Neural Response to Interactive Faces Is Associated with Clinical Characteristics in ASD and Typical Development

Background: Individuals with autism spectrum disorder (ASD) tend to exhibit atypical gaze and neural responses to faces. Most studies have focused on perception of static faces, although social difficulties in ASD are most apparent in interactive contexts. The P100 and the N170 are event-related potentials that chart static facial-processing, but their connection to the clinical phenotype during reciprocal facial-processing is unknown.

Objectives: The current study investigated the relationship between neural responses to interactive faces and clinical characteristics. Using integrated eye-tracking (ET) and electroencephalography (EEG), we measured visual attention and neural response (P100, N170) to dynamic faces in ASD and typically-developing (TD) children. We sought to identify relationships between neural responses and clinical characteristics. We hypothesized that atypical facial-processing patterns would correlate with higher levels of clinical dysfunction.

Methods: Participants included school-aged children matched on age and IQ with ASD (n = 19) or TD (n = 19). EEG was recorded with a 128-channel sensor-net, and ET was recorded with an EyeLink-1000 remote camera system. In a gaze-contingent paradigm, participants viewed an arrow that cued them to look at the eyes or mouth of a face that then dynamically responded by opening either its eyes or mouth. Social function was measured using multiple standardized measures including the Multidimensional Anxiety Scale for Children (MASC), Childhood Anxiety Sensitivity Index (CASI), Vineland Adaptive Behavior Scales 2^nd Edition (VABS-II), and Social Responsiveness Scale Parent Report (SRS-P). 

Results: Preliminary data indicated that, when viewing reciprocal eye gaze, the TD group (but not the ASD group) exhibited significant correlations between higher levels of anxiety (MASC Total and CASI Average scores) and greater left hemisphere P100 (r = 0.713, p = 0.002) and N170 (r = -0.619, p = 0.01), respectively. In addition, only in the TD group, greater social adaptive functioning (VABS Social Standard Score) was correlated with greater left hemisphere P100 (r = 0.658, p = 0.002). Lesser impairment in cognition (SRS-P Cognition T-Score) was correlated with greater P100 in the right hemisphere for TD (r = -0.504, p = 0.039) but in the left hemisphere for ASD (r = -0.492, p = 0.045).

Conclusions: This study employed an innovative gaze-contingent paradigm to more realistically assess neural responses to simulated social interactions and their relationship with the clinical phenotype. Individuals with higher anxiety may be more vigilant in anticipating (P100) and recognizing (N170) eye contact in the TD group. In addition, greater social functioning predicted increased attention (P100) toward eyes in the TD group. The lack of these relationships in the ASD group is consistent with a lack of neural specialization in processing social information. Finally, lateralization differences may reflect different functional brain circuitry between groups. Data collection is ongoing, and analyses in progress examine visual attention in relation to brain activity and clinical characteristics in these groups and an additional group of unaffected siblings. Preliminary findings indicate potential benefits of using ET-EEG biomarkers to develop better diagnostic and treatment options.