Overnight dynamics in scale-free and oscillatory spectral parameters of NREM sleep EEG: can polysomnography provide an insight into sleep regulation?

Sleep is an inherent feature of human neurobehavioral functioning. However, clinical and research practices largely rely on consensual criteria of sleep-wake characterization. In turn, quantitative EEG measures were shown to be largely redundant and non-standardizable. We aim to transcend this issue by relying on the parametrization of NREM sleep EEG based on the power law scaling of the Fourier spectra (scale-free and oscillatory measures). We hypothesize that spectral slopes are characterized by sufficient convergent and criterion validity in order to be considered as alternative and more standardizable measures of sleep intensity, whereas spectral peak frequencies reveal a U-shaped overnight distribution indicating putative circadian influences. Findings indicate that spectral slopes flatten over successive sleep cycles, reproduce known age and region-effects, correlating reliably with slow wave activity (SWA). Furthermore, its interindividual variability was proven to be less than of SWA. Frequency of the largest peak was shown to roughly correspond to spindle frequency as the known age, sex, and regional effects were evident. Besides, the peak frequencies in the frontopolar region were compatible with an assumed circadian modulation. In conclusion, the appropriate parameterization of the NREM sleep EEG Fourier spectra reliably reflects processes proposed in the two-process model of sleep regulation.