脑电的非线性检测与双谱分析

The Nonlinear Detection and Bispectral Analysis of EEG Signals

本文应用高阶谱技术构成以微机为核心的脑电非线性检测及双谱分析系统。采用非高斯参数模型对脑电信号进行建模，利用模型参数检测脑电信号的非线性特征，分析脑电的参数化双谱估计，分别讨论双港技术的离散实现和系统的硬件设置。用多种临床脑电验证了该方法的有效性。由于采用了双谱技术，脑电背景噪声的影响得到有效的抑制。本文分析结果表明：高阶谱技术为分析脑电信号和其它医学信号提供了一条新途径。

Higher-order spectral (HOS) technology is applied to the nonlinear detectionand bispectral analysis of EEG signals based on personal computer system. Non-Gaussian ARmodel is proposed to the modelling of EEG signals. The model parameters are used for thenonlinear detection and the parametric bispectral estimation of the EEG signals. The bispectraldiscrete implementation and the analysis system are also discussed. The effect of the back-ground noise is effectively rejected due to the bispectral technology. The results presented inthis paper show that the polyspectral technique proposes a new way for analyzing the EEG andother medical signals.