Sleep Deprivation in Mice: Looking Beyond the Slow Wave Rebound

Sleep is a fundamental process supporting the dynamic regulation of neural function. Emerging methods have proposed that the aperiodic components of brain signals (such as the spectral slope, spectral intercept, and spectral knee), in addition to entropy-based measures, offer robust empirical markers of neural states. The present study investigates the sensitivity of these broadband spectral metrics in comparison to classical band-limited measures, specifically slow wave activity (SWA; 0.75-4.5 Hz), in a 9-day mouse sleep deprivation paradigm involving baseline, sleep restriction, and recovery phases (open-source database). Spectral parameters were computed using the FOOOF algorithm. Results indicate that SWA differentiates between baseline and rebound sleep only during NREM episodes. In contrast, both the spectral slope and spectral intercept capture sleep deprivation-related changes during both REM and NREM sleep, suggesting these fractal measures reflect sleep homeostasis across stages. Given the shift of the spectral knee towards higher frequencies in mice (~8-10 Hz) as compared to humans (generally around 1 Hz), eliminating the overlap of the spectral slope with the traditional SWA range in these rodents, homeostatic regulation appears to be not strictly bounded to the lower frequencies (0.75-4.5 Hz). Normalised spectral entropy did not differentiate between baseline and recovery sleep, potentially due to its sample size sensitivity. These findings support the empirical utility of broadband spectral parameters in assessing sleep-wake dynamics and highlight their potential to complement or surpass traditional band-limited metrics.