Objective characterization and differentiation of sleep states in healthy newborns and newborns-at-risk by spectral analysis of heart rate and respiration rhythms.

Best parameters of power and coherence spectra of heart rate fluctuations (HR) and respiration rhythms (RR) were established to differentiate sleep states in healthy newborns and newborns-at-risk by means of multivariate variance and discriminant analysis. Nine healthy newborns and 20 newborns with low risk features were examined polygraphically. Long-term-variability of both HR and RR and coherence between HR and RR in the frequency range of 0.26-0.97 Hz, corresponding to Respiratory Sinus Arrhythmia, allow the best differentiation of neonatal sleep states S 1 and S2. Differentiation of healthy newborns and newborns-at-risk by these parameters was not possible. Thus, in studies dealing with low risk features in newborns sleep states must be previously classified. State 1 and 2 represent different autonomic organisations. State 1 is a neuro-vegetative, relatively stable state. State 2 shows cyclic increases of coherence between HR and RR within the frequency range of Respiratory Sinus Arrhythmia. These properties are related to autonomic brain stem functions and were absent in 6 out of the 20 newborns-at-risk.