PhD Scientific Days 2023

Mental Health Sciences IV.

ECoG Spectral Parameters of Homeostatic Sleep Regulation

Text of the abstract

In the frequency domain the surface neurophysiological (EEG, ECoG) signal can be decomposed into two components: an aperiodic background activity described by a 1/f function and periodic perturbations superimposed on the background activity. An index of the aperiodic activity is the spectral slope derived by curve fitting over the log-log transformed power spectrum, reflecting on the ratio of slow to fast oscillations. Recent evidence suggests the aperiodic power parameter is a promising biomarker of sleep homeostasis. The present study investigates the spectral slope computed using the fooof (fitting oscillations & one over f) algorithm on an animal model (Mus musculus, C57BL/6 and 129S4/SvJae hybrids). The mice were exposed to a 9 day-long continuous ECoG recording containing a sleep deprivation paradigm with baseline, sleep deprivation, and recovery phases. The spectral slope appears to adequately capture the homeostatic sleep regulatory adjustments related to prolonged wakefulness and the subsequent compensation. Furthermore, the spectral slope quantifies more accurately the neural changes caused by sleep deprivation relative to the classical band/bin-wise approaches (within the slow wave range in the current case), and also reflects on the high inter- individual variability of neurophysiological signals.