Evaluation of maturity of sleep states in preterm infants using conventional and amplitude-integrated electroencephalography

Abstract

ObjectiveTo determine whether sleep state maturity can be estimated accurately using conventional electroencephalography (cEEG) or amplitude-integrated electroencephalography (aEEG) features concerning sleep in neurologically unimpaired preterm infants. MethodsA total of 51 preterm infants were monitored with cEEG-polygraphy and simultaneous aEEG. Sleep state maturity of EEG corresponded to specific postmenstrual age (PMA). PMA on cEEG was blindly estimated according to cEEG patterns (indicated as background continuity, frequencies, and voltages) as well as developmental markers in specific states. PMA on aEEG was blindly estimated based on the cycling score (cycling representing sleep state transitions) according to a pre-established scoring system. ResultsA total of 51 EEGs recorded between 32 and 37 weeks PMA were analysed. A significant relationship between estimated PMA (ePMA) and actual chronological PMA (cPMA) was shown by linear regression both on cEEG (r = 0.93, β = 0.98, 95% confidence interval (CI) 0.87–1.09, p < 0.001) and aEEG (r = 0.85, β = 0.83, 95% CI 0.69–0.98, p < 0.001). The estimation gap (defined as ePMA minus cPMA) was between −2 and +2 weeks both on cEEG and aEEG. The percentage of estimation gap between −1 and +1 weeks was 96% for cEEG, which was higher than the estimate of 88% for aEEG. ConclusionEstimated maturity of sleep state was well correlated with cPMA both on cEEG and aEEG. PMA corresponding to state maturity could be estimated within two weeks of actual cPMA using either of these two tools. However, cEEG had higher accuracy compared with aEEG in the evaluation of sleep state maturity.