一种分析睡眠脑电图的时频特征图

使用Choi-Williams分布对一段睡眠脑电图（EEG）信号进行时频变换,利用局部频谱的特征估计各个时间间隔里的波形,并得到局部频谱的特征曲线,整段EEG信号中所有时间咪上的频谱特征曲线组成一种时频特征图,使用该时频特征图分析睡眠EEG,不仅能够统计该段EEG信号中各种基本波形的出现情况,而且可以观察EEG信号中每个基本波形的变化方式,通过时频特征图对采集的实际睡眠EEG数据进行分阶实验,结果表明,时频特征图可以作为一种分析睡眠EFG有效工具,有良好的应用前景。

Time-Frequency Characteristic Map for Analyzing SleepElectroencephalogram

By computing the time frequency distribution of a segment of sleep EEG signals in Choi Williams method, the basic EEG patterns in each interval of the signals are recognized based on the characteristic of local spectrum, and the characteristic curves of the local spectrum are obtained. The combinations of all characteristic curves of the local spectra in the EEG signals make up a time frequency characteristic map (TFCM). Analyzing the sleep EEG with application of TFCM, the statistic of the basic patterns in a segment of the EEG signals may be estimated, and the variation of the basic patterns can be observed. The staging result of the sleep EEG data by using TFCM shows that the method works well.