Incoprea Theory
Volume III · 2026

Theory 21/Sleep & Chronobiology /Sleep physiology/3 min

Sleep Is Local Before It Is Global

Sleep is treated as a whole-brain switch. The tiredness may start in one patch of cortex.

The claim

Sleep pressure accumulates in the cortical regions that were used, so sleep is spatially inhomogeneous and the transition to sleep begins locally: performance lapses should be predicted by local slow-wave activity rather than by time awake.

01The evidence

  • Local sleep has been demonstrated in awake rats, with individual cortical regions showing off-periods characteristic of sleep while the animal is behaviourally awake.PMC, local sleep in awake rats
  • Local build-up of sleep pressure has been linked to mind wandering, which is the behavioural signature a spatially inhomogeneous process would produce.Physiology and Behavior

02The tension

Sleep is modelled as a global state with a global drive, measured by time awake and by EEG averaged over the scalp. That predicts that performance should degrade uniformly and that sleep should begin everywhere at once. Neither is what the recordings show.

03The theory

Wakefulness imposes a metabolic and synaptic load on the neurons that are active. The load is local, so the recovery need is local. Slow waves during subsequent sleep are largest in the regions that were most active, which is the mechanism by which the debt is repaid.

If the debt is local, then the transition to sleep is not a switch but a percolation of local off-periods. Behaviour fails when enough regions are off at once, which is why lapses are intermittent and why they cluster in the tasks that use the most-depleted regions.

The prediction is that lapses should be predicted by local slow-wave activity in the task-relevant region rather than by time awake, and that a task using one region should impair performance on that task before others.

The idea also explains why sleep deprivation effects are task-specific: the person is not uniformly impaired, they are impaired where the local debt is largest.

04What this is not

Global sleep-homeostasis models

Those models treat sleep pressure as a whole-brain variable indexed by time awake and scalp-averaged EEG. This treats it as a spatial field, and predicts task-specific impairment that a global drive cannot produce.

05Predictions

  • P1Performance lapses should be predicted by local slow-wave activity in the task-relevant region better than by time awake.
  • P2Deprivation should impair the most-used task before others, at matched effort.
  • P3Recovery sleep should be localised to the regions that were most active.

06What would refute it

Uniform impairment across tasks with no regional signature would leave the global account standing.

07What it would change

It would make fatigue management task-specific, and change how shift-work schedules and safety limits are designed.

08Evidence base

  1. Local sleep in awake rats — PMChttps://pmc.ncbi.nlm.nih.gov/articles/PMC3085007/
  2. Local build-up of sleep pressure and mind wandering — Physiology and Behaviorhttps://www.sciencedirect.com/science/article/pii/S0006295221000745