Search

Purpose of the study. To develop a method for analyzing the seasonal and cyclic components of R-R intervals by estimating the autocorrelation function (ACF). Material and methods. The study consists of conducting a practical analysis of ACF as an application to R-R intervals, divided into three groups: all patients are patients with atrial fibrillation. The first group had tachycardia; the second had bradycardia; the third group - are patients without specific features. It should be noted that the age of the patients, according to which the analysis was carried out, is from 40 years and older. Results. The module developed at VBA (Visual Basic for Applications) enables to apply the methods of primary processing of the studied data, to create the ACF with its graphic representation for rather large values of the lag variable τ=1,100. Samples of 1000 values were examined. The most optimal data cleaning methods have been selected to get rid of excessive noise. A detailed analysis of ACF was carried out in the three groups of patients. Impressive research statistics have been compiled. Conclusions. In people with signs of tachycardia, all ACF ratios were not significant. In cases with severe tachycardia, ACF coefficients were positive, and had a decreasing trend. In the series with signs of bradycardia, almost all ACF coefficients were significant. ACF decreased with increasing lag τ. It can be concluded that in this group of time series there is a marked constant linear trend. In the group of patients with normal heart rate dynamics, insignificant ACF coefficients were obtained, which indicates the absence of any tendency. The coefficients were concentrated along the axis of the lag, had both positive and negative values. The oscillating process indicates only a strong stochastic component of the studied data.

Background. Today, when most processes and phenomena have a periodic structure that is not always homogeneous and depends on a number of irregular events, the classical spectral analysis based on the Fourier transform does not always allow an objective and reliable investigation. A promising alternative to the classical spectral analysis can be a wavelet transform, based on a well-localized function in both time and frequency domains, which allows us to analyze irregular processes. The purpose of the research was to analyze the potentials of the wavelet transform for the analysis of the spectral characteristics of cardiological time series with some heterogeneity; to develop a software application that allows performing wavelet analysis of data and building wavelet spectra, which are illustrative and easy in interpreting the obtained results. Material and methods. 37 cardiological time series were studied. We developed an application for the application package Mathematika 5.0 that allows building and analyzing the wavelet spectra of the studied data. Results. An alternative method for analyzing the spectral characteristics of the studied cardiological time series has been developed and clearly demonstrated. It allows observing the dynamics of the change in periods over the entire observation interval without performing the interval splitting, as would be required by the classical Fourier analysis. Conclusion. For the studied groups of patients the identified periods, found with the help of the wavelet transform have been presented. The advantages of using wavelets over the classical spectral analysis, which in cases of unstable time series does not allow tracking the dynamics of the changing in the data spectrum, have been demonstrated.