Heart rate variability was studied in normal and abnormal newborn infants during sleep in relation to the EEG background. The background EEGs of the full-term newborn infants were classified into normal, minimal depression, mild depression, moderate depression, burst-suppression and flat patterns in order of increasing severity. Nonsequential and sequential histograms were produced with the aid of a computer from some 500 consecutive RR intervals of the ECG recorded on a magnetic tape during polygraphic recordings. The seuential curves mainly showed fast oscillation during quiet sleep and slow oscillation during active sleep in normal infants. The infants who had moderately depresed EEGs presented pronounced oscillation. In those whose EEGs displayed burst-suppression and flat patterns, the sequential curves showed very little heart rate variability. They were often flat (fixed heart rate). Thus, heart rate variability was found to be one of the valuable parameters in evaluating abnormality of the CNS in newborn infants.