I. Introduction

HEART rate and arterial blood pressure variability signals constitute a unique probe for the assessment of cardiovascular control mechanisms [i.e., primarily of autonomic nervous system (ANS)] in both physiological and clinical conditions. There have been a lot of studies [1] (and references therein) on the control mechanisms on the basis of cardiovascular variability signals, using the standard power spectrum density functions [2]. Usually 256–512 consecutive cardiac cycles [one cardiac cycle being given by the R-R interval obtained in electrocardiogram (ECG) recording] are taken for analysis. Previous studies [1] have shown that the high-frequency fluctuations around the respiratory rate (around 0.2-0.4 Hz in healthy human subjects) are mediated by the parasympathetic nerves of the ANS and the low-frequency fluctuations (around 0.1 Hz) are mediated by both the sympathetic and parasympathetic nerves. Conventional Fourier analysis can only provide the global or averaged information over the entire duration of the data. Information over the evolution of the spectral components cannot be obtained. So Fourier analysis on short data segments have been used to capture the temporal evolution of the different spectral components. Such analyses depend heavily on the length of the data segments, since shorter time intervals would decrease the resolution of the spectrum, whereas the longer duration affects the stationarity of the time series. Compared to the voluminous research efforts of heart rate variability (HRV) signals, less attention has been paid to the beat-by-beat blood pressure waveform variations. Ohtomo et al. [3] provided the three-dimensional spectral array of the beat-by-beat blood pressure signal using maximum entropy spectral estimation. They found that in healthy subjects the frequencies corresponding to the spectral peaks fluctuate in an erratic fashion which reflects chaoticity. In some pathological cases [e.g., subjects with supraventricular premature contraction (SVPC)], new spectral components lower than the fundamental mode appear, which is caused by the intermittent occurrences of SVPC. Sherebrin et al. [4] analyzed photo-plethysmographic signal in frequency domain and determined how the pulse profile varied with the age; the decrease of power in the harmonics of the peripheral pulse profile with age was proposed to be an useful noninvasive measure of aging and vascular disease.