脑机接口中基于SOBI的EEG预处理

EEG preprocessing method based on SOBI in BCI

目的针对脑机接口(brain computer interface,BCI)中脑电信号(electroencephalography,EEG)包含的伪迹以及信号源可能服从多个高斯分布,本文提出一种基于二阶盲辨识(second-order blindidentification,SOBI)的盲源分离去除伪迹方法。方法首先,含有伪迹的多个导联EEG信号采用联合近似对角化和数据白化,计算出混合矩阵,同时分解成数目相等的若干个独立分量。然后,根据伪迹信号特有的直观特性,将分解出含有伪迹的独立分量置零,剩余分量通过混合矩阵,进行逆向投影重构,得到去除伪迹后EEG信号。最后,对3名实验者的实验数据,从处理时间和识别精度两方面进行检验。结果本文中提出的SOBI方法相比于常用的独立成分分析(independent component analysis,ICA),在单个样本处理时间上,分别缩短了1691 ms、1770 ms和2308 ms;在识别精度上,分别提高33%、5%和10%。结论 SOBI能快速有效地去除伪迹信号,为BCI中EEG的在线处理奠定了基础。

ObjectiveFor the artifact signal of electroencephalography（ EEG） in brain computerinterface（ BCI）, this paper presents an artifact removal method based on second-order blind identification（SOBI） in blind source separation.MethodsFirstly,joint approximate diagonalization and data whitening areutilized for multiple-channel. Meanwhile the mixing matrix is calculated and these EEG signals are decomposedinto an equal number of independent component. Then,some independent components containing artifacts needto be set zero based on experience. And the remaining components are reversely projected and reconstructedwith the mixing matrix to obtain EEG signals that artifacts are removed. Finally,the proposed method is testedfrom two aspects including the processing time and recognition accuracy based on three sets of experimentaldata.ResultsThe proposed method has better performance than the commonly used independent componentanalysis（ICA）. The processing time of one trial is shortened by 169 1 ms,177 0 ms and 230 8 ms,and therecognition accuracy is increased by 3 3%,5 % and 10%.ConclusionsThe proposed SOBI can quickly andeffectively remove artifact signals,which may lay the foundation for online processing of EEG in BCI.