Quantitative metrics of social response for autism diagnosis

Abstract

Social robots recognize and respond to human social cues with appropriate behaviors. The capabilities used to build social robots can be uniquely applied to assist in the diagnosis and treatment of autism, a pervasive developmental disorder which results in selective impairment of social abilities. This paper outlines some of the ways in which social robots can provide unique perspectives to address critical problems in diagnosing autism. We provide preliminary data and observations on how this result can be achieved based on three years of immersion in a clinical research group that performs diagnostic evaluations of more than 130 children per year.