锋电位检测信号的多元小波去噪方法研究

Multivariate Wavelet Denoising Method for Neuronal Spike Signals

神经元锋电位(spike)信号是研究大脑信息编码的基础,具有宽带、高频和小幅值特点,易受噪声干扰。为了提高检测信号的信噪比,根据微电极阵列记录信号中通道之间噪声相关性较强的特点,采用多元小波去噪方法对锋电位检测信号进行了噪声抑制,并基于仿真和实测数据将其与主成分(PCA)去噪算法、小波-PCA联合去噪算法进行了比较。仿真和实测数据结果表明,多元小波去噪方法不仅可以有效提高spike检测信号的信噪比,而且可以降低spike波形的畸变,为小幅值spike信号的检测和下一步分析研究奠定了良好的基础。

Spikes which are the basis of the research of brain information are sensitive to noise because of broadband, high frequency and small amplitude signals. Based on the strong correlations among the noises in different channels, a new denoising method, multivariate wavelet denoising method, was developed for improving signal-to-noise ratio （SNR） of the spike signals. The proposed methods were evaluated and compared with both the principal component analysis （PCA） denoising method and PCA-wavelet combined denoising method using both real and simulated data sets. The result of simulation and real data shows that this method can not only improve SNR but also reduce spike waveform distortion, and that it is important for the detection and the next step analysis research of spikes.