基于心率变异率特征值的心律失常评估研究

Study of Algorithms and Analytical Evaluation Based on Eigenvalues of ECG Signals

心率变异率作为一种基于心电信号的疾病分析方法,是临床医学上具有重要参考价值的参数指标。该文深入研究了心率变异率特征值提取的时域分析方法、频域分析方法和非线性分析方法,针对心律失常的特点,在时域特征值中引入了pNNx等心率变异率指标,在非线性特征值中引入了多尺度样本熵。采用PKU-IMS心电数据库的窦性心律数据与MIT-BIH数据库的心律失常数据,提取了窦性心律组与心律失常组的心率变异率特征值,当选取95%的置信区间时,时域分析的特征值nn50,pnn50,nn100和pnn100,频域分析的特征值vlfp, lfp, hfp和lf2hf,以及非线性分析的特征值τ≥4时,可显著区分窦性心率与心律失常。由于心率变异率分析也应用于糖尿病、脑血管、呼吸系统等疾病的辅助诊断,因此,该心率变异率特征值分析方法有望推广至相关疾病的评估。

Heart rate variability is a method of disease analysis based on ECG signals and an important reference parameter in clinical medicine. The time domain analysis method, frequency domain analysis method and non-linear analysis method of extracting the HRV eigenvalues are thoroughly investigated. According to the characteristics of arrhythmia, the pNNx are introduced in the time domain eigenvalues and the multi-scale sample entropy is introduced in the non-linear eigenvalues. Using sinus rhythm data from the PKU-IMS ECG database and arrhythmia data from the MIT-BIH database, the characteristic value of HRV of the sinus rhythm group and the arrhythmia group were extracted. When 95% confidence intervals were selected, the eigenvalues of nn50,pnn50,nn100 and pnn100 in the time domain analysis, the eigenvalues of vlfp, lfp, hfp and lf2 hf in frequency domain analysis, and the eigenvalues of τ which greater than or equal to 4 in the nonlinear analysis can distinguish sinus heart rate and arrhythmia significantly. Since the analysis of HRV is also applied to the auxiliary diagnosis of diabetes, cerebrovascular and respiratory diseases, the analysis method of characteristic value of HRV studied is expected to be extended to the evaluation of related diseases.