Leveraging Reward Mechanisms in ASD Behavioral Treatment: A Computer Game Intervention

Background: Consensus is growing that differences in motivation and reward mechanisms are central to the etiology of ASD. These differences not only affect the development of social skills in ASD, but also diminish the effectiveness of behavioral treatments (which typically rely on reward to influence behavior). There is significant untapped potential in using computer game platforms to facilitate social skill enhancement in ASD, as children with ASD (and neurotypicals) generally find these platforms highly rewarding.

Objectives: This presentation focuses on the development and testing of an eye gaze-controlled space adventure game designed to enhance social abilities in ASD. Initial piloting focuses on 1) user experience data, 2) within-game metrics of social skill changes, and 3) the development of naturalistic eye contact assessments for a subsequent clinical trial.

Methods: This National Science Foundation-supported initiative represents a partnership between the Children’s Hospital of Philadelphia, University of Pennsylvania, and BioStream Technologies, LLC. This team has developed an immersive, interactive space adventure game that uses consumer-grade eye tracking technology to teach children to make and hold eye contact with game characters. To date, the game has been administered to 99 children with and without ASD (ages 4 to 16), across a total of 269 game sessions, including 52 children with ASD across a total of 172 game sessions.

Results: User experience data suggest a high level of engagement and favorable responses to the game. Preliminary analyses on a subset of data (N = 18) indicate that within-game metrics are useful indicators of social skill, with ASD individuals showing diminished eye contact during initial, incidental interactions with game characters. Preliminary data also indicate that the game’s facial expression exercises yield increased emotion recognition scores after 4 or 5 gameplay sessions (p < .001, n=24). Analysis of mutual eye gaze data (during a live social interaction) are ongoing, and involve computer vision and machine learning analyses of video data.

Conclusions: BioStream’s game has proven to be safe and well tolerated by all participants to date. The vast majority of participants liked the game, which is essential to assuring that it can be played for a sufficient duration to facilitate behavior changes. Preliminary results suggest that the game provides social skill metrics that have significant potential as change measures (as well as cross-sectional assessment). Next steps in this line of research include the analysis of novel outcome data, and the implementation of a Phase II clinical trial.