基于脑电和肌电多特征的自动睡眠分期方法

发布时间：2018-06-15 13:18

本文选题：睡眠分期 + 脑电　；参考：《计算机工程》2017年10期

【摘要】：为实现准确的自动睡眠分期,且满足泛化能力的需求,基于脑电(EEG)和肌电(EMG)多特征,提出一种自动睡眠分期方法。以MIT-BIH多导睡眠数据库中样本的EEG和EMG为分析对象,采用离散小波变换对原始数据进行滤波预处理,提取EEG的α,β,θ,δ节律波和高频成分的能量比,利用样本熵算法提取EEG的非线性特征。将特征参数输入支持向量机分类器中进行样本训练与分类识别。实验结果表明,该方法的分期准确率可以达到92.94%,相比基于EEG的睡眠分期方法平均准确率提高3.96%,交叉验证平均准确率达82.68%,具有较好的泛化能力。
[Abstract]:In order to realize accurate automatic sleep staging and meet the requirement of generalization ability, an automatic sleep staging method was proposed based on the features of EEG and EMG. The EEG and EMG of the samples in MIT-BIH polysomnography database were analyzed. The original data were filtered by discrete wavelet transform, and the energy ratios of 伪, 尾, 胃, 未 rhythm waves and high frequency components of EEG were extracted. The nonlinear feature of EEG is extracted by sample entropy algorithm. Input feature parameters into support vector machine classifier for sample training and classification recognition. The experimental results show that the accuracy of this method can reach 92.940.Compared with the sleep staging method based on EEG, the average accuracy is increased 3.96, and the average accuracy of cross-validation is 82.68, which has better generalization ability.
【作者单位】：上海大学机电工程与自动化学院;中国科学院苏州生物医学工程技术研究所;
【基金】：国家自然科学基金(61433016) 苏州市科技计划项目(ZXY201427,ZXY201429)
【分类号】：R740;TN911.6

【相似文献】