The heart rate responds dynamically to various intrinsic and environmental stimuli. The autonomic nervous system is said to play a major role in this response. Multifractal analysis offers a novel method to assess the response of cardiac interbeat intervals. Twenty-four hour ECG recordings of RR interbeat intervals (of 48 elderly volunteers (age 65-94), 40 middle-aged persons (age 45-53) and 36 young adults (age 18-26)) were investigated to study the effect of aging on autonomic regulation during normal activity in healthy adults. Heart RR-interval variability in the very low frequency (VLF) band (32-420 RR intervals) was evaluated by multifractal tools. The nocturnal and diurnal signals of 6 h duration were studied separately. For each signal, the analysis was performed twice: for a given signal and for the integrated signal. A multifractal spectrum was quantified by the h(max) value at which a multifractal spectrum attained its maximum, width of a spectrum, Hurst exponent, extreme events h(left) and distance between the maxima of a signal and its integrated counterpart. The following seven characteristics are suggested as quantifying the age-related decrease in the autonomic function ('int' refers to the integrated signal): (a) h(sleep)(max) - h(max)(wake) > 0.05 for a signal; (b) h(int)(max) > 1.15 for wake; (c) h(int)(max) - h(max) > 0.85 for sleep; (d) Hurst(wake) - Hurst(sleep) < 0.01; (e) width(wake) > 0.07; (f) width(int) < 0.30 for sleep; (g) h(int)(left) > 0.75. Eighty-one percent of elderly people had at least four of these properties, and ninety-two percent of young people had three or less. This shows that the multifractal approach offers a concise and reliable index of healthy aging for each individual. Additionally, the applied method yielded insights into dynamical changes in the autonomic regulation due to the circadian cycle and aging. Our observations support the hypothesis that imbalance in the autonomic control due to healthy aging could be related to changes emerging from the vagal function (Struzik et al 2006 IEEE Trans. Biomed. Eng. 53 89-94).