Quantification of Poincare' maps for the evaluation of heart rate variability

Poincare' plots (each normal RR interval is plotted against its next one) have been proposed for the study of heart rate variability. The plots submitted to independent observers for visual scoring were found to fall into four distinct patterns. The authors have quantified this method by identifying each cluster with: (a) geometrical descriptors, that is the radii of its central ellipse of inertia, whose axes coincide with the directions of major symmetry of the cluster; (b) image distribution parameters, that is the number of frequency peaks in the plot and their distance from a 45 degree's diagonal; (c) scanning parameters, that are the width and the limits of the plot moving from low to high values of RR intervals.<<ETX>>