Toward Differential Diagnosis of Autism Spectrum Disorder using Multimodal Behavior Descriptors and Executive Functions

Publication date: Available online 13 December 2018

Source: Computer Speech & Language

Author(s): Chin-Po Chen, Susan Shur-Fen Gau, Chi-Chun Lee

Abstract

Varied manifestations of social communication deficits, atypical prosody, and restricted and repetitive behaviors are often observed in individuals with autism spectrum disorder (ASD). The pervasiveness and heterogeneity in ASD have made it an increasingly important interdisciplinary research domain. The categorizations in ASD, ie. Autistic Disorder, High-functioning autism, Asperger' syndrome, has varied throughout the past versions of Diagnostic and Statistical Manual of Mental Disorders (DSM) in order to have a better description of ASD. Using computational approach in characterizing these neuro-developmental disorders is, therefore, important for characterizing relevant behavior constructs consistently with potential wide applicability. In this work, we propose to compute signal-derived multimodal behavior descriptors of ASD subjects during dyadic interactions of Autism Diagnostic Observation Schedule (ADOS), and we further examine these behavior features' discriminatory power in differentiating between the three groups in ASD: Autistic Disorder (AD), Asperger' syndrome (AS), and High-functioning autism (HFA). Additionally by combining assessment of ASD subject's executive functions, i.e., measured by Cambridge Neuropsychological Test Automated Battery (CANTAB), the classification accuracy improved further especially on AD versus AS. Finally, we found a moderate correlation between turn-taking duration in our computed behavior features and measures of the Rapid Visual Information Processing in CANTAB.

from Speech via a.sfakia on Inoreader https://ift.tt/2Gf9ENG