Longitudinal study of early mild cognitive impairment via similarity-constrained group learning and self-attention based SBi-LSTM
作者：
Highlights：
• We devise a similarity-constrained group sparse network (SGN) for multi-center brain functional connectivity network (BFCN) construction, which can learn the similarity among features and reduce the useless features simultaneously.
• We design a SBi-LSTM framework to utilize the longitudinal information for EMCI detection.
• We explore a self-attention mechanism to find the most discriminative features to improve the detection performance.
摘要
•We devise a similarity-constrained group sparse network (SGN) for multi-center brain functional connectivity network (BFCN) construction, which can learn the similarity among features and reduce the useless features simultaneously.•We design a SBi-LSTM framework to utilize the longitudinal information for EMCI detection.•We explore a self-attention mechanism to find the most discriminative features to improve the detection performance.
论文关键词：Mild cognitive impairment,Similarity-constrained group learning,SBi-LSTM,Self-attention,Longitudinal study
论文评审过程：Received 11 November 2021, Revised 20 June 2022, Accepted 13 July 2022, Available online 22 July 2022, Version of Record 23 August 2022.
论文官网地址：https://doi.org/10.1016/j.knosys.2022.109466
