Gaze Contingent Games to Modify Neural Response to Eye Contact in ASD

Background: Decreased eye contact represents a common manifestation of social dysfunction in autism spectrum disorder (ASD). Prior electrophysiological research on gaze perception in ASD has focused on passive observation of social information, failing to address the interactive nature of eye contact. This research has identified atypicalities in event-related potential (ERP) indices of face processing in individuals with ASD. Behavioral treatments designed to improve eye contact are resource intensive.  Automated intervention methods, such as computer games, are potentially more efficient but have historically lacked the contingency and synchrony that characterize true social interactions and are considered requisite for social skills treatments.

Objectives: The current study integrated eye tracking (ET) and electroencephalographic (EEG) recording in the context of an interactive, gaze-mediated computer game with the goal of modifying neural responses to reciprocal eye contact in ASD. We predicted that (1) eye movements to faces and gaze-cued objects would change in the context of a rewarding game and that (2) neural responses to faces and eyes would be enhanced as the game shaped more appropriate gaze behavior. 

Methods: EEG was recorded from 10 high-functioning adolescents with ASD (mean age = 13.3) and 10 matched TD controls (mean age = 14.9) using a 128 electrode net. ET was recorded with a remote eye tracker (Eyelink-1000). Trials began with a peripherally-presented crosshair followed by a centrally-presented face. Contingent upon the participant’s fixation, the face responded by looking at the participant or looking to a target in one of four corners of the screen.  Participants earned points for maintaining eye contact or following gaze. ERPs were time-locked to face movement; the N170 was extracted from occipital electrodes and the P300 was extracted at central electrodes.

Results: Individuals with ASD displayed an attenuated N170 response to eye movement relative to TD peers (-.08μV versus -6.0μV , respectively) and failed to display differential neural response to direct versus averted gaze. TD individuals exhibited more negative N170s (-3.2μV) compared to individuals with ASD (-1.9uV) when viewing faces with reciprocal gaze. ET indicated that, overall, individuals with ASD looked less to the eyes than TD participants (p =.01). Compared to TD counterparts, individuals with ASD looked between rather than at the eyes (p = .08), and they did not show an expected bias to look to the eye contralateral to their dominant eye (p=.016). Analyses in progress investigate gaze shift and neural response to eye contact across game trials. 

Conclusions: Results revealed a marker of shared gaze that is attenuated in individuals with ASD. Individuals with ASD displayed reduced sensitivity to socially relevant changes in gaze even when these changes were contingent on their own behavior. Specifically, individuals with ASD showed reduced fixation to the eyes; however, even when looking to the eyes, they exhibited aberrant patterns of brain activity. This experiment holds promise for (1) evaluating the utility of automated, socially interactive behavioral interventions and (2) assessing brain activity during intervention, with the potential to modify interventions in-vivo in response to brain activity and behavior.