{"title":"Motor Task Learning in Brain Computer Interfaces using Time-Dependent Regularized Common Spatial Patterns and Residual Networks","authors":"H. Sadreazami, G. Mitsis","doi":"10.1109/newcas49341.2020.9159807","DOIUrl":null,"url":null,"abstract":"This work proposes a method for motor task recognition in brain computer interfaces (BCI). The proposed method is realized by EEG signals classification using time-dependent regularized common spatial patterns and deep residual networks. Unlike other existing methods, the proposed method relies on both the spectral and temporal features by preserving the temporal resolution of the spatially-filtered EEG signals. These features are projected onto an image representation and fed into a residual network for a hierarchical feature learning and classification. Experiments are carried out on benchmark datasets taken from BCI competitions to evaluate the performance of the proposed method and to compare it with other existing methods. The binary classification results of the proposed method demonstrate a superior performance in classification accuracy compared to other existing methods.","PeriodicalId":135163,"journal":{"name":"2020 18th IEEE International New Circuits and Systems Conference (NEWCAS)","volume":null,"pages":null},"PeriodicalIF":0.0000,"publicationDate":"2020-06-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"3","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"2020 18th IEEE International New Circuits and Systems Conference (NEWCAS)","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1109/newcas49341.2020.9159807","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
引用次数: 3
Abstract
This work proposes a method for motor task recognition in brain computer interfaces (BCI). The proposed method is realized by EEG signals classification using time-dependent regularized common spatial patterns and deep residual networks. Unlike other existing methods, the proposed method relies on both the spectral and temporal features by preserving the temporal resolution of the spatially-filtered EEG signals. These features are projected onto an image representation and fed into a residual network for a hierarchical feature learning and classification. Experiments are carried out on benchmark datasets taken from BCI competitions to evaluate the performance of the proposed method and to compare it with other existing methods. The binary classification results of the proposed method demonstrate a superior performance in classification accuracy compared to other existing methods.