Russian Researchers Develop AI Model for Early Autism Diagnosis
The system has already demonstrated diagnostic accuracy of 80%.
Researchers at Plekhanov Russian University of Economics have developed a new AI model designed to improve autism diagnosis. The team used hypergraph neural networks to analyze functional MRI (fMRI) scans. The AI architecture can model complex, multiway interactions between different regions of the brain. The system has already demonstrated diagnostic accuracy of 80%. The experiments used data collected from multiple research centers.
"This significantly outperforms both traditional machine learning methods and conventional graph convolutional networks, matching the best results reported worldwide in this field," said Elena Pitsik, Lead Researcher at the Research Institute of Applied Artificial Intelligence and Digital Solutions at Plekhanov Russian University of Economics.
The researchers believe the new approach opens broad opportunities for advancing the early diagnosis of autism.
