Nonstationary Heart Rate Variability in Respiratory Tests

Abstract

A new method was proposed for processing a nonstationary heart rate by using frequency-modulated signals rather than amplitude-modulated signals equally spaced over several points of time as in the conventional method. A frequency-modulated signal is a set of identical Gaussian peaks that coincide with the true time points of heart beats. A continuous wavelet transform was used to quantitatively describe the heart rhythm signal. A test with controlled breathing was performed as an example and included three consecutive stages: rest, rhythmic breathing at a specified frequency, and exhalation. Tachograms recorded during the breath test was found to be a nonstationary signal with the alternation of peaks of different spectral ranges. A system of quantitative parameters was developed to describe the dynamics of changes in the spectral properties of the tachogram in transitional areas. A static clustering by the effect of the respiratory test and a dynamic clustering in order to identify the time points when the autonomic nervous system is stressed were performed for all subjects. The article discusses the prospects of using the method as a means to analyze the transient effects in various functional tests and as biofeedback that would help to change the heart rhythm.