In adults and older children, heart rate variability (HRV) is frequently used to study autonomic function noninvasively. Normal values of HRV in newborn infants, however, are not widely available. This problem may be partially attributed to the lack of standardization of different methods. This study assessed HRV in normal newborn infants using 24-hour Holter monitoring. From 1997 to 2000, we prospectively evaluated frequency- (spectral analysis), geometric-, and time-domain indexes of HRV in normal term infants. Ninety-six asymptomatic infants who were <72 hours old were studied. Frequency-domain parameters (power in the high, low, very low, ultra low, and total frequency domains), a geometric parameter (HRV triangular index), and time-domain parameters (SDNN, SDANN, SDNNi, r-MSSD, s-NN50) are reported as means +/- SD, medians, and 5th and 95th percentiles to establish the normative values for newborns. A high degree of correlation (r > or = 0.85, p <0.0001) was noted among the 3 vagal tone dependent parameters, such as high-frequency power (frequency domain), r-MSSD, and s-NN50 (time domain). Our study supports the use of vagal dependent time-domain parameters like r-MSSD and sNN50 as surrogates for high-frequency power in newborns. Because the data are reported as means +/- SD, medians, and 5th and 95th percentiles, their use facilitates the study of parasympathetic and sympathetic activity in comparable populations.