Sleep.

While you sleep, the brain consolidates memory, clears metabolic waste (the glymphatic system), and rebalances hormones, metabolism, immunity, and emotion. Chronic short or broken sleep raises the risk of heart disease, high blood pressure, type-2 diabetes, dementia, depression, and early death. No supplement stack outperforms simply sleeping well — which is why sleep, not the latest peptide, is the real foundation of health. Evidence ●●●●●

The single most important truth in this whole guide: you cannot biohack your way out of too little sleep. Gadgets and supplements optimise the margins — they don't replace hours in bed. The genuine short-sleeper gene (DEC2/ADRB1) is real but vanishingly rare; 'I only need five hours' is almost always accumulated debt you've stopped noticing, because the sense of impairment fades long before the impairment does. Evidence ●●●●○

Three things, in order: enough (duration), regular (same times every day), and undisturbed (deep, unfragmented). Nailing those three matters far more than optimising any exotic metric — a perfect magnesium stack can't rescue a chaotic schedule or an untreated snore. Get the fundamentals first; then tune the margins with everything that follows.

Aim for 7–9 hours of actual sleep; both too little and habitually too much track with worse health (a U-shaped curve). But timing regularity may matter as much as duration — going to bed and waking at consistent times predicts lower mortality even after accounting for how long you sleep (the Sleep Regularity Index work). Consistency is the most underrated lever in sleep, and the cheapest. Evidence ●●●●○

You cycle through light (N1/N2), deep (N3, slow-wave), and REM roughly every 90 minutes. Deep sleep — physical restoration and glymphatic clearance — is front-loaded early in the night; REM — emotional processing and memory — is back-loaded toward morning. That's why cutting sleep short mostly steals REM, and going to bed too late steals deep. Rough norms: ~13–23% deep, ~20–25% REM. Evidence ●●●●○

Sleep efficiency (time asleep ÷ time in bed; aim >85%), sleep-onset latency (aim 10–20 min — falling asleep the instant your head hits the pillow signals debt, not virtue), and WASO (wake after sleep onset). Plus the recovery signals: HRV and resting heart rate (higher overnight HRV and lower RHR mean better recovery), respiratory rate, overnight SpO2, and body temperature. Read trends over weeks, never single nights. Evidence ●●●○○ as consumer metrics

Sleep debt is real and it accumulates. You can partially repay it with recovery sleep, but weekend catch-up doesn't fully undo weekday restriction, and the performance hit keeps creeping up while your sense of how impaired you are flatlines (the classic chronic sleep-restriction studies). The lesson: it's far cheaper to avoid building the debt than to repay it. Evidence ●●●●○

The Oura ring is the enthusiast favourite: strong for sleep/wake, duration, resting heart rate, HRV, respiratory rate, and temperature trends — weaker at exact stage-by-stage scoring (as are all of them). Whoop centres on recovery, strain, and HRV; the Apple Watch now flags possible sleep apnea from breathing disturbances (an FDA-cleared feature); Garmin and Fitbit track well too. All are good at duration and trends, unreliable at precise deep/REM splits. Evidence ●●●○○

The Eight Sleep Pod actively heats and cools each side of the bed through the night and tracks you contactlessly; under-mattress sensors (Withings and others) track with nothing on your body. Their real value is the temperature control and no-wearable convenience more than pinpoint staging — the thermal regulation is a genuine lever, the sleep-stage graph is a rough estimate. Evidence ●●●○○

Validated against clinical sleep studies, consumer trackers are decent at total sleep time and sleep/wake but tend to overestimate sleep and get the stages wrong; a single night's numbers are noisy. Worse, fixating on the score causes 'orthosomnia' — tracker-induced anxiety that actively worsens sleep. Watch weekly trends, not last night's verdict. Evidence ●●●●○ on the accuracy limits

Your core temperature has to drop about 1°C to fall and stay asleep, so a cool room (~18°C / 65°F, roughly 16–20°C) helps — too warm is one of the most common hidden sleep-wreckers. Warm feet paradoxically help: dilating the hands and feet dumps core heat, which is why socks or a warm bath before bed speed sleep onset. Cooling mattress pads take this further. Evidence ●●●●○

Even modest light suppresses melatonin, and a controlled study found sleeping in a dimly lit room raised overnight heart rate and worsened next-morning insulin resistance versus full darkness. Blackout curtains, an eye mask (shown to improve next-day alertness and memory), and covering stray standby LEDs are cheap, high-yield fixes. Evidence ●●●●○

