Simulating healthy human heart rate: a Markovian model

This study is performed within the scope of Computers in Cardiology Challenge 2002 on simulating 24 hours RR interval time series. We construct a computational model to characterize short- and long-term complex dynamics of healthy human heart rate. The cardiac dynamics is simplified via map the increase and decrease of the interbeat interval to 1 and 0, respectively. The probability of transition from current binary sequence to next state is then determined by following two factors: 1. Prior history of binary sequences, and 2. Current value of RR interval. Probability tables were constructed from real data. We used 8-bits and 2-bits binary sequences to simulate short- and long-term heart rate fluctuations. The magnitude of increment was chosen randomly. Finally, we implemented two simple functions to simulate the circadian rhythm and temporal structures of fluctuations during rapid eye movement stage sleep. The model reached a score of 0.689 in event 1 (entry 142). In summary, our preliminary study indicated Markovian model may apply to different levels of physiologic regulations. Further study is needed to examine the correlation with physiologic mechanisms.