Power spectral analysis of two-channel EEG in very premature infants undergoing heat loss prevention

Abstract

To evaluate whether wearing a wool cap, a routine practice used to prevent heat loss in premature infants, affects interpretation of electroencephalogram spectral analysis. Eighteen premature infants (median gestational age 28 weeks, range 23-32) without neurological complications were randomized to two channel (C3, C4 referred to Cz) digital electroencephalogram recordings with (90 min) and without (90 min) wearing wool cap, at 4 days of life. Electroencephalogram was analyzed automatically by measurement of burst suppression ratio and asymmetry index and by Fast Fourier Transform to calculate total absolute spectral power; relative spectral power in the δ (0.5-3.5 Hz), θ (4-7.5 Hz), α (8-12.5 Hz), and β (13-30 Hz) frequency bands; spectral edge frequency; and mean dominant frequency. The use of wool cap had no effect on all electroencephalogram parameters considered. Gestational age showed an effect on relative spectral power of all considered bands, spectral edge frequency and mean dominant frequency, while no effect was seen on burst suppression ratio and asymmetry index. Neonates born at gestational weeks lower than 28 had significantly higher relative power in the δ band and lower relative power in the α and β bands. Heat loss prevention using wool cap does not affect interpretation of spectral electroencephalogram. Spectral values in our group of very premature infants without neurological complications correspond to normal data reported in the literature. Maturation changes consist of reduction of relative power of the δ band, spectral edge frequency and mean dominant frequency.