基于心电信号的睡眠阶段的辨识

本研究以睡眠中的心电图信号为基础,计算出心率变异性的时域和频域参数,根据自律神经系统的变化与心率变异性的关联性,对睡眠的不同阶段进行辨识,从而实现睡眠质量监测的居家化。睡眠质量辨别算法以监督式倒传递类神经网络为核心,通过SDNN、RMSSD、SDSD、NN50、p NN50、HFnorm、VLF百分比、5 min TP等8个特征值,进行睡眠的5个阶段的辨别。实验通过686组数据测试发现,隐藏层神经元数目为30,性能目标为40,为最佳参数设定,其中对睡眠中Stage1阶段的识别率可达93.33%。

Sleep-stage identification based on electrocardiogram

The study calculated the parameters of heart rate variability (HRV) in the time domain and frequency domain based on electrocardiogram (ECG) during sleep and established a sleep-stage identification model through the association between the change of the autonomic nervous balance and HRV,which could be integrated with ECG equipment to achieve the sleep quality evaluation in the smart house.The study identified the five stages of sleep with the supervised back-propagation neural network whose eight parameters were SDNN,RMSSD,SDSD,NN50,pNN50,HF-norm,VLF percentage and 5min TP.Through testing 686 groups of data,the results presented that the number of hidden layer neurons as 30,performance goals as 40 were the best parameter setting,and the recognition rate of Stage1 in sleep phase reached 93.33%.