AUTHOR=Feng Quan , Huang Yongjie , Long Yun , Gao Le , Gao Xin 

TITLE=A Deep Spatiotemporal Attention Network for Mild Cognitive Impairment Identification

JOURNAL=Frontiers in Aging Neuroscience

VOLUME=Volume 14 - 2022

YEAR=2022

URL=https://www.frontiersin.org/journals/aging-neuroscience/articles/10.3389/fnagi.2022.925468

DOI=10.3389/fnagi.2022.925468

ISSN=1663-4365

ABSTRACT=Mild cognitive impairment (MCI) is a clinical state that serves as an early warning for Alzheimer's disease (AD). The subtle and slow changes that exist between the brain structures of MCI patients and normal controls (NCs) make them lack effective diagnostic methods. Therefore, the identification of MCI is a challenging task. However, the prediction of human brain tissue structure by functional brain network (FBN) analysis is a new method emerging in recent years, which provides sensitive and effective medical biomarkers for the diagnosis of neurological diseases. Therefore, to address this challenge, we propose a novel deep spatiotemporal attention network (DSTAN) framework for MCI identification based on brain functional networks. Specifically, we first extract spatiotemporal features between brain function signals and FBN by designing a spatiotemporal convolution strategy (ST-CONV). Then, on this basis, to further capture the brain nodes with strong correlation with MCI, we introduce an attention mechanism for learning. Finally, we fuse spatiotemporal features for MCI identification. The whole network is trained in an end-to-end manner. Extensive experiments show that our proposed method outperforms the current baselines and state-of-the-art methods significantly, achieving a classification accuracy of 84.21%.