睡眠-觉醒节律紊乱下注意诱发脑电特征分析与识别

EEG Characteristics Analysis and Recognition Induced by Attention Under Sleep-wake Rhythm Disorder

目的研究特殊职业人员睡眠-觉醒节律紊乱下执行空间注意定向任务时脑电信号的变化特性。方法实验室模拟36 h完全睡眠剥夺和自然生物节律紊乱实验,采集20名受试者完成静息态任务和空间注意定向任务的绩效数据和脑电数据,然后对绩效数据和脑电数据进行分析并采用线性判别式分析法、最近临近法、支持向量机法对脑电数据进行分类。结果在完全睡眠剥夺后受试者的反应时显著延长、正确率显著降低(P<0.05),在自然生物节律紊乱后反应时和正确率都有所恢复,正确率没有恢复到常规水平;与常规环境相比,36h完全睡眠剥夺环境和自然生物节律紊乱环境下注意力水平和注意方向的脑电样本熵特征均发生变化,基于径向基核函数的支持向量机分类器(SVM-R)的准确率更高。结论在睡眠剥夺和自然生物节律紊乱环境下大脑复杂度会变小;样本熵脑电特征提取方法和SVM-R分类方法可以作为特殊职业人员执行注意任务时的个体作业脑电变化检测及分类的有效方法。

Objective To study the characteristics of EEG signals of special occupations when performing spatial attention orientation tasks in sleep-wake rhythm disorders.Methods The experiments of total sleep deprivation and natural biological rhythm disorder were simulated in the laboratory,and the per-formance data and EEG data of were collected when subjects completed resting state tasks and spatial attention orientation tasks.Then the data were analyzed and classified by classification methods,which was LDA,KNN and SVM.Results After total sleep deprivation,the reaction time(RT)of the subjects was significantly prolonged and the correct rate(CR)decreased significantly.After natural biological rhythm disorder occurred,both RT and CR recovered to a certain extent,but the level of CR did not return to the normal.Compared with the routine environment,under the 36-hour total sleep deprivation environment and natural biological rhythm disorder environment,sample entropy features of attention level and attention direction changed,and the accuracy of feature classifier and support vector machine classifier(SVM-R)based on radial basis kernel function was higher.Conclu-sion In conditions of sleep deprivation and natural biological rhythm disorder,the complexity of brain will be reduced.Sample entropy EEG feature extraction method and SVM-R classification method can be used as effective methods to detect and classify the EEG changes for special professionals when at-tention tasks are performed.