脑电逆问题中求逆方式与信号去噪的研究

Study on inversing algorithm and noise removing in the inverse problem of EEG

提出了一种空间总集全局寻优的求逆方式，改进了在脑电逆问题的研究中，传统的时空全局寻优求逆及静态求边方式的不足。为了能够利用单次或少量几次刺激对脑电诱发响应信号进行定位求逆，提出了空间因子求逆方式及相关噪声抵消法，有效地消除了脑电诱发响应中强背景噪声的影响。仿真研究表明，利用这些方法，在信噪比为－１０ ｄＢ的情况下，仍能对诱发响应信号进行准确的定位和提取。

In this paper, a new algorithm called spatial ensemble optimizing method is presented which reduces the sensitivity to noise in the study of EEG's inverse problem as compared to conventional temporal-spatial optimizing method and static inversing algorithm. In order to use a singletrial or a few number of trials to localize the source of evoked potential, a spatial factor inversing algorithm and a coherent noise eliminating method are proposed which effectively reduce the affect of the strong background noise. By simulation, these novel methods give encouraging results when the SNR is about - 10 dB.