文献类型：期刊文献

英文题名：SincNet-Based Hybrid Neural Network for Motor Imagery EEG Decoding

作者：Liu, Chang[1];Jin, Jing[2];Daly, Ian[3];Li, Shurui[1];Sun, Hao[1];Huang, Yitao[1];Wang, Xingyu[1];Cichocki, Andrzej[4,5,6]

机构：[1]East China Univ Sci & Technol, Minist Educ, Key Lab Smart Mfg Energy Chem Proc, Shanghai 200237, Peoples R China;[2]East China Univ Sci & Technol, Shenzhen Res Inst, Minist Educ, Key Lab Smart Mfg Energy Chem Proc, Shanghai 200237, Peoples R China;[3]Univ Essex, Sch Comp Sci & Elect Engn, Brain Comp Interfacing & Neural Engn Lab, Wivenhoe Pk, Colchester CO4 3SQ, Essex, England;[4]Skolkovo Inst Sci & Technol SKOLTECH, Moscow 143026, Russia;[5]Polish Acad Sci, Syst Res Inst, PL-01447 Warsaw, Poland;[6]Nicolaus Copernicus Univ, Dept Informat, PL-87100 Torun, Poland

年份：2022

卷号：30

起止页码：540

外文期刊名：IEEE TRANSACTIONS ON NEURAL SYSTEMS AND REHABILITATION ENGINEERING

收录：;EI(收录号：20221011757917);WOS:【SCI-EXPANDED(收录号:WOS:000769969000001)】；

基金：This work was supported in part by the National Key Research and Development Program under Grant 2017YFB13003002, in part by the Grant National Natural Science Foundation of China under Grant 61573142 and Grant 61773164, in part by the Program of Introducing Talents of Discipline to Universities (the 111 Project) under Grant B17017, in part by the Shuguang Project funded by Shanghai Municipal Education Commission and Shanghai Education Development Foundation under Grant 19SG25, in part by the Ministry of Education and Science of the Russian Federation under Grant 14.756.31.0001, and in part by the Polish National Science Center under Grant UMO-2016/20/W/NZ4/00354.

语种：英文

外文关键词：Electroencephalography; Feature extraction; Convolution; Kernel; Convolutional neural networks; Filter banks; Task analysis; Brain-computer interface; motor imagery; SincNet; neural network; gated recurrent unit

摘要：It is difficult to identify optimal cut-off frequencies for filters used with the common spatial pattern (CSP) method in motor imagery (MI)-based brain-computer interfaces (BCIs). Most current studies choose filter cut-frequencies based on experience or intuition, resulting in sub-optimal use of MI-related spectral information in the electroencephalography (EEG). To improve information utilization, we propose a SincNet-based hybrid neural network (SHNN) for MI-based BCIs. First, raw EEG is segmented into different time windows and mapped into the CSP feature space. Then, SincNets are used as filter bank band-pass filters to automatically filter the data. Next, we used squeeze-and-excitation modules to learn a sparse representation of the filtered data. The resulting sparse data were fed into convolutional neural networks to learn deep feature representations. Finally, these deep features were fed into a gated recurrent unit module to seek sequential relations, and a fully connected layer was used for classification. We used the BCI competition IV datasets 2a and 2b to verify the effectiveness of our SHNN method. The mean classification accuracies (kappa values) of our SHNN method are 0.7426 (0.6648) on dataset 2a and 0.8349 (0.6697) on dataset 2b, respectively. The statistical test results demonstrate that our SHNN can significantly outperform other state-of-the-art methods on these datasets.