Unpredictable noise causes micro-arousals you don't remember but that fragment sleep and spike blood pressure. Earplugs, a solid door, and a steady background sound to mask sudden noises all help. (The sound question — whether white or pink noise actually improves sleep — gets its own section later.) Evidence ●●●○○

Get bright light into your eyes soon after waking, ideally outdoors (even an overcast sky is many times brighter than indoor lighting) for 10–30 minutes. Morning light anchors and strengthens your circadian rhythm, making you alert now and sleepy at the right time tonight. It's the single most powerful — and completely free — circadian tool you have. Evidence ●●●●○

In the last 2–3 hours before bed, dim the lights and cut bright, blue-rich screen light, which suppresses melatonin and pushes your clock later. Use warm, dim lamps, night modes, and low brightness. Blue-blocking glasses may help a little, but the evidence is genuinely mixed — don't over-rely on them; simply dimming everything does more. Evidence ●●●○○

Even light through closed eyelids matters: a controlled study found a single night in a moderately lit room raised heart rate and worsened next-morning glucose handling versus sleeping in the dark. Kill the standby LEDs, block the window, and mask your eyes if the room can't be made fully dark. Evidence ●●●○○

Caffeine has a half-life of about 5–6 hours, so an afternoon coffee is still half-active at bedtime; a controlled study found caffeine taken even 6 hours before bed measurably cut sleep. Practical rule: none within ~8–10 hours of sleep (for many, a cut-off around early afternoon). It works by blocking adenosine — the 'sleep pressure' that should build all day — so late caffeine steals from the very drive that puts you under. Evidence ●●●●○

Alcohol helps you fall asleep and then wrecks the second half of the night: it fragments sleep, suppresses REM, and worsens snoring and apnea. A nightcap is one of the most reliable ways to feel unrested despite technically 'sleeping.' If you drink, leave several hours before bed and keep it moderate. Evidence ●●●●○

A large meal close to bedtime raises core temperature, spikes glucose, and can cause reflux — all working against sleep. Aim to finish eating about 3 hours before bed (though don't go to bed starving either — a small, boring snack is fine). Nicotine, like caffeine, is a stimulant best kept out of the evening. Evidence ●●●○○

A hot bath or shower 1–2 hours before bed is one of the best-evidenced routines: warming your skin drives blood to the surface so your core temperature drops faster afterward, and a meta-analysis found it cuts the time to fall asleep by about a third. A sauna does the same. Counter-intuitive but real — get warm to get sleepy. Evidence ●●●●○

Regular exercise is one of the strongest sleep improvers, and the old 'never exercise at night' rule is largely debunked: evening exercise is fine for most people, with one exception — vigorous exercise ending within about an hour of bed can delay sleep. Morning or afternoon training has the bonus of reinforcing your circadian rhythm through activity and light. Evidence ●●●●○

Give yourself 30–60 minutes of low-stimulation wind-down: dim light, no work, no doomscrolling. What sabotages the landing: screens in bed, late heavy meals, a nightcap, late vigorous exercise, intense work or arguments, and clock-watching. Keep the bed for sleep (and sex) only, so your brain keeps linking the two. Evidence ●●●○○

The big secret of bad sleep is that you can be pulled to the edge of waking dozens — even hundreds — of times a night without remembering any of it. Each micro-arousal fragments the deep and REM sleep that actually restore you, and briefly spikes heart rate and blood pressure. You register '8 hours in bed'; your brain got far less continuous sleep than that. Evidence ●●●●○

The common causes, roughly in order of how often they're the real problem: disordered breathing (snoring, apnea, airway resistance — the big one, covered next), alcohol, an irregular schedule or jet lag, a too-warm or too-bright room, stress and a racing mind, a full bladder (nocturia), restless legs, and simple aging (deep sleep naturally declines). Most 'bad sleepers' have one or two of these, not a mystery. Evidence ●●●○○

Fragmented sleep doesn't just make you tired; the repeated arousals and sympathetic surges raise blood pressure, strain the heart, worsen glucose control, and blunt recovery. That's why untreated fragmentation — especially from breathing — is a genuine health issue, not merely a comfort one. It's also why 'fixing' it can change how you feel far more than any supplement. Evidence ●●●●○

Snoring is turbulent airflow vibrating the soft tissues of a narrowed airway. Benign primary snoring — no pauses, no daytime sleepiness — is mainly a partner problem. But loud snoring plus witnessed pauses, gasping, morning headaches, nocturia, or daytime sleepiness is a red flag for apnea and warrants a proper test. Snoring that worsens sharply with alcohol or lying on your back points the same way. Evidence ●●●●○

