Time-frequency Analysis Of Ventricular Repolarization Variability Based On ST Interval

According to the statistics of the World Health Organization,there are currently 17million people in the world who die of heart disease each year,which is equivalent to 32people who die of heart disease every minute.There are more than 200 million people with cardiovascular diseases in China.The number of people who die from various types of coronary heart disease is estimated to exceed 1 million every year.The number of people who die of hypertension complications is estimated to exceed 1.5 million.Therefore,the study of ECG signals has important implications for the treatment of cardiac diseases.According to studies,ventricular repolarization variability?VRV?problems are related to myocardial ischemia,hypertension,myocardial infarction and many other diseases,and ventricular repolarization wave variation has been observed in most patients with high-risk arrhythmia..At present,VRV analysis methods mainly include domain analysis,frequency domain analysis,and time-frequency analysis.Because some ventricular diseases show similar trends in the time domain,time domain analysis is prone to errors.Although frequency domain analysis solves this problem,because VRV is non-stationary signal,frequency domain analysis assumes that the VRV signal is Smooth signal,resulting in errors.The time-frequency analysis method analyzes the VRV signal from the perspective of time and frequency domain analysis.In this paper,a method is proposed to remove baseline drift and suppress high frequency interferences based on wavelet transform.After locating the S-waves and T-waves,the ST segment is extracted,and the 1Hz cubic spline interpolation is resampled to obtain the signal for analysis.utilizing Short Time Fourier Transform?STFT?and Wavelet transform,the instantaneous average frequencies of the bands including HF,MF and LF are obtained.Using the above method,the short-time Fourier transform time-frequency analysis was performed on 30 cases of heart disease patients and 10 cases of normal subjects.The average instantaneous frequency of each case was obtained,and the data on healthy individuals and cardiac diseases were obtained.Bilateral t-tests were performed on the data obtained for the healthy and heart disease groups.Statistical analysis showed that there was a significant difference between the two groups of data in the low-frequency part.The p-value was 3.91977*10^-6.The data was far less than 0.05,suggesting that the short-time Fourier transform of ECG can effectively distinguish health in the low-frequency range.Same as above,the above data was analyzed by using wavelet transform.Statistical analysis showed significant differences between the two groups of data in the low-frequency part.The p-value difference was 6.58786*10^-10,which was far less than 0.05,indicating that the wavelet transform was used for time-frequency analysis of ECG signals.It is statistically significant at low frequencies and can be used to identify differences between healthy and diseased groups.It is observed that both time-frequency analysis can effectively discriminate between two groups of samples in the low-frequency part,among which the difference using wavelet transform is more significant.