In the present study, we reinvestigated the question of whether human heart rate variability (HRV) is fractal in nature. Ten healthy volunteers participated in either of two studies conducted while beat-by-beat long-term HRV (8,500 heartbeats) was recorded for 2-3 h in the quiet, awake state in the supine position. In the first study, five subjects were tested four times each to evaluate the basic fractal nature of human HRV. The other five subjects were examined for the effects of oral propranolol (2 x 80 mg/day) on the fractal property of HRV in the second study. HRV data were analyzed by coarse-graining spectral analysis to break down their total power into harmonic and nonharmonic (fractal) components. The harmonic component was further divided into low (0.0-0.15 Hz; LF)- and high (> 0.15 Hz; HF)-frequency components. From these spectral components, %Fractal, %LF, and %HF as functions of total power were calculated. The fractal component was used to calculate the spectral exponent, beta. The %Fractal of human resting HRV was 85.5 +/- 4.4% (mean +/- SD, n = 20). The beta for the fractal HRV was 1.08 +/- 0.18 (n = 20). With propranolol, these basic properties of fractal HRV dynamics remained unchanged despite an increase in the mean RR interval (placebo, 912 +/- 111 ms; propranolol, 1,134 +/- 133 ms, P < 0.05) and a change in the harmonic spectral shape evaluated by LF/HF (placebo, 2.76 +/- 1.57; propranolol, 1.82 +/- 0.81, P < 0.05). For short-term data, less power was extracted as fractal because of the absence of the very low frequency component, yet the beta and LF/HF were unchanged from long-term data. These findings indicate that 1) the observed inversely proportional frequency (1/f) spectrum in human resting HRV is due to underlying random fractal dynamics and 2) the sympathetic nervous system seemed to play a minor role in modulating the fractal HRV dynamics.