External nasal strips (Breathe Right) and internal nasal dilators flare the nostrils and genuinely reduce snoring and improve nasal breathing for congested breathers — but they do not treat sleep apnea (meta-analysis shows no meaningful effect on the apnea index or oxygen levels). Useful comfort aids for a stuffy nose; not a treatment for a collapsing airway. Evidence ●●●○○

Taping the lips to force nasal breathing is a viral trend built on thin, mixed evidence (one small mild-apnea study was positive; a 2024 review found limited benefit and possible harm). The safety point matters: mouth taping in undiagnosed apnea can be dangerous — if your airway obstructs, an open mouth is an escape valve for air, and taping it shut removes it. Get tested before you tape. Evidence ●●○○○

The highest-yield, mostly-free fixes address the airway itself: don't sleep on your back (positional therapy roughly halves supine events), avoid evening alcohol (it relaxes the airway muscles), lose excess weight, and treat nasal congestion (allergies, a deviated septum). These change what's collapsing, not just the sound it makes. Evidence ●●●●○

In OSA the airway repeatedly collapses during sleep, each event ending in a gasp or arousal. Severity is graded by the apnea-hypopnea index (AHI): 5–15 mild, 15–30 moderate, over 30 severe. It affects an estimated ~1 billion adults — the majority undiagnosed — and untreated it independently drives high blood pressure (especially resistant), atrial fibrillation, stroke, insulin resistance, and higher mortality. Evidence ●●●●●

You can have significant airway trouble that a standard apnea count calls 'normal.' Upper airway resistance syndrome (UARS) and effort-related arousals fragment sleep without meeting the apnea threshold — classically in younger, thinner people, often women, the opposite of the stereotype. If you have every symptom of disordered breathing but a 'normal' home test, this is the leading reason — and it takes a full lab study to catch. Evidence ●●●○○

Here's the part most people miss: you don't need a full apnea to be harmed. A transient dip in blood oxygen (or just airflow limitation) triggers a micro-arousal that fragments sleep, fires the fight-or-flight system, spikes blood pressure, and blunts HRV — potentially hundreds of times a night. The oxygen desaturation index (ODI) counts these dips, and even sub-apnea nocturnal desaturation raises cardiovascular risk. Not every desaturation is 'sleep apnea' — but every one taxes your sleep and your heart. Evidence ●●●●○

CPAP is the gold standard and highly effective. Alternatives include mandibular advancement (oral) appliances, positional therapy, myofunctional (airway-muscle) exercises, and ENT surgery. For the obesity-driven type, significant weight loss helps — and GLP-1 drugs (tirzepatide) are now FDA-approved for moderate-severe apnea after roughly halving events in trials. Hypoglossal-nerve stimulation (Inspire) is an option for the CPAP-intolerant. Evidence ●●●●●

Home sleep-apnea tests using peripheral arterial tonometry (PAT) — the WatchPAT One (single-use) and WatchPAT 300 — read finger arterial tone, oxygen, and movement to estimate your apnea index, and they correlate well with a full lab study (r ≈ 0.87–0.89). Ordered through a clinician, a PAT test is the strongest at-home option for moderate-to-severe apnea. Evidence ●●●●○

Cheaper, always-on screening: an overnight recording pulse oximeter (e.g. Wellue O2Ring) shows whether your oxygen is dipping and reports an ODI; Oura flags breathing disturbances; the Apple Watch has an FDA-cleared apnea notification (high specificity, modest sensitivity — a positive alert is meaningful, a negative doesn't rule it out). Excellent for surfacing a problem; none is a diagnosis. Evidence ●●●○○

A full attended overnight study records brain waves, eye movements, muscle tone, heart, airflow, breathing effort, oxygen, and position — the only test that scores true sleep time, arousals, effort-related arousals (RERAs), and central vs. obstructive events. It's the right test when you have strong symptoms but a normal home result, or when UARS, central apnea, or another sleep disorder is suspected. Evidence ●●●●●

Home tests are validated for moderate-to-severe apnea and tend to under-count events (they measure over time in bed, and can't score effort-related arousals). So they can miss mild disease and UARS. Per the guidelines, a 'normal' home test in someone with real symptoms should lead to a lab study — not to reassurance. Evidence ●●●●○

