基于小波变换的脑电高阶奇异谱分析

Higher Order Singular Spectrum Analysis of EEG Based on Wavelet Transform

奇异谱分析是数字信号分析的一种新方法。信号的奇异谱反映信号的奇异特征。但奇异谱分析方法是基于二阶统计的方法 ,反映的是信号时间上和空间上的一种线性相关关系。因而很难反映非线性信号内在的非线性关系。本文提出一种基于小波变换和高阶统计的奇异谱分析的新方法 ,并将其运用于正常脑电和癫痫患者的脑电信号分析中。实验结果表明 ,正常脑电和癫痫脑电的奇异谱有明显的不同。

Singular spectrum analysis (SSA) is a novel way to analyze digital signals. Singular spectra of signals reflect the singular features of signals. SSA is a method which based on two order statistic and reflects the linear correlation on spacetime of signals. So the intrinsic nonlinear correlations of nonlinear signals are difficult to be reflected by SSA. A new method based on wavelet transform and higher order singular spectrum analysis (HSSA) is proposed in this paper, and we use this method for analysing EEG for both normal subjects and epileptic subjects. The results show that the singular spectra of normal subjects are significantly different from that of epileptic subjects.