Forecasting of ventricular tachycardia using scaling characteristics and entropy of heart rate time series

Abstract

In this paper we investigated heart rate time series stored by implantable cardioverter defibrillators in order to the short-term forecast ventricular tachycardia (VT). Standard heart rate variability (HRV) parameters, compression based entropy functions, and HRV scaling characteristics (detrended fluctuation analysis, Higuchi’s fractal dimension) were analyzed in 29 VT time series and compared to individually acquired control time series. We found no differences in standard HRV parameters but significant changes in the entropy function as well as in the scaling characteristics of HRV before the onset of VT. In conclusion, HRV analysis might provide markers for early detection of forthcoming VT.