The best-supported, in rough order: glycine 3 g before bed (lowers core temperature, speeds onset); magnesium (glycinate, ~200–400 mg elemental — best if you're deficient); ashwagandha (KSM-66, 300–600 mg for 8+ weeks — the best-evidenced adaptogen, working by lowering stress and cortisol, ideal for the 'wired and tired'); and L-theanine 200 mg or oral lavender (Silexan) 80 mg for anxiety-driven insomnia. All genuine, all modest, all safe. Evidence ●●●○○

GABA (a pill probably can't reach the brain — the mechanism marketing is misleading), isolated apigenin (chamomile's active compound, but no human trial of the isolate), valerian and L-tryptophan (the sleep guidelines recommend against both), and taurine/magnolia (animal data only). Tart cherry is mildly real but comes with a lot of sugar in juice form. Evidence ●●○○○

DSIP (delta sleep-inducing peptide) and epitalon are marketed for sleep, but the human evidence is thin-to-absent — DSIP's few positive studies were tiny, decades old, given intravenously, and never replicated — and they're unregulated gray-market injectables with no approved use. Interesting names; no real evidence, and real purity/dosing risk. Evidence ○○○○○ in humans

Increasingly sold as a bedtime nasal spray, especially in perimenopause. A systematic review found “little or no sleep-promoting influence” in humans, and in animals it tends to promote wakefulness; the one real signal is shorter events in sleep apnea, from two small studies. Any benefit is indirect, through stress. Evidence ●○○○○ for sleep — see the Oxytocin guide.

Melatonin signals biological night; it doesn't knock you out. That's why it shines for circadian problems — jet lag, delayed sleep phase, shift work — and only modestly helps ordinary insomnia (the sleep guidelines actually recommend against it for chronic insomnia, because as a hypnotic it barely works). Match it to a clock problem, not a can't-switch-off problem. Evidence ●●●○○

The classic MIT work found a physiological 0.3 mg dose restored sleep as well as 3 mg — while the higher dose pushed melatonin to unnatural levels that lingered into the day, causing grogginess. The 3–10 mg gummies on shelves are 10–30× physiological. For most purposes, 0.3–0.5 mg, timed correctly, is the target — more is not better, it's arguably worse. Evidence ●●●○○

To shift your clock earlier, take a small dose in the early evening (a few hours before your natural sleep time), not at bedtime. And beware quality: US analyses found actual melatonin content ranging from about −83% to +478% of the label, with some gummies containing unlabelled CBD or serotonin. Buy USP-verified, and use tiny doses. Evidence ●●●●○ on the quality problem

Dual orexin receptor antagonists (daridorexant/Quviviq, lemborexant/Dayvigo, suvorexant/Belsomra) block the brain's 'stay awake' signal rather than forcing sedation, so they preserve natural sleep architecture better and carry the lowest dependence of the hypnotics. They're the emerging near-first-line drug when one is genuinely needed — the main barriers are cost and insurance coverage. Evidence ●●●●○

Z-drugs (zolpidem/Ambien, eszopiclone) are effective and fast but bring tolerance, dependence, and an FDA boxed warning for complex sleep behaviours (sleep-driving and -eating, even on a first dose). Benzodiazepines suppress deep and REM sleep and are habit-forming. Both are for short-term or situational use, not nightly. Low-dose doxepin (3–6 mg) and, off-label, trazodone are gentler maintenance options. Evidence ●●●○○

The over-the-counter 'PM' sleep aids (diphenhydramine/Benadryl, doxylamine) stop working within days as tolerance builds, and their anticholinergic load is linked, at chronic high use, to a higher dementia risk. Fine for the occasional single night; not for regular use, and best avoided in older adults. Evidence ●●●○○

Light shifts your clock, and the direction of travel decides when to seek and avoid it. Going east (the harder direction, because your natural clock runs slightly long) you want bright morning light at your destination and to avoid late-evening light; going west, the opposite — seek evening light, avoid early morning. Getting the timing wrong pushes the clock the wrong way, which is why apps like Timeshifter that compute a personalised light/melatonin schedule are genuinely useful. Evidence ●●●●○

For flights crossing 5+ time zones, a small dose of melatonin (0.5–3 mg) at destination bedtime meaningfully reduces jet lag — a Cochrane review put the number-needed-to-treat at 2, one of the best in sleep medicine. Meal timing helps too: fast on the plane and eat your first proper meal at the destination's normal breakfast time to anchor your body's food clock. Evidence ●●●●○

