P102 Can EEG spectral analysis distinguish children with narcolepsy from those with idiopathic hypersomnolence and subjective sleepiness?

Abstract Background EEG spectral analysis provides a measurement of sleep micro-architecture and is a more sensitive measure of sleep disruption than conventional sleep macro-architecture. We aimed to examine the use of this technique in distinguishing children with narcolepsy and idiopathic hypersomnolence (IH) from subjectively sleepy children with a negative multiple sleep latency test. Methods Relative power, the absolute power in each frequency band divided by the total absolute power, was calculated for delta power (0.5 Hz–3.9 Hz), theta power (4 Hz–7.9 Hz), alpha power (8 Hz–11.9 Hz), sigma power (12 Hz–13.9 Hz) and beta power (14 Hz–30 Hz). A mean value for each frequency was calculated for each 30 s epoch then averaged for each sleep stage within each child. Results 28 children with Narcolepsy, 11 with IH and 26 with subjective sleepiness were included and individually matched for age and sex with a control child. In N2, Theta power was higher in the Narcolepsy group (median 13.5%, IQR 11.2-15.5) compared to the subjectively sleepy group (11.1%, 9.0-12.8, p<0.05). During N3, Delta power was higher in the subjectively sleepy group (93.3%, 91.9-93.9 vs 83.3%, 65.8-93.9, p<0.001), and Theta power (4.8%, 4.2-6.3 vs 12.6% 4.8-22.5, p<0.001), Alpha power (0.9%, 0.7-1.3 vs 2.8% 0.9-5.8, p<0.001), and Beta power were lower (0.3% 0.3 -0.4 vs 1.1% 0.3-4.2, p<0.05) compared to their controls. Conclusions Differences in sleep micro-architecture could not distinguish between children with narcolepsy and IH but may be able to distinguish children with narcolepsy from subjectively sleepy children.