Abstract Attention to faces and eye contact are key behaviors for establishing social bonds in humans. In Autism Spectrum Disorders (ASD) a neurodevelopmental disturbance characterized by poor communication skills, impaired face processing and gaze avoidance are critical clinical features for its diagnosis. The biological alterations underlying these impairments are not clear yet. Using high-density electroencephalography coupled with multi-variate pattern classification and group blind source separation methods we searched for face- and face components-related neural signals that could best discriminate neurotypical and ASD visual processing. First, we isolated a face-specific neural signal in the superior temporal sulcus peaking at 240ms after stimulus onset. A machine learning algorithm applied on the extracted neural component reached 74% decoding accuracy at the same latencies, dissociating the neurotypical population from ASD subjects in whom this signal was weak. Further, by manipulating attention to face parts we found that the signal-evoked power in neurotypical subjects varied as a function of the distance of the eyes in the face stimulus with respect to the viewers’ fovea, i.e. it was strongest when the eyes were projected on the fovea and weakest when projected in the retinal periphery. Such selective face and face-components neural modulations were not found in ASD individuals although they showed typical early face related P100 and the N170 signals. These findings show that dedicated cortical mechanisms related to face perception set neural priority for attention to eyes and that these mechanisms are altered in individuals with ASD.

A neural marker of eye contact highly impaired in autism spectrum disorder