Метод выделения значимых компонент для оценки вариабельности формы пульсовых волн

Abstract

Background and Objectives: The conventional approach to the quantification of the pulse wave is based on the assessment of the features of its shape within each beat to beat heart interval. Usually, a set of indices is calculated (such as heart efficiency index, reflection index, stiffness index), which are determined by the reference points of the wave contour. We have developed an alternative method aimed to analyze the variability of the pulse waveform regardless of the variability of its rhythm. A distinctive feature of the method is that the classical spectral analysis tool – the Fourier series expansion in harmonic functions – is used not for frequency analysis, but for describing the features of the pulse waveform, regardless of its frequency-time characteristics. Materials and Methods: The data is represented by the signal as separate sets of fragments, each of which corresponds to one cardiointerval. Further, information about the current heart rate, including its variability, is removed from the data. We do this by resampling each cardiointerval by the same predetermined number of samples. The shape variability of each waveform is quantified by calculating the amplitudes and phases of the harmonics of each waveform as a representative of a strictly periodic sequence. This is an important point, since the original pulse wave signal belongs to the class of random signals and for it, as for the whole, the amplitude and phase of the Fourier spectra are not determined. As a result of the procedure, the analysis of the pulse waveform for each cardiointerval is reduced to the analysis of the amplitude and phases of the required number of harmonics. The method described above was used to process the data of an experiment aimed at quantitatively calculating the relationship between the central and distal pulses. The measurements were carried out on a group of 16 healthy volunteers aged 20–35 years after being in a calm state for 20 minutes. Then, rheographic signals were synchronously recorded from three points on the human skin surface (aortic region, wrist and distal phalanx of the finger). Results: The study revealed significant differences in the stability of the shape of the pulse waves recorded in different parts of the vascular bed, which is expressed in different degrees of variability of their main components. It is important that the central pulse has a smaller number of significant harmonics in comparison with the distal one, and is more stable at the first 4 harmonics containing the main signal power. Conclusion: We proposed and tested a method for analyzing the variability of the pulse waveform, based on the harmonic analysis of the re-sampled signal for each of the cardiointervals, aimed at studying the variability of the pulse waveform separately from the variability of its rhythm. The obtained quantitative data on the stability of the harmonics of the central pulse wave indicate the prospects for further development of the transfer function method in the problem of restoring the shape of the central pulse based on distal measurements.