Scale-specific and scale-independent measures of heart rate variability asrisk indicators

We study the correlation properties of heartbeat fluctuations using scale-s pecific variance (root-mean-square fluctuation) and scaling (correlation) e xponents as measures of healthy and cardiac impaired individuals. Our resul ts show that the variance and the scaling exponent are uncorrelated. We fin d that the variance measure at certain scales is well suited to separate he althy subjects from heart patients. However, for mortality prediction the s caling exponents outperform the variance measure. Our study is based on a d atabase containing recordings from 428 individuals after myocardial infarct (MI) and on database containing 105 healthy subjects nd 11 heart patients. The results have been obtained by applying two recently developed methods ( DFA-Detrended Fluctuation Analysis and WAV-Multiresolution Wavelet Analys is) which are shown to be highly correlated.