Relation between neonatal behavioral states and heart rate variability

Abstract At the end of pregnancy, fetal behavioral states can be determined by examining the fetal heart rate patterns alone. How-ever, drawing conclusions on fetal heart rate variability (HRV) in different states which have themselves been deter-mined on the basis of heart rate patterns may be self-predictive. In order to better understand the relationship between HRV and behavioral states, we examined the HRV in neonates in whom the behavioral states could be determined not using heart rate patterns but on the basis of Prechtl's criteria. Five newborns were observed and their behavior proto-colled in a quiet environment for 44-187 min. (median: 122 min.) within 2-7 (3) days after birth at term. Electrocardio-gram (ECG), electroencephalogram (EEG) and video recordings were simultaneously obtained. Behavioral state epi-sodes of 3 min duration were determined according to standard criteria for quiet sleep (S1), active sleep (S2), quiet awake (S3) and active awake (S4). For each episode, RR intervals, their standard deviation (SDNN) and root mean square of successive differences (RMSSD) were calculated. A total of 129 episodes were identified (N in S1 / S2 / S3 / S4: 44 / 36 / 21 / 28). The HRV measures values differed with respect to behavioral state (p<0.001): RR interval was longer in the quiet states, SDNN tended to increase over S1 to S4 while RMSSD tended to decrease. Linear discriminant analysis based on the HRV measures classified 67% of the episodes in the correct state (S1 / S2 / S3 / S4: 84% / 72% / 91% / 18%). All states could be well separated except S4 from S2 and S3. The results suggest that time domain HRV measures may help identify behavioral states except for S4 in neonates. This should be considered when determining fe-tal behavioral states in at-term fetuses on the basis of their heart rate patterns.