Determination of Vigilance States and Spectral Analysis

Vigilance states for consecutive 4-s epochs were classified by visual inspection, and blind to genotype, according to standard criteria: wakefulness (high and variable EMG signal, low-amplitude EEG signal), NREM sleep (NREMS; high EEG amplitude dominated by slow waves, low EMG), and REM sleep (REMS; low EEG amplitude, theta oscillations and muscle atonia). Vigilance states were analyzed offline using Neuroscore Software (Data Sciences International) with the EEG and EMG signals modulated with a high-pass (3 dB, 0.5 Hz) and a low-pass (50 Hz) analog filter. For LD or DD conditions, continuous recordings were analyzed and time spent in each vigilance state was expressed as a percentage of the total recording time over various intervals (1–24 h). All DD recordings were started after at least 7 days of constant conditions. The mean number of individual bouts of vigilance states were grouped as a function of their duration (8–12, 16–28, 32–60, 64–124, 128–252, 256–508, 512–1020, >1024 s) per hour during LD or DD. The mean duration of NREMS bouts was compared between ZT6-12 on baseline day and following 6 h of sleep deprivation (SD).

EEG power spectra were computed for consecutive 4 s epochs over a circadian cycle by a fast-Fourier transform (frequency range: 0.5–49.80 Hz; resolution 0.24 Hz; Hanning window function). Genotypic differences were determined in DD over a complete circadian cycle and expressed as a percentage of total EEG power within each vigilance state for each mouse. The time course of spectral activity was also computed in 2 h bins during LD for delta (1–4 Hz) during NREMS and during/post 6 h SD and calculated as a percentage of the mean 24 h baseline for each mouse. Epochs containing EEG artifacts were discarded from the analysis.