Normalized correlation dimension for heart rate variability analysis.

In this paper we use the concept of large-scale dimension densities to analyze heart rate variability data. This method uses a normalized Grassberger-Procaccia algorithm and estimates the dimension in the rather large scales of the system. This enables us to analyze very short data. First we re-analyze data from the CIC 2002 challenge and can completely distinguish between real data and computer-generated data using only one parameter. We then analyze unfiltered data for 15 patients with atrial fibrillation (AF), 15 patients with congestive heart failure (CHF), 15 elderly healthy subjects, and 18 young healthy subjects. This method can completely separate the AF group from the other groups and the CHF patients show significant differences compared to the young and elderly healthy volunteers. Furthermore, differences are evident in the dimensionality between day and night for healthy persons, but not for the CHF patients. Finally, the results are compared to standard heart rate variability parameters.