基于EMD和样本熵的脑电信号睡眠分期研究

摘要

对于睡眠脑电的人工分期仍有不足的问题，提出一种基于脑电节律经验模态分解和样本熵结合的睡眠分期方法。首先利用经验模态分解对睡眠脑电信号进行去噪和基本节律提取，获得针对不同睡眠状态下脑电节律的样本熵，通过统计其样本熵均值，发现脑电节律δ波和θ波于不同睡眠状态下，样本熵均值不相等且方差较小，能有效表示不同睡眠阶段的特征，说明样本熵方法能较好的实现脑电信号的特征提取，用来区分睡眠各分期，为利用睡眠进行监护与医学治疗提供了一种可行的方法。%A method of sleep stages,based on EMD and sample entropy of EEG rhythm,is put forward to solve the shortage of artificial staging of sleep EEG.Firstly,EMD is used to denoise sleeping EEG and extract the basic rhythm.Secondly,the sample entropy of EEG rhythm under different sleep states is computed.Finally,after the statistics of mean and variance of the sample entropy,the mean of sample entropy of wave δand wave θis different under different sleeping stages of EEG rhythm,which indicates that the sample entropy analysis of sleep EEG rhythm can be used to effectively characterize different sleeping stages and extract feature of EEG.It provides a new way to make use of sleeping EEG for care and treatment.