振幅整合脑电图在正常睡眠脑电分期中的应用  

Sleep staging in normal sleep:An amplitude-integrated EEG analysis

作　　者：叶大田[1,2] 李慧[1,2] 彭诚[2] 

机构地区：[1]清华大学生物医学工程系,北京100084 [2]清华大学深圳研究生院,深圳518055

出　　处：《清华大学学报（自然科学版）》2014年第8期1098-1104,共7页Journal of Tsinghua University(Science and Technology)

基　　金：康复工程与健康科技深港联合创新平台资助项目(ZYB200907070060A)

摘　　要：该文提出了振幅整合脑电图用于正常年轻人睡眠脑电分期的方法。记录了13例正常年轻人约8小时睡眠脑电数据,分为训练组(6例)和测试组(7例)。计算训练组每一例的振幅整合脑电图(aEEG);提取aEEG的上边带曲线作为其特征曲线;提取不同分期的aEEG上边带中位数和四分位距特征;将这些特征进行综合统计分析,得出aEEG在不同睡眠期的边界和波动范围的数值指标;利用此指标对训练组和测试组的脑电数据进行睡眠自动分期。测试组和训练组的分期结果与ZEO系统结果有较好的一致性,证明了aEEG的一组特征值作为睡眠分期决策指标的可行性。This paper describes a sleeping staging analysis method based on the amplitude-integrated EEG（aEEG）.Single channel EEG of 13 young volunteers was recorded for eight hours during their normal sleep with the volunteers divided into a training group（6cases）and a test group（7 cases）.The aEEG signals were obtained from the original EEG signals.The upper margins of the aEEG signals were extracted as the characteristic curve,with the medians and interquartile ranges of the upper margin calculated for different sleeping stages as the characteristic parameters for sleeping stage discrimination. A statistical analysis gave indices for the characteristic parameters for different sleeping stages for an automatic sleeping stage analysis system.Experimental verification indicated that this method is consistent with ZEO system results.Thus,this method based the aEEG characteristic parameters holds promise for sleeping stage analysis in clinical applications.

关 键 词：脑电图(EEG) 振幅整合脑电图(aEEG) 睡眠分期 上下边带 四分位数 

分 类 号：R319[医药卫生—基础医学]