A nonlinear signal processing approach to model heart rate variability

Mathematically modeling and generating the time series (RR-intervals) for heart rate variability (HRV) has been an on-going research activity for some time. This is of use, not just in artificial electrocardiogram (ECG) generation, but also in order to both gain an insight into the heart's operation and for disease diagnosis. First presented in 1972, the Zeeman equations (which model the beating of the heart) were an important contribution to this research area. But some biologists may disagree with aspects of the proposed model-e.g., because there is no consideration of sympathetic and parasympathetic influences on the heart rate. So in this paper, we propose new developments to the original Zeeman equations (as regards the sympathovagal balance), in order to bring them closer to the biologist's idea of a suitable model for heart rate generation. Finally, simulations illustrate these improvements in the resultant HRV modeling.