Automatic diagnosis of autism spectrum disorders in children through resting-state functional magnetic resonance imaging with machine vision.

Abstract

Background: Autism spectrum disorders (ASDs) are neurodevelopmental disorders characterized by impairments in social interactions, communication, repetitive behaviors, and restricted interests. Magnetic resonance imaging (MRI) has been increasingly used to identify common patterns in individuals with autism for classification purposes. This study aims to develop an intelligent system for diagnosing ASD in children using resting-state functional magnetic resonance imaging (fMRI) and machine learning algorithms.

Methods: This study proposes a method for classifying children with ASD versus healthy control (HC) using resting-state fMRI. This study used images from 26 autistic children and 26 controls, aged 5 to 10 years. Image features were extracted from both groups, and the children with ASD were classified from the HCs using support vector machine (SVM), random forest (RF), K-nearest neighbor (KNN), and artificial neural network (ANN) algorithms.

Results: Our experimental results reveal that the proposed method accurately detects ASD using the ABIDE dataset and achieves accuracy of 88.46%, 73.07%, 82.69%, and 90.38% with SVM, RF, KNN and ANN algorithms, respectively.

Conclusions: Diagnosing autism through clinical evaluations is time-consuming and relies on expert expertise, highlighting the importance of intelligent diagnosis for this disorder. In this study, we developed an intelligent system that demonstrated high accuracy in ASD diagnosis using resting-state fMRI and machine learning techniques.