A Machine Learning Approach for Grading Autism Severity Levels Using Task-based Functional MRI

Autism is a developmental disorder associated with difficulties in communication and social interaction. Autism diagnostic observation schedule (ADOS) is considered the gold standard in autism diagnosis, which estimates a score explaining the severity level for each individual. Currently, brain image modalities are being investigated for the development of objective technologies to diagnose Autism spectrum disorder (ASD). Alterations in functional activity is believed to be important in explaining autism causative factors. This paper presents a machine learning approach for grading severity level of the autistic subjects using task-based functional MRI data. The local features related to the functional activity of the brain is obtained from a speech experiment. According to ADOS reports, the adopted dataset is classified to three groups: Mild, moderate and severe. Our analysis is divided into two parts: (i) individual subject analysis and (ii) higher level group analysis. We use the individual analysis to extract the features used in classification, while the higher level analysis is used to infer the statistical differences between groups. The obtained classification accuracy is 78% using the random forest classifier.

I. Introduction

Autism spectrum disorder (ASD) is a neuronal developmental disorder associated with a range of symptoms that differ in severity as social, sensorimotor, and communicative deficits [1]–[3]. ASD is diagnosed at the age of three, but some characteristics can be noticed as early as 12 months, Especially with the progress of medical imaging and the latest state-of-the-art machine learning approaches [4], [5]. Different modalities such as, structural magnetic resonance imaging (sMRI), functional magnetic resonance imaging (fMRI) and diffusion tensor imaging (DTI), are widely used for analysing brain's structural and functional characteristics [6], [7].