Related Experiment Video
Updated: Jun 16, 2026
Identification of Disease-related Spatial Covariance Patterns using Neuroimaging Data
Published on: June 26, 2013
MAN-GNN: An interpretable biomarker architecture for neurodevelopmental disorders
Qiulei Han1, Hongbiao Ye2, Miaoshui Bai3
1College of Computer Science and Technology, Changchun University, Changchun, 130022, China; Ministry of Education, Key Laboratory of Intelligent Rehabilitation and Barrier-Free for the Disabled, Changchun University, Changchun, 130022, China; Jilin Provincial Key Laboratory of Human Health Status Identification and Function & Enhancement, Changchun, 130022, China; Jilin Rehabilitation Equipment and Technology Engineering Research Center for the Disabled, Changchun, 130022, China.
This study introduces a novel graph neural network framework to analyze neuroimaging data for neurodevelopmental disorders. The approach enhances understanding of brain activity, improving diagnostic accuracy for conditions like ADHD and ASD.
Area of Science:
- Neuroscience
- Artificial Intelligence
- Medical Imaging
Background:
- Neurodevelopmental disorders (NDDs) present overlapping symptoms and heterogeneity, hindering understanding of underlying neurobiology.
- Current diagnostic methods rely on subjective reports, limiting insight into inter-patient variability.
- Accurate neurobiological markers are crucial for NDDs like Attention-Deficit/Hyperactivity Disorder (ADHD) and Autism Spectrum Disorder (ASD).
Purpose of the Study:
- To develop a graph neural network (GNN) framework integrating neuroimaging data for NDDs.
- To enhance the analysis of nonlinear, temporal, and dynamic features in brain neural activity.
- To improve classification accuracy and interpretability for NDDs.
Main Methods:
- Incorporation of the Neurodynamics Rössler system to simulate nonlinear brain activity from static neural signals.
- Integration of spatial and topological brain network features for enhanced discrimination.
- Application of adaptive controllers and cross-site adversarial learning to improve noise resistance and generalization.
Main Results:
- The GNN framework demonstrated superior classification accuracy compared to existing methods.
- The model showed significant improvements in identifying neurobiological differences in NDDs.
- Experimental validation on ADHD and ASD datasets confirmed the framework's effectiveness.
Conclusions:
- The proposed GNN framework offers a promising approach for neuroimaging biomarker research in NDDs.
- The framework provides enhanced interpretability, aiding in the auxiliary diagnosis of NDDs.
- This method advances the understanding of neurobiological mechanisms underlying NDD heterogeneity.
