Classification of Mild Cognitive Impairment Based on Dynamic Functional Connectivity Using Spatio-Temporal Transformer
Jing Zhang, Yanjun Lyu, Xiaowei Yu, Lu Zhang, Chao Cao, Tong Chen, Minheng Chen, Yan Zhuang, Tianming Liu, Dajiang Zhu·January 27, 2025
Summary
A novel framework employing a spatial-temporal transformer for dynamic functional connectivity analysis from rs-fMRI data introduces a contrastive learning strategy to predict mild cognitive impairment, a precursor to Alzheimer's disease. This method outperforms existing approaches, offering potential for early AD identification. By leveraging contrastive learning, the framework enhances robustness and accuracy in MCI prediction, reducing reliance on labeled data. Ablation studies confirm the synergy of temporal and spatial blocks for optimal performance, achieving a 9.7% increase in F1 score and 89.1% overall accuracy compared to current methods.
Advanced features