Self-supervised graph contrastive learning with diffusion augmentation for functional MRI analysis and brain disorder detection

Xiaochuan Wang, Yuqi Fang, Qianqian Wang, Pew-Thian Yap, Hongtu Zhu, Mingxia Liu

Abstract: Highlights•Creating a self-supervised graph contrastive learning framework for fMRI analysis and brain disease detection to alleviate the small-sample-size problem.•Designing a graph diffusion augmentation strategy to preserve the integrity of original BOLD signals.•Conducting extensive experiments on two rs-fMRI datasets with 1,230 subjects.