Multi-channel spatio-temporal graph attention contrastive network for brain disease diagnosis

Chaojun Li, Kai Ma, Shengrong Li, Xiangshui Meng, Ran Wang, Daoqiang Zhang, Qi Zhu

Abstract: Highlights•A deep graph model is introduced for DBNs analysis, enabling higher-order spatio-temporal information propagation.•A multi-modal method integrates structural and functional brain connectivity into a unified graph representation.•A graph attention network with contrastive loss improves spatio-temporal feature discrimination.•Our method outperforms state-of-the-art approaches on epilepsy and ADNI datasets, and demonstrates potential in identifying neurological biomarkers.