多导信号的相空间重构及在脑电信号分析中的应用

A Multi-channel Phase-space Reconstruction Method and its Application to EEG

本研究提出了一种基于时间延迟和序列相关性的多导信号相空间重构方法。根据信号的相关系数重组多导联信号序列,并利用信号的可确定性选取时间延迟参数重构相空间。对仿真数据的研究表明这种重构方法在信号的确定性和相关维数计算中具有良好的性能,受噪声、延迟量和嵌入维数等参数变化和单导重构相比影响较小,结果更稳定和可靠。对实验思维脑电数据的非线性分析得到了脑电和思维复杂性,关联性等性质的联系。该方法适用于短时多导含噪信号的非线性分析和脑电的在线研究等。

A multi-channel phase-space reconstruction method was proposed in the paper based on rearranging signal serials by the correlation coefficients and selecting time delay by signal determinism. The study of determinism and calculation of correlation dimension on simulative data indicated the recoustraotion method exhibit a good performance to the noise disturbance, the preferences of time delay and embedding dimension, the analysis results were stable and reliable. The nonlinear analysis of mental EEG gave the relationship between EEG signals and mental complexity. This method can be useful for analyzing multi-channel signals with noise, short time series and real-time EEG on-line.