Heart rate variability characterization using correlation dimension

Heart rate variability (HRV) has become a useful tool for analyzing cardiovascular autonomic control from the electrocardiogram (ECG). Detection and analysis of HRV allows a quantitative and noninvasive method to obtain reliable and reproducible information on autonomic modulation of heart rate. Nonlinear methods (e.g., correlation dimension) are promising tools for HRV assessment. This paper presents a reliable estimation of correlation dimension (D₂) for nonlinear HRV analysis. A modified GP correlation algorithm and the selection of parameters are presented. This nonlinear dynamical index is then employed for HRV characterization of the healthy subjects and hypertension patients. Results show that D₂ in daytime of a healthy person is generally smaller than that at night, whereas the hypertension patients exhibit no specific trend between the D₂ in daytime and that at night.