OFF-periods reduce the complexity of neocortical activity during sleep

Abstract

Complexity of electroencephalographic signals decreases during slow-wave sleep (SWS), however, the neural mechanism for this decrease remains elusive. Here, we show that this complexity reduction is caused by synchronous neuronal OFF-periods. We analyse in-vivo recordings from neocortical neuronal populations, finding that OFF-periods in SWS trap cortical dynamics, making the population activity more recurrent, deterministic, and less chaotic than during REM sleep or Wakefulness. Importantly, when we exclude OFF-periods, SWS becomes indistinguishable from Wakefulness or REM sleep. In fact, we show that spiking activity for all states has a universal scaling compatible with critical phenomena. We complement these results by a critical branching model that replicates our experimental findings, where we show that forcing OFF-periods to a percentage of neurons suffices to generate a decrease in complexity that replicates SWS.