ECG Signal Analysis Based On Two Kinds Of Phase Entropy

As an important organ of vertebrate,the heart occupies an extremely important position in the structure of human body,but cardiovascular and cerebrovascular diseases are the first of the three kinds of chronic non-communicable diseases that cause the most deaths.In recent years,cardiovascular diseases have received more and more attention.Electrocardiogram(ECG)is regarded as an important reference for the detection of diseases in clinical medicine.Due to the nonlinear characteristics of ECG signals,nonlinear analysis of ECG signals has become a research hotspot.In the previous nonlinear algorithms,the main method to explore the information exchange between the two systems is the transfer entropy,and the algorithm performs well in detecting the information flow of the dynamic system.The common algorithm is permutation entropy,which is simple and convenient,and has a good degree of discrimination.It is widely used in biological signals,speech signals and other related fields.However,the above two algorithms analyze the signal characteristics from the aspect of signal amplitude,and phase synchronization and amplitude correlation are two independent variables.Related studies show that phase mode contains more information than amplitude mode coding.Therefore,this paper studies the information representation method based on phase information and analyzes the advantages of two kinds of phase entropy in ECG signal detection:The main work and innovations of this paper are as follows:First of all,the phase transfer entropy algorithm was first applied to the analysis of ECG signals,and the effectiveness of the phase transfer entropy algorithm was verified using analog signals.The effect of signal length on the phase transfer entropy of healthy people in MIT-BIH ECG database is analyzed.The variation of phase transfer entropy with age is analyzed by using phase transfer entropy.The phase transfer entropy of ECG signal in two patients with arrhythmia was simulated.The research shows that exploring phase transfer entropy can detect abnormal ECG signal to a certain extent.Secondly,according to the basic principle of permutation entropy,a phase permutation entropy algorithm with more sequence features is introduced,the principle and calculation process of the algorithm are given.The effect of the permutation entropy algorithm and the phase permutation entropy algorithm on the detection effect of the abrupt signal is illustrated intuitively by the simulation signal.Permutation entropy and phase permutation entropy are applied to the analysis of three kinds of arrhythmia ECG signals.The results show that,compared with the permutation entropy,the phase permutation entropy shows a more obvious jump to the simulation signal and ECG mutation,and can better distinguish the change of ECG signal.Thirdly,the characteristics of phase transfer entropy and phase permutation entropy in signal characterization are compared and analyzed,and the effects of noise on phase transfer entropy and phase permutation entropy are observed by changing the SNR.The research results show that phase permutation entropy is more sensitive to noise,while phase transfer entropy has better robustness.Comparing the entropy values of two phase entropies in detecting the pre-onset and away from the onset of atrial fibrillation,and the detection of atrial tachyarrhythmia with two phase entropies,the advantages and disadvantages of the two phase entropies in the detection of electrocardiographic signals are obtained.It is one of the key issues to excavate the dynamic features contained in the phase of ECG signal to detect lesions better and more sensitively.The results of this paper have some significance for the nonlinear detection of ECG signals.