Dynamic coarsened spatio-temporal graph convolutional networks for fMRI classification of addiction-induced sleep disorders

Substance use disorder is frequently characterized by persistent sleep disorders that hinder cognitive recovery. Accurately identifying these disruptions requires methods capable of tracking both spatial interactions and the temporal evolution of brain activity. While functional magnetic resonance imaging (fMRI) offers valuable insights into broad neural patterns, conventional fMRI classification methods typically rely on static connectivity and predefined knowledge. These conventional approaches may fail to account for the complex, time-varying nature of brain networks, thereby posing significant challenges for accurate fMRI classification. To overcome these limitations, this paper proposed the Dynamic Coarsened Spatio-Temporal Graph Convolutional Network, a data-driven framework designed to model the brain as a time-evolving directed graph and extract highly distinct data features. The framework achieves this by integrating two key components: a dynamic causal learning module that maps evolving connections over time, and a noise-reducing Gaussian mixture model. Experimental results indicate that the proposed approach outperforms existing methods in classifying fMRI data related to addiction-induced sleep disorders, offering a robust solution for high-dimensional fMRI analysis in clinical neurology.