基于脑电模糊熵的呼吸暂停事件自动检测方法

Automatic detection method for apnea events based on fuzzy entropy of EEG signals

针对医用多导睡眠监测仪不便于家用普及的问题,设计了一种基于单一脑电信号的呼吸暂停事件检测算法,利用巴特沃斯滤波器对脑电信号进行频带分解,采用基于模糊熵的方法在子带信号中提取特征参数,对比呼吸暂停和正常呼吸期间对应脑电模糊熵的变化,建立机器学习分类模型对呼吸暂停事件和正常呼吸事件进行分类,并采用来自2个独立数据库的55名被试的睡眠脑电数据对该方法的有效性进行验证。结果表明:该方法在公共数据和临床数据中分别取得了93.25%和94.50%的准确率,证明了脑电模糊熵可有效地表征呼吸暂停期间的脑电特性,为便携式呼吸暂停检测设备的实现提供了理论基础及技术支持。

In view of that the medical polysomnography is not convenient for home use,an apnea event detection algorithm based on single electroencephalogram signals was designed.The signals were decomposed by the Butterworth filter,and then the characteristic parameters were extracted from the sub-band signals using the fuzzy entropy-based method.After that,the changes of electroencephalogram fuzzy entropy during apnea and normal breathing were compared,and then a machine learning classification model was established to distinguish apnea from normal respiratory events.The total of 55 subjects′sleep electroencephalogram data from two independent databases were used to verify the validity of this method.The results showed that the method achieved 93.25%and 94.50%accuracy in public data and clinical data,respectively.This study demonstrated that fuzzy entropy can effectively characterize electroencephalogram properties during apnea.In addition,the research also provided theoretical basis and technical support for the realization of portable apnea detection equipment.