children. That means the patterns for “this is a face” and “this is an object” overlapped more in the autistic group. The Yale summary added that signals became more face-specific with age in neurotypical children, while autistic children in the sample did not show the same age-related sharpening. Those results point to a developmental difference, not just a snapshot gap.
How the team tested neural distinctness
The study used a machine-learning method to check how well a computer could tell if a child was viewing a face or an object based on the recorded brain pattern. The classifier worked on signals taken from all 128 electrodes. Differences in classification performance by diagnosis supported the claim that face-related signals were less distinct in the autistic group. Reports also noted evidence that some differences appeared before the well-known N170 response, which occurs about 170 milliseconds after a face appears on screen.
That timing detail broadens the target for future work. Many past studies focused on the N170 as the main marker for face processing. Signs of earlier differences suggest atypical processing could begin sooner in the visual stream. If confirmed, that could guide new tasks and interventions that look at earlier windows in perception. It could also help scientists design tests that pick up subtle differences missed by older, narrower measures.
Where this fits in the broader science
Face research in autism has long shown group differences, yet with mixed results over time. A 2023 meta-analysis of 38 studies and 1,764 participants found a reduced face inversion effect in autistic groups on average, but it also found wide variation across tasks and measures. Large biomarker efforts have urged broader samples, stronger methods, and repeatable tasks to sort out real signals from noise. Whole-scalp recordings and larger cohorts aim at exactly that goal.
The Autism Biomarkers Consortium for Clinical Trials has reported progress with face tasks in electroencephalography and has pushed for measures that are reliable and scalable for clinics. Reviews still warn that small samples and varied methods can stall progress. They recommend standardized protocols and clear reporting so that results can be checked and used in real trials. The new study’s size and methods align with those calls, which helps place the findings in an improving research pipeline.
What the study does not claim
The reports describe brain patterns, not daily behavior. They do not say that less distinct electroencephalography signals equal poorer face recognition in real life. They do not report eye-tracking results, social skills measures, or whether the brain patterns predict day-to-day outcomes. The summaries also do not provide exact effect sizes or the final sample breakdown, which limits judgment of the result’s size until the full paper is reviewed. One short caution is that cross-sectional age patterns are not the same as tracking the same children over time.
For families, teachers, and clinicians, the practical message is careful but hopeful. Larger, more complete brain measures are starting to map how face processing grows in childhood and how it may differ in autism. That map can support earlier help that respects each child’s needs and strengths. It also calls on institutions to back research that is transparent, repeatable, and tied to real outcomes. People across the political spectrum agree on this much: science should serve families directly and clearly.
Sources:
sciencedaily.com, cadena3.com, medicine.yale.edu, ua.news, topics.consensus.app, indiaautismcenter.org
