单导联脑电信号P300分类研究

Research on the classification of P300 in single channel EEG

针对单导联脑电(EEG)信号,提出一种基于约束独立分量分析(cICA)的P300特征提取方法。将单导联EEG信号分段及少次叠加平均后形成混合EEG数据,40次靶刺激对应EEG叠加平均构成的P300模板作为参考信号,经cICA算法只分解出一个含P300的独立分量,提取出3维特征向量送入线性分类器。针对实测EEG数据进行方法验证,单导联5试次情况下P300识别正确率达82.64%,信息传输率达16.80 b/m。实验结果表明,提出的方法在单导联较少试次情况下,能够有效提取P300,且对EEG导联位置在顶区范围不严格挑剔。

Aiming at the single channel electroencephalogram （EEG） signal,a feature extraction method for P300 based on constrained independent component analysis （cICA） was proposed.The single channel EEG signal is segmented and averaged with a few trials to compose a mixed EEG data,and the forty EEG epoches corresponding to the target stimuli are averaged to construct a P300 template,which is used as the reference signal.The cICA algorithm is used to pick up only one independent component that contains P300,and a three dimensional feature vector is extracted and sent to the linear classifier.The proposed method is tested on the experimentally measured EEG data; and under the condition of single channel and five trials,the correct recognition rate for P300 reaches 82.64％ and the information transfer rate reaches 16.80b/m.The experiment result shows that the proposed method can extract P300 effectively under the condition of single channel and less trials without strict requirement for the EEG channel position in parietal or central lobe.