Project Details

A system for social training and progress tracking for children with Autism Spectrum Disorder

Haik Kalantarian, MS, PhD

Summary

BACKGROUND: Autism spectrum disorder (ASD) is a developmental disability affecting 70 million children worldwide, including 750 thousand American children under the age of ten. Many children with ASD struggle to make eye contact, recognize facial expressions, and engage in social interactions with their peers. GAP: The increasing incidence of this condition and limited resources have resulted in long waitlists that delay diagnosis and treatment beyond the timeframe in which intervention can have maximum impact. HYPOTHESIS: Our hypothesis was that a charades-style mobile game, Guess What?, can address social deficits through a form of mobile Discrete Trial Training and Pivotal Response Treatment. METHODS: Our study included 21 children (7 ± 1.7 years) who played for an average of 22 sessions over a period of four weeks. We used the SRS-2 as our primary outcome measure. RESULTS: The study demonstrated strong preliminary evidence of high participant engagement (adherence across the four weeks was over 100%), and a measurable impact on standard measures of socialization. These sessions also generated more than 400 videos, amounting to over 10 hours of social video and a host of app and sensor data. IMPACT: These encouraging outcomes demonstrate that Guess What? provides important therapeutic gains to core deficits of autism predictive of long-term positive quality of life outcomes while generating data to support progress tracking for families and iterative machine-learning model enhancement for precision healthcare. Continued development of this platform will facilitate the process of tracking progress on a more granular scale, and give way to more efficient and personalized treatment. Website Link: guesswhat.stanford.edu

Publications:

Kalantarian, Haik, Khaled Jedoui, Peter Washington, Qandeel Tariq, Kaiti Dunlap, Jessey Schwartz, and Dennis P. Wall. "Labeling images with facial emotion and the potential for pediatric healthcare." Artificial intelligence in medicine 98 (2019): 77-86.

Kalantarian, Haik, Khaled Jedoui, Peter Washington, and Dennis P. Wall. "A Mobile Game for Automatic Emotion-Labeling of Images." IEEE Transactions on Games (2018).