基于复杂性测度的睡眠脑电分期处理方法研究  

A Sleep EEG Segmentation Process Based on Complexity Measure

作　　者：张泾周[1] 马颖颖[1] 李婷[1] 周钊[1] 苗治平[1] 

机构地区：[1]西北工业大学自动化学院,西安710072

出　　处：《中国生物医学工程学报》2009年第3期367-371,共5页Chinese Journal of Biomedical Engineering

摘　　要：提高临床脑部及神经系统疾病的早期诊断水平,及时予以控制和治疗,是降低脑疾病对人类危害最有效的途径。探讨复杂性测度在睡眠脑电分期中的应用,主要利用加窗的Lempel-Ziv复杂度处理算法对经采用小波变换滤波算法滤除生理干扰后的睡眠脑电信号进行分期处理,并与没有加窗的Lempel-Ziv复杂度处理算法的仿真处理结果进行比较。结果表明:加窗的Lempel-Ziv复杂度算法能更好地将睡眠脑电不同状态分开,在一定程度上减少由脑电的非平稳性带来的计算上的片面性,同时兼顾各期睡眠脑电状态的不均匀性,在很大程度上满足临床的应用要求。It is an effective way to decrease the incidence of brain illness by improving the diagnostic efficiency of diseases of clinical cerebrum and nervous system and controlling the state of illness in early stage. This paper investigated the application of complexity measure in sleep EEG segmentation. Firstly, we eliminated the physiological interference in the EEG signal such as sharp pulse, baseline drift and the other high-frequency signal interference by wavelet transform. Secondly, based on the theory of coarse graining processing, the EEG signal sequences were transformed to binary sequences. Finally, the time-window Lempel-Ziv complexity measurement was used to identify different stages of sleep EEG, and the results were compared with that of common complexity as well. To a certain extent, the proposed method not only reduced the unilateralism of calculation brought by the nonstationarity of EEG but also took into account the unevenness of the states. The accuracy for distinguishing the sleep EEG stages was improved.

关 键 词：睡眠脑电 复杂性测度 粗粒化处理 睡眠分期 

分 类 号：R318[医药卫生—生物医学工程]