Title: Autism detection using machine learning

Authors: U.B. Mahadevaswamy; Rohan Ravikumar; Rachana Mahadev; Kadaparthi Varun Rao; K.S. Anurag

Addresses: Department of Electronics and Communication Engineering, Sri Jayachamarajendra College of Engineering, Mysuru-570006, Karnataka, India ' Department of Electronics and Communication Engineering, Sri Jayachamarajendra College of Engineering, Mysuru-570006, Karnataka, India ' Department of Electronics and Communication Engineering, Sri Jayachamarajendra College of Engineering, Mysuru-570006, Karnataka, India ' Department of Electronics and Communication Engineering, Sri Jayachamarajendra College of Engineering, Mysuru-570006, Karnataka, India ' Department of Electronics and Communication Engineering, Sri Jayachamarajendra College of Engineering, Mysuru-570006, Karnataka, India

Abstract: Autism is one of the most eminent bio-neurological developmental disorders characterised by disablement in social-communication and behavioural aspects like repetitive behaviours and restrictive interests. Present methods of autism detection are based on behavioural analysis of the subject. These methods are dependent on the skills of the examiner and are laborious. This paper proposes a method for autism detection by virtue of structural MRI of brain obtained from the ABIDE-2 dataset and well-trained machine learning models. The prediction was done based on supervised learning methods namely Random Forest and Logistic Regression. These methods yielded accuracy of 0.82 and 0.80 respectively which are comparatively better than previous works. VBM analysis was conducted on the same dataset with SPM8 in order to obtain the difference in brain structure of typical controls and autism affected subjects and a resultant glass brain image was obtained.

Keywords: autism; ABIDE-2; s-MRI; VBM; voxel based morphometry analysis; random forest; logistic regression.

DOI: 10.1504/IJBRA.2021.117947

International Journal of Bioinformatics Research and Applications, 2021 Vol.17 No.4, pp.375 - 387

Received: 03 Jul 2019
Accepted: 30 Jun 2020
Published online: 05 Oct 2021 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article