Autism Spectrum Disorder Studies Using fMRI Data and Machine Learning: A Review

Machine learning methods have been frequently applied in the field of cognitive neuroscience in the last decade. A great deal of attention has been attracted to introduce machine learning methods to study the autism spectrum disorder (ASD) in order to find out its neurophysiological underpinnings. In this paper, we presented a comprehensive review about the previous studies which applied machine learning methods to analyze the fMRI data of autistic individuals and the typical controls from all-round process, including features constructed from original fMRI data, feature selection methods, involved machine learning methods and factors for high classification accuracy, as well as critical conclusions. Applying different machine learning methods and fMRI data acquired from different sites, classification accuracies were obtained ranging from 60% up to 97% while informative brain regions and networks were determined and discussed. Through thorough analysis, higher classification accuracies were found to usually occur in the studies which involved task-based fMRI data, single dataset for some selection principle, effective feature selection methods or advanced machine learning methods. Though current studies are definitely challenging from multi-site datasets, continuous efforts on them were made to reveal more information about ASD. In the future, developed feature selection and machine learning methods combined with multi-site dataset or easy-operated task-based fMRI data appear to have the potentiality to serve as a promising diagnostic tool for ASD, which requires joint and unremitting efforts of ASD researchers all over the world.