Using gaussians to model increment distribution of the long-term R-wave interval in healthy humans

The fluctuation in the long-term R-wave interval recording can range from a few milliseconds to approximately 1 s in healthy humans. The large fluctuation has been characterized by the 1/f-like power spectrum. This implies scale invariance symmetry of the signal, which is closely related to the family of probability density function of the increment process. In this paper, the distribution of the increment of the R-wave interval was studied. The increment of the R-wave intervals typically evolves from a stretch-exponential-like distribution in the small time interval to a near gaussian distribution in the large time interval. The model proposed by Hausdorff and Peng, consisting of summing a set of gaussian components, can be used to explain the transition. However, it requires further modification in order to capture the experimental data properly. The new model suggests, among other things, multiple scaling in HRV.