Study On Non-linear Dynamics Analysis For Heart Rate Variability

Study of Heart Rate Variability (HRV) can help us to understand the physiological phenomena and pathological physiological mechanism of activity of disease. To attain the goal of clinical application, Feasibility and suitability of non-linear dynamics analysis means for HRV is discussed by analyzing the non-linear dynamics characteristic of HRV signal systematically. By this way, it is proposed to evaluate thecardiac activity and to find some new means and measures offering further earlier clinical diagnose.18 groups of ECG data (2000 to 3000 points) and a group of healthy ECG data (2000 points) are derived from MIT-BIH ECG database. These data are sorted into five classes according to the proportion of the ECG pathological waveform such as normal beat data, fusion of paced and normal beat data, fusion of ventricular and normal beat data, bundle branch block beat data, blocked APB data and etc. In five different classes, first ECG signal is filtered, then decomposed and reconstructed with wavelet transform experimentally. With the method mentioned above, the high precision RR interval sequence (this is HRV signal, too) is finally gained. After summarizing and researching the principle of nonlinear dynamics analysis means, especially from reckoning time sequence (HRV signal) with the method of nonlinear dynamics analysis means with the following analysis technique such as Pioncare section, fractal dimension, lyapunov exponent and approximate entropy, a series of methods are developed which can evaluate the activity and its change to research the HRV system and study behavior of nonlinear dynamics activity of HRV signal. Thereby it comes to a significant conclusion experimentally. (1) Pioncare section, fractal dimension, lyapunov exponent and approximate entropy are adopted to reckon and analyze the HRV signal of healthy people. It shows that these parameters are 6.53,6.42,6.17,6.93 and 6.47 separately (fractal dimension value). The data indicates that HRV is a chaos signal and proves that it is an effective means to analyze HRV signal. (2) Five classes of different pathological HRV data are analyzed relatively, the results show that: 1) The parameters are different from different pathological conditions. 2) The illness parameters are smaller than the normal parameters as a whole. 3) Under the same pathological class, the value drop of parameters indicates an even severe disease. It implies that these indexes not only distinguish normal people from patients but also offer the detailed information of cardiac activity. Also it can offer the more new information for clinic. (3) From the program running efficiency of every analyzing-parameter method (My project is accomplished with MATLAB software tool), running efficiency of pioncare section is highest. The running time of lyapunov exponent is about 5 to 7 minutes. The running time of fractal dimension is slower and the approximate entropy is slowest.The nonlinear dynamics parameters of HRV are an effective index to evaluate the cardiac autonomic nervous system function indirectly. It has the advantages such as non-invasion, high precision and quantitative measurement. It's a valuable means and an important technique of forecasting an acute disease such as coronary heart disease, myocardial infarction and etc. Study of nonlinear dynamics analysis means for HRV has wide clinical application and study foreground.