脑电Laplacian空域滤波的改进算法

Improved EEG Laplacian spatial filtering algorithm

Laplacian空域滤波通过被关注的通道与相邻通道信号转换,能有效地提高脑电信号的信噪比,经典的Laplacian空域滤波矩阵由通道之间的距离确定。该方法依赖于相应通道定位的准确性,实际应用中不能保证电极精准安放并存在个体差异,影响了该方法的应用效果。为了提高该算法的通用性,采用高斯模型改进经典Laplacian空域滤波算法,根据用户调试数据确定模型相关参数,进而确定空域滤波矩阵。采用国际脑机接口竞赛BCI competition 2005 IV数据集验证滤波后的效果,总体比经典Laplacian方法识别准确度提高10个百分比,采用改进方法滤波后不同类间距离增大、同类间距离减小,提高了信号的信噪比。

Through channel and adjacent channels signal conversion,Laplacian spatial filter can effectively improve the EEG signal to noise ratio. Laplacian spatial filtering method was determined by the distance between the channels. The method is tigthly dependent on the accuracy positioning of the corresponding channel,but it is very difficult to put electrode on accuracy position in practical application. Otherwise the method recognition rate sharply declined. In order to improve the robustness of the algorithm,this paper proposed an improved Laplacian spatial filtering algorithm method,which determined spatial Laplacian filter coefficients according to signal calibrating. It used brain-computer interface competition（ BCI competition） 2005 IV data sets to validate the spatial filtering method. The result demonstrates that it improves the signal-to-noise ratio. The method recognition rate has generally increased by 10 percentage than Laplacian method. It demonstrates the improved method increases the distance between different classes,decreases between similar distance and improves the signal to noise ratio.