Analysis of long term heart rate variability: Methods, 1/f scaling and implications

The use of spectral analysis techniques to quantify short-term heart rate fluctuations on the order of seconds to minutes has helped define the autonomic contributions to beat-to-beat control of heart rate. The authors used similar techniques to quantify the entire spectrum (0.00003-1.0 Hz) of heart rate variability during 24-h ambulatory ECG monitoring. The ECG from standard Holter monitor recordings from normal subjects was sampled with the use of a phase-locked loop, and a heart rate time series was constructed at 3 Hz. Frequency analysis of the heart rate signal was performed after a nonlinear filtering algorithm was used to eliminate artifacts. A power spectrum of the entire 24-h record revealed power that was inversely proportional to frequency, 1/f, over four decades from 0.00003 to 0.1 Hz (period approximately 10 h to 10 s). Displaying consecutive spectra calculated at 5-min intervals revealed marked variability in the peaks at all frequencies throughout the 24 h, probably accounting for the lack of distinct peaks in the spectra of the entire records.