Sleep Is Plumbing

Sleep advice is usually delivered as a moral instruction — you should sleep more — with a vague gesture at "recovery". The mechanism is far more specific and far more interesting than that, and knowing it changes which parts of sleep hygiene you actually bother with.

#The brain has no lymph vessels

Every other tissue in your body clears metabolic waste through the lymphatic system. The brain doesn't have one. What it has instead is a bulk-flow washing mechanism, and it's mostly plumbing.

Cerebrospinal fluid is driven along the spaces surrounding arteries by arterial pulsation. It crosses into the brain's interstitium through aquaporin-4 water channels concentrated at the endfeet of astrocytes, washes through the tissue, and exits along perivenous routes.

The critical part: during slow-wave NREM sleep, noradrenergic tone falls and the interstitial space physically expands by roughly 60%. That drop in hydraulic resistance sharply increases the exchange of cerebrospinal and interstitial fluid. The brain isn't metaphorically cleaning itself. It is being flushed, and the valve is deep sleep.

#From mice to humans

For about a decade this was almost entirely rodent work, which meant every popular sleep article was extrapolating from mice and mostly not saying so.

That changed. A 2026 Nature Communications randomised crossover trial — sleep versus sleep-deprivation nights, using an in-ear wearable measuring brain parenchymal resistance by impedance spectroscopy alongside EEG — was the first to link human sleep-stage physiology directly to plasma clearance of amyloid-beta and tau. Sleep-linked glymphatic activity explained 49–98% of additional variance in biomarker clearance beyond baseline models. The strongest predictors were reduced parenchymal resistance, increased cerebrovascular compliance, and higher EEG delta power.

Delta power is the variable. That's the measurable output of slow-wave sleep, and it's what alcohol suppresses, what late meals disturb, and what fragmented sleep removes.

Worth one honest caveat: the trial inferred clearance from plasma biomarker appearance rather than imaging flow directly, and several groups have published serious challenges to the glymphatic model itself, arguing that the observed transport is better explained by diffusion plus vascular pulsation than by directed bulk flow. Sleep obviously matters for cognition. Whether glymphatic clearance specifically is the mechanism is less settled than the coverage suggests.

#The other thing deep sleep does

The largest growth-hormone pulse of the 24-hour cycle happens at the onset of slow-wave sleep, driven by increased hypothalamic GHRH neuron activity against reduced somatostatin tone. GH then drives hepatic IGF-1 production.

Fragmenting deep sleep doesn't move this pulse to later. It reduces it — markedly, in selective slow-wave suppression studies.

#Why the clock matters more than the duration

Here's the finding that should reorganise most people's sleep hygiene: regularity predicts mortality better than duration does. A consistent 7.5 hours beats an average 8.5 hours assembled from a ragged schedule.

The mechanism is entrainment. The master clock in the suprachiasmatic nucleus is set by intrinsically photosensitive retinal ganglion cells containing melanopsin, most sensitive around 480nm — blue. These project directly to the SCN and are what suppress pineal melatonin. Morning light advances the clock and sets that evening's melatonin onset roughly 14–16 hours later.

Which means the highest-leverage sleep intervention happens in the morning, not at bedtime. You aren't choosing when to sleep. You're choosing when the system will let you.

#What actually follows from this

• Fix the wake time, seven days a week. Bedtime standard deviation under 15 minutes is the target.
• Daylight in the eyes within 30 minutes of waking, outdoors. Through a window loses an order of magnitude of intensity.
• Cold, dark room. Sleep onset requires core temperature to fall, achieved by dumping heat through the hands and feet — a cold room accelerates the gradient.
• No alcohol. It suppresses precisely the slow-wave sleep that drives both the clearance and the GH pulse. This is the single largest avoidable input.
• Last meal about three hours before bed, and no caffeine after early afternoon — adenosine is still accumulating behind the receptor blockade, and caffeine's half-life is around five hours in most people.

One adjustment worth naming: if you're under 18, the consensus recommendation is 8–10 hours, not the adult 7–9. Adolescent sleep need is genuinely higher, and adolescent circadian phase is genuinely delayed — which argues for a later fixed wake time, not for abandoning the fixed part.

---

This is one lever out of a larger system. The full protocol — every mechanism, every dose, every source tier, and an append-only record of everything it has gotten wrong — is at /protocol.