Theta and alpha EEG oscillations reflect sleep need -- except during the wake maintenance zone

Abstract

Increasing time spent awake results in accumulated sleep need, a process known as sleep homeostasis. Sleep homeostasis combines with a 24 h circadian rhythm to determine when and for how long we sleep. Both sleep homeostasis and the circadian rhythm substantially affect spectral power of the wake electroencephalogram (EEG), but not in ways predicted by current models. Specifically, these models hypothesize that time spent awake increases neuronal synaptic strength, which increases synchronization and should therefore increase oscillatory activity. However, the dominant wake EEG oscillations, measured as theta (4-8 Hz) and alpha power (8-12 Hz), do not follow the predicted buildup in homeostatic sleep pressure with time awake. This is due to a limitation of spectral power analysis, which does not distinguish between changes in the amplitude of oscillations from changes in the quantity of oscillations present in the signal. We wished to determine whether the amplitudes of EEG oscillations would specifically reflect homeostatic sleep pressure, independently from changes in quantity. We collected data from 18 young healthy adults during a 4-h sleep / 24-h extended wake paradigm. We indeed found that theta and alpha oscillation amplitudes reflect homeostatic sleep pressure, increasing along a saturating exponential function with time awake. Instead, theta quantities increased linearly with time awake, and alpha quantities decreased. Notably, theta and alpha amplitudes temporarily decreased during the wake maintenance zone (WMZ), a 3-4 h time window just before bedtime when it is difficult to fall asleep. Using pupillometry, we also found that mean pupil diameter increased during this window, while variance decreased. These results suggest that the WMZ is dependent on an alerting signal from the ascending arousal system. The WMZ therefore counteracts the observed build-up in homeostatic sleep pressure reflected in EEG amplitudes by temporarily desynchronizing cortical activity.