Noise Analysis of Average Blood Pressure in Terms of Aggregation of First-Order Autoregressive Process Models

Abstract

Blood pressure fluctuations reflect the presence of a variety of dynamic response of the cardiovascular control systems to physiological perturbations to cardiovascular homeostasis. As in most biological processes, these fluctuations exhibit 1/f^α characteristic of the power spectrum under physiologic conditions. In recent studies of human blood pressure variability (BPV), the slope α of the 1/f^α relationship has been found to change in accordance with pathologic alternations of BPV, e.g. a selective loss of high frequency fluctuations associated with the inhibition of the control of arterial BP by means of autonomic nervous system. However, current experiments do not specifically address the origin of the corresponding change in the slope. In the present study, a time-discrete model of 1/f^α fluctuation generator consists of the aggregation of first order autoregressive (AR(1)) processes with different values of process parameters was developed to analyze the relationship between pathological conditions and the values of the BPV spectral exponent α. Results show that the present model well reproduced inverse power-low spectrum lines shifted more negatively in high frequency components, those are specific for BPV spectrum of denerved arteries. Spectral analysis using the proposed model may provide a useful means of identify and/or diagnose diseases those are apparently unrelated to the pathologic states of cardiovascular system.