单次运动想象脑电的特征提取和分类

为了实现脑-计算机接口(BCI)系统,对运动想象脑电信号的特征进行了提取和分类.将大脑C3,C4处采集的二路运动想象脑电信号分成4段,分别建立六阶AR参数模型进行功率谱估计,再对每段数据的功率谱求和构造特征矢量,提供给误差反向传播算法进行左右手运动想象脑电模式分类.结果表明,该方法提取的特征向量较好地反应了运动想象脑电信号的事件相关去同步(ERD)和事件相关同步(ERS)的变化时程.另外,该方法识别率高,复杂性低,适合在线脑-计算机接口的应用.

Feature extraction and classification of single trial motor imagery EEG

In order to realize the brain-computer interface(BCI) system,features of motor imagery EEG(electroencephalogram) were extracted and classified.First,the motor imagery EEG signals sampled from the C3 and C4 position of the brain were divided into four segments.Next,six-order AR parameter model was used for power spectrum estimation of each segmentation EEG and the summation of power spectrum was calculated to construct the feature vector.Then,the error back propagation algorithm was utilized to classify the electroencephalogram pattern of left and right hand motor imagery.The results show that the eigenvector extracted by proposed method effectively reflects the event-related desynchronization(ERD) and event-related synchronization(ERS) time course changes of motor imagery EEG.In addition,the proposed method obtains a high recognition rate and a low complexity so as to be utilized in online BCI system.