Title: NREM and REM: cognitive and energetic effects in thalamo-cortical sleeping and awake spiking model

Abstract: Sleep is known to play a central role in learning and cognition, yet the mechanisms underlying its role in stabilizing learning and improving energetic management are still to be clarified. It is characterized by patterns of cortical activity alternating between the stages of slow wave sleep (NREM) and rapid eye movement sleep (REM). In this work, we investigate the role of sleep-like cycles on a biologically-plausible two-area thalamo-cortical spiking neural network capable of experiencing AWAKE-NREM-REM brain states. We train the network on a handwritten digits classification task and show the beneficial effects of sleep on energy consumption and classification performance, exploiting its effects on internal structure, memory representation and consolidation. We show the ability of sleep to shape the synaptic distribution, thus affecting the neuronal firing rate and creating novel multi-area associations.