摘要
探索了小波包崎和BP神经网络在识别左右手想象运动中的作用。采用脑一机接口2003竞赛中Graz科技大学提供的脑电数据,计算C3、C4电极8~16Hz频带脑电信号的小波包墒,将其作为反应想象左右手运动的特征量,用BP神经网络对大脑想象左右手运动任务进行分类,最大分类正确率可达88.57%,与使用线性判别式算法分类结果相比,效果更好。脑电信号小波包熵随时间的变化与事件相关去同步和事件相关同步现象相一致,可在线识别左右手想象运动,为大脑运动意识任务的特征提取及肢残患者的临床康复提供了新思路。
It is to explore the role of BP neural network and wavelet packet entropy ( WPE ) in recognition of left and right hands imagination movements. The WPEs of C3 and C4 electrodes between 8-16Hz were calculated respectively using the data of Graz University of Technology provided in BCI Competition 2003, and were defined as the feature vector reacting to the left and right hands' motions imagined. The left and right hands motor imaginary tasks were classified by BP neural network, the satisfactory results were obtained with the highest classification accuracy of 88.57% ,it is better than the classification using linear discriminant algorithm. The WPE of EEG changing with time was coincident with event-related desynchronization and event-related synchronization. The left and right hands motor imaginary tasks could be recog- nized on line, and this provides new approaches for exacting the feature of brain motor consciousness tasks and for the clinical rehabilitation of the disabled patients.
出处
《计算机应用与软件》
CSCD
2009年第8期78-81,共4页
Computer Applications and Software
基金
甘肃省高等学校研究生导师科研项目计划(0710-05)
关键词
脑电信号
特征提取
小波包熵
BP神经网络
分类
electroencephalogram (EEG) Feature extraction Wavelet packet entropy BP neural network Classification