Slow-wave sleep · Lobeworks/17

Slow-wave sleep is the deepest stage of non-REM sleep, the one in which the cortex fires in large synchronous waves of about one to two per second, and it is the stage most closely tied to the stabilisation of memories and, by one hypothesis, to the nightly reset of synaptic strength.

Slow-wave sleep. Slow-wave sleep is the deepest stage of non-REM sleep, the one in which the cortex fires in large synchronous waves of about one to two per second, and it is the stage most closely tied to the stabilisation of memories and, by one hypothesis, to the nightly reset of synaptic strength.

In the scalp record it is scored as stage N3 when waves of 0.5 to 2 Hz larger than 75 microvolts fill at least a fifth of a 30-second epoch (EEG). Behind each wave the cortical neurons alternate together between a silent down state and an active up state (the slow oscillation). The thalamus adds bursts of 11 to 16 Hz, the sleep spindles, and the hippocampus its brief sharp-wave ripples. Adults spend about 15 to 25 % of the night here, most of it in the first two cycles; it shrinks with age.

It hosts the replay. Ripples tend to arrive inside spindles, which ride the up states of the slow oscillation, a nesting proposed to let the hippocampus hand the day's episodes to the cortex (memory consolidation).

The synaptic homeostasis hypothesis gives it a second job. Waking learning strengthens many synapses, which costs energy and space and would saturate; Tononi and Cirelli proposed that slow-wave sleep scales them down across the board, keeping the strongest. Electron microscopy in mouse cortex found synapses about 18 % smaller after sleep, sparing the largest. It is a hypothesis, argued against by results where sleep strengthens selected synapses.

Theta belongs elsewhere. The 4 to 8 Hz theta rhythm marks active exploration and REM sleep, not this stage; delta is the band of slow-wave sleep.

Whether it also washes the brain is disputed (cerebrospinal fluid).

In slow-wave sleep the cortex stops listening and starts filing.

Cut off from the senses, it falls into one shared rhythm, and inside that rhythm the day's traces are replayed, kept or pared down.

Questions: Does sleep wash the brain? Maybe, and the evidence points both ways. The glymphatic hypothesis holds that cerebrospinal fluid flows into the tissue along arteries and out along veins, carrying away waste such as amyloid, and a 2013 mouse study measured the space between cells growing by about 60 % in sleep with clearance rising to match. A 2024 mouse study using a different method found clearance reduced in sleep and anaesthesia instead. The two disagree on measurement more than on anatomy, and neither has been shown directly in the human brain, so sleep cleans the brain is a claim to hold loosely. What does deep sleep do for yesterday's memories? It replays them. During slow-wave sleep the hippocampus fires the sequences it recorded during the day again, compressed into sharp-wave ripples of a fraction of a second, and those ripples tend to fall inside thalamic spindles that ride the up states of the cortex's slow waves. That nesting is thought to be the moment the cortex receives the episode and links its pieces directly, so the memory depends less on the hippocampus over time. In rats, cutting the ripples short during sleep impairs the memory the next day; in people, a stretch of sleep rich in deep sleep protects word pairs learned before it better than one rich in REM sleep (Plihal and Born, 1997). What is the synaptic homeostasis hypothesis? It is the proposal, by Giulio Tononi and Chiara Cirelli, that sleep exists partly to pay back the synaptic cost of being awake. Learning during the day strengthens many synapses, and strengthening is expensive in energy and space and would end in saturation, with every synapse strong and none informative. Slow-wave sleep, with the cortex offline and firing in synchronous waves, would scale them down across the board, a general weakening related to long-term depression, keeping the relative differences and sparing the strongest. Electron microscopy of mouse cortex found synapses about 18 % smaller after a few hours of sleep, the largest spared. It stays a hypothesis: other studies find sleep strengthening and forming selected synapses after learning, and both may happen at once.