一种基于WRCE-CSP的运动想象电位识别方法

针对脑一机接口系统中运动想象脑电信号（Electroencephalography，EEG）的模式识别问题，提出了加权节律成分提取（WeightedRhythmicComponentExtraction，WRCE）与共空间模式（CommonSpacePattern，CSP）相结合的特征提取方法，并使用Fisher线性判别分析进行分类。采用2003年的BCI竞赛数据Datasetm对该方法进行评估，测试数据的分类正确率达到86．13％，比使用传统CSP方法进行特征提取时的分类正确率提高了5．71％，表明该方法可有效地应用于运动想象EEG的模式识别中。

WRCE-CSP Based Pattern Recognition Method for Motor Imagery Potential

According to the motion imagery Electroencephalograph(EEG) pattern recognition in Brain-Computer Interface(BCI),the paper puts forward the weighted rhythmic component extraction based common space pattern method to the feature extraction,and uses Fisher Linear Discriminate Analysis(Fisher-LDA) for classification.Applying Data set Ⅲ of BCI competition 2003 to evaluate this method,the classification accuracy of test data is 86.13%,5.71% better than the classification accuracy of common space pattern feature,which proves that this method is effective in EEG pattern recognition of motion imagery.