A Fusion-Based Machine Learning Approach for Autism Detection in Young Children Using Magnetoencephalography Signals

Kasturi Barik; Katsumi Watanabe; Joydeep Bhattacharya; Goutam Saha

doi:10.1007/s10803-022-05767-w

A Fusion-Based Machine Learning Approach for Autism Detection in Young Children Using Magnetoencephalography Signals

J Autism Dev Disord. 2023 Dec;53(12):4830-4848. doi: 10.1007/s10803-022-05767-w. Epub 2022 Oct 3.

Authors

Kasturi Barik¹, Katsumi Watanabe², Joydeep Bhattacharya³, Goutam Saha¹

Affiliations

¹ Department of Electronics and Electrical Communication Engineering, Indian Institute of Technology Kharagpur, Kharagpur, India.
² Faculty of Science and Engineering, Waseda University, Tokyo, Japan.
³ Department of Psychology, Goldsmiths, University of London, London, UK. j.bhattacharya@gold.ac.uk.

Abstract

In this study, we aimed to find biomarkers of autism in young children. We recorded magnetoencephalography (MEG) in thirty children (4-7 years) with autism and thirty age, gender-matched controls while they were watching cartoons. We focused on characterizing neural oscillations by amplitude (power spectral density, PSD) and phase (preferred phase angle, PPA). Machine learning based classifier showed a higher classification accuracy (88%) for PPA features than PSD features (82%). Further, by a novel fusion method combining PSD and PPA features, we achieved an average classification accuracy of 94% and 98% for feature-level and score-level fusion, respectively. These findings reveal discriminatory patterns of neural oscillations of autism in young children and provide novel insight into autism pathophysiology.

Keywords: Autism spectrum disorder; Biomarker; Brain oscillations; Classification; MEG; Preferred phase angle.

MeSH terms

Autism Spectrum Disorder* / diagnosis
Autistic Disorder* / diagnosis
Brain
Child
Child, Preschool
Humans
Machine Learning
Magnetoencephalography / methods