幂律调制多尺度熵及其在生物信号分析中的应用研究

Power-law modulated multiscale entropy and its application research in biological signal analysis

目的:探讨幂律调制多尺度熵(pMSE)在生物信号分析中的应用效果及其对疾病诊断和康复的临床价值。方法:结合信号的自相似和分形理论,对多尺度熵(MSE)算法的不足进行改进,提出pMSE算法,并在三种生物信号(添加了不同程度仿真白噪声的生理信号;健康、充血性心力衰竭和心房颤动患者的心跳间期信号;清醒和疲劳脑电信号)中分别进行测试和对比,评估pMSE算法的应用效果。结果:pMSE有效改进了MSE的一些不足,在仿真噪声信号分析中获得了更加正确和清晰的比较结果,在3种心跳间期信号分析中取得了更加准确和直观的区分效果,在清醒和疲劳脑电信号分析中取得了更高的分类准确率(t=2.30,P<0.05)。结论:pMSE在多种生物信号分析中取得了比MSE更好的效果,可以更准确地度量生物信号复杂度,对临床疾病诊断和康复具有潜在的应用价值。

Objective:To investigate the application effect of power-law modulated multi-scale entropy(pMSE)in the analysis of biological signal and its clinical value in diagnosis and rehabilitation of disease.Methods:Combined the self-similarity and fractal theory of signal to improve the disadvantages of MSE and put forward the pMSE algorithm.And the pMSE was measured and compared in three kinds of signals(physiological signals in that simulative white noise of different levels added,interval signals of heartbeat from healthy people,patients with congestive heart failure(CHF)and patients with atrial fibrillation,electroencephalogram(EEG)of patients in waking and fatigue state)respectively so as to assess the applied effect of pMSE algorithm.Results:The proposed pMSE effectively improved some disadvantages of the MSE,and it obtained more correct and clear comparative results in the analysis of simulative noise signals,and it also obtained more accurately and intuitively differentiated effects in the analysis of three kinds of heartbeat interval signals,and it also gained higher accuracy of classification in the analysis of the EEG signals of patients in waking and fatigue state respectively(t=2.30,P<0.05).Conclusion:Compared with MSE,the proposed pMSE can obtain better effects in the analysis of various biological signals and it can measure the complexity of biological signals more accurately.Therefore,it has potentially application value in diagnosis and rehabilitation of clinical disease.