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

Zahra Khandan Khadem-Reza ^1,2, Reza Ahmadi Lashaki ³, Mohammad Amin Shahram ^1,4

⁰Medical Physics Research Center, Mashhad University of Medical Sciences, Mashhad, Iran.

Quantitative Imaging in Medicine and Surgery +

|

July 3, 2025

View abstract on PubMed

Summary

This study developed an intelligent system using resting-state fMRI and machine learning to accurately diagnose autism spectrum disorder (ASD) in children, achieving up to 90.38% accuracy. This offers a faster, data-driven approach to ASD diagnosis.

Area Of Science

Neuroscience
Medical Imaging
Artificial Intelligence

Background

Autism spectrum disorders (ASDs) are neurodevelopmental conditions affecting social interaction, communication, and behavior.
Magnetic resonance imaging (MRI) is explored for identifying patterns in individuals with autism.
Current clinical diagnosis of ASD is time-consuming and requires expert evaluation.

Purpose Of The Study

To develop an intelligent system for diagnosing ASD in children.
To utilize resting-state functional magnetic resonance imaging (fMRI) data.
To apply machine learning algorithms for classification.

Main Methods

Classified children with ASD versus healthy controls (HC) using resting-state fMRI.
Analyzed data from 26 autistic children and 26 controls (ages 5-10).
Employed Support Vector Machine (SVM), Random Forest (RF), K-Nearest Neighbor (KNN), and Artificial Neural Network (ANN) algorithms.

Main Results

The system achieved high accuracy in ASD detection on the ABIDE dataset.
Accuracies were 88.46% (SVM), 73.07% (RF), 82.69% (KNN), and 90.38% (ANN).
Artificial Neural Network (ANN) demonstrated the highest classification performance.

Conclusions

An intelligent system using resting-state fMRI and machine learning shows high accuracy for ASD diagnosis.
This approach offers a potentially more efficient and objective method for identifying ASD in children.
The findings highlight the utility of advanced computational techniques in neurodevelopmental disorder research.

Keywords:

Autism spectrum disorder (ASD) children functional magnetic resonance imaging (fMRI) machine vision resting-state