The Correlation Dimension of heart rate variability reflects cardiac autonomic activity

Abstract

Background: Linear methods of time series analysis such as summary statistics and frequency domain parameters have been used to measure heart rate variability (HRV). More recently, nonlinear methods including the correlation dimension (CDim) have been used to evaluate HRV. The aim of this study was to examine the effect of autonomic perturbations on the CDim. Methods: The CDim was calculated from 2000 data points (RR intervals) collected over a relatively short period of time (20-40 min) in 12 healthy subjects aged between 20 and 40 years (mean 30 ± 2 years) during: (a) supine rest; (b) head up tilt (sympathetic activation, parasympathetic nervous system activity withdrawal); (c) intravenous infusion of atropine (parasympathetic nervous system activity withdrawal); and (d) following overnight administration of low dose transdermal scopolamine (parasympathetic nervous system augmentation. Results: The CDim was determined at rest (7.8 ± 0.3) and found to be significantly reduced during tilt (5.9 ± 0.4, P < 0.01) and atropine administration (4.2 ± 0.4, P < 0.01) and possibly increased by scopolamine (8.3 ± 0.5, NS). Conclusions: The changes following these interventions suggest that CDim can accurately measure cardiac autonomic nervous system activity.