LAUSR

An increasing number of studies argue for the necessity of an objective method to quantify autism spectrum disorder (ASD) symptoms to replace time-consuming assessments and interviews by trained clinicians. Several attempts at digital phenotyping have been reported, some of which have succeeded in predicting ASD risk based on quantified symptoms by machine learning. Color information analysis is a promising tool for digital phenotyping of ASD symptoms. The potential field of application ranges from detection of atypicality in autonomic function and pupillary response to evaluation of gastrointestinal symptoms. However, some technical and ethical hurdles remain that hamper the ready application of color analysis to digital phenotyping of ASD.