Autoregressive parameterization of cardiovascular control: Comparison with Fourier techniques

A model for cardiovascular regulation is used to analyze time series data using causal and non-causal techniques. Although noncausal Fourier techniques such as cross-spectral analysis are useful in characterizing physiologically open-loop processes, such as the effect of respiration (RESP) on heart rate (HR), they are difficult to interpret in a feedback containing system, such as the closed-loop relationship between HR and blood pressure (BP). Using autoregressive moving average (ARMA) analysis, postural changes led to an increase in arterial baroreflex gain that was observed at frequencies <0.15 Hz and a broadband decrease in mechanical feedforward gain. These changes were not well described by noncausal cross-spectral analysis.