Many methods have been proposed during the last two decades for the assessment of autonomic nervous system function by quantification of the heart-rate variation (HRV). Relatively little has been written about the HRV in relation to physiological models of the heart-rate regulation. The integral pulse frequency model (IPFM) is a simple model that describes the genesis of heartbeats under the influence of the autonomic nervous system. By comparing simulated HRV data generated with the IPFM model with data from healthy volunteers we found similarities indicating that the model accurately reflects real data. Furthermore, we found a considerable difference between HRV measurements based on beat-by-beat heart-rate and measurements based on the heartbeat interval. Our results suggest that the commonly used electrocardiographic RR interval representation of heart-rate variation might possess an inherent nonlinear, mean heart-rate-dependent property indicating that analysis directly based on RR intervals can give biased results with respect to the underlying autonomic activity. The conclusions embrace all measurements of variation that are directly based on RR intervals including simple indices as well as higher-level quantification such as spectral analysis.