小鼠视觉感受区电位信号(LFP)与视觉刺激之间关系研究

Research on Correlation Between LFP and Visual Stimulation in Mice's Visual Perception Area

针对小鼠脑电波与行为之间关系研究中涉及弱脑电波信号处理、成分分离及提取、脑电波与对应行为分析等问题,首先处理了脑电波信号,根据呼吸机理建立了"呼"、"吸"动作电位触发下的脑电波模型.其次,通过独立分量分析方法获得睡眠状态小鼠视觉感受区LFP信号模型,采用区间估计和Rayleigh熵检验结合讨论了提取的相关脑电波与呼吸信号的锁相性,分析了LFP信号的周期性,建立了多元线性回归模型,用最小二乘法从理论上讨论了其线性关系,并检验了所建二次模型周期变化的相关性.然后,通过平均经验模态分解及基于互相关性的伪分量检验算法,改善了模态混叠,建立了完整的脑电波信号分离模型.最后,分析图形因素的图像特征,通过提取视觉刺激曲线、功率谱相关频段,采用叠加平均信号法,以LFP节律随视觉信号锁相程度判断不同视觉刺激引起的LFP成分变化.

This paper mainly did research on correlation between LFP and visual stimulation in Mice＇s visual perception based on the processing of weak-signal.First,on the basis of respiratory mechanism,the present study processed the collected brain signals（electroencephalogram,EEG） and established a EEG triggered by action potentials of expiration and inspiration.Next,this study sought to acquire the LFP model of sleeping mice＇s visual receptive fields via independent component analysis（ICA）.Combined interval estimation with Rayleigh entropy test,we discussed the phase-locking of extracted related brainwave and respiratory signal.We adopted wavelet transformation to analyze the periodicity of LFP signals,built a multiple linear regression model,discussed their linear relationship theoretically using the least squares method and tested the correlation of periodic variation of the built quadratic model.In addition,the modal aliasing was improved and a complete EEG Separation Model was established based on Ensemble Empirical Mode Decomposition（EEMD） and Pseudo Component Test on the basis of cross correlation.Finally,image features were extracted and analyzed.A method called spike-triggered average method（STA method） was used to deal with the extracted visual stimulation curve and the related frequency band of power spectrum.Changes of LFP components aroused by different visual stimulation were judged by the extent of the phase-locking correlation between LFP rhythm and the visual signals.