Unfolding the overnight dynamics in human sleep features plays a pivotal role in understanding
sleep regulation. Studies revealed the complex reorganization of the frequency composition
of sleep electroencephalogram (EEG) during the course of sleep, however the scale-free
and the oscillatory measures remained undistinguished and improperly characterized
before. By focusing on the first four non-rapid eye movement (NREM) periods of night
sleep records of 251 healthy human subjects (4–69 years), here we reveal the flattening
of spectral slopes and decrease in several measures of the spectral intercepts during
consecutive sleep cycles. Slopes and intercepts are significant predictors of slow
wave activity (SWA), the gold standard measure of sleep intensity. The overnight increase
in spectral peak sizes (amplitudes relative to scale-free spectra) in the broad sigma
range is paralleled by a U-shaped time course of peak frequencies in frontopolar regions.
Although, the set of spectral indices analyzed herein reproduce known age- and sex-effects,
the interindividual variability in spectral slope steepness is lower as compared to
the variability in SWA. Findings indicate that distinct scale-free and oscillatory
measures of sleep EEG could provide composite measures of sleep dynamics with low
redundancy, potentially affording new insights into sleep regulatory processes in
future studies.