A 2024 study found a single large dose of creatine (~0.35 g/kg) improved cognition and reduced fatigue during a night of sleep deprivation — a promising buffer for the wrecked days of travel. Be honest about the limits: it was tiny (15 people), a one-off megadose (which can upset the stomach), and never tested for jet lag specifically. Worth watching, not established — and ordinary daily creatine (3–5 g) is safe and useful for other reasons but won't reproduce that acute rescue. Evidence ●●○○○

For chronic insomnia, cognitive behavioural therapy for insomnia (CBT-I) is the recommended first-line treatment — ahead of any drug. It's about as effective as sleeping pills in the short term and more effective long-term, because it fixes the habits and thoughts that perpetuate insomnia rather than masking them, with no dependence. It's increasingly available as effective apps (Sleepio, Somryst). Evidence ●●●●●

Slow, exhale-emphasised breathing shifts you toward the parasympathetic 'rest' state that permits sleep. A Stanford trial found five minutes a day of 'cyclic sighing' (a double inhale, then a long exhale) beat mindfulness for calming — and the 4-7-8 pattern and the single physiological sigh work the same way. Free, fast, and genuinely calming when the mind won't settle. Evidence ●●●○○

Mindfulness meditation has a real trial behind it for older adults' sleep, and NSDR / yoga nidra (guided 'non-sleep deep rest') is a benign way to rest deeply after a bad night and to drift off. Just don't believe the claim that NSDR replaces actual sleep — that part is untested and probably untestable. Evidence ●●●○○

Brief awakenings between sleep cycles are physiologically normal, and for most of human history people even slept in two chunks with a wakeful hour in between. A single night waking is not evidence that anything is broken — and treating it as a crisis is exactly what keeps you awake. The calm framing is itself part of the fix. Evidence ●●●○○

The rules: don't clock-watch and don't check your phone (both spike arousal). Do a few slow, long exhales or a body scan in bed to down-regulate; try cognitive shuffling — picturing a stream of random, neutral images to crowd out anxious thought. If you're still wide awake after ~20 minutes, get up, go to a dimly lit room, and do something boring until you're sleepy again — no bright light, no screens, no food, no work. This keeps your brain from learning that bed = lying awake frustrated. Evidence ●●●●○

The usual culprits: alcohol (as it clears, cortisol and heart rate rebound and fragment the second half of the night), stress and an early cortisol rise, a blood-sugar dip, needing to pee, and — if it comes with snoring or gasping — sleep apnea. A regular 3am wake with daytime symptoms is worth investigating, not just enduring. Evidence ●●●○○

The daytime toolkit: get bright light and movement early to lift alertness; use caffeine strategically in the morning and early afternoon (but not late, or you'll wreck tonight); and if you need it, take a short nap under ~20–30 minutes, early in the day — the 'NASA nap' that boosts alertness without the grogginess of deep sleep. Just don't nap long or late. Evidence ●●●○○

The strong urge is to go to bed super early, sleep in, and nap a lot to 'make it up.' Resist it. The mild tiredness from one bad night is your ally — it makes tonight's sleep come faster and deeper. Sleeping in and long naps bleed off that pressure and set up a second bad night, which is exactly how one bad night becomes chronic insomnia. Keep a fixed wake time. Evidence ●●●●○

Treating a bad night as a disaster creates the very anxiety that blocks the next night's sleep — a self-perpetuating loop. Remind yourself you can function, keep perspective, and let a single bad night just be a single bad night. The people who spiral into chronic insomnia are usually the ones who panic about the first few. Evidence ●●●○○

Relaxing music before bed is the best-evidenced of the lot: a Cochrane review found it improves how rested people feel (a moderate effect on subjective sleep quality). And any steady sound works as a masker — smoothing over the sudden noises that would otherwise jolt you awake in a noisy environment. Both are cheap, safe, and reasonable. Evidence ●●●○○

The honest read on the rest: white noise has weak, mixed evidence as a sleep improver (some studies even found it made sleep worse) — it's a masker, not a brain hack. Binaural beats, brown noise, and ASMR are popular but poorly supported. If they relax you, enjoy them — just don't believe the 'brainwave entrainment' claims. Evidence ●●○○○

The genuine science is closed-loop acoustic stimulation: quiet pink-noise pulses delivered in precise sync with your slow brain waves during deep sleep can boost that deep sleep (the memory benefit is shakier). But the magic is the timing, not the sound — a plain pink-noise playlist isn't the same thing; it needs a device that reads your brainwaves and times the pulses to them. Evidence ●●●○○ in the lab

The far-out options with genuine evidence: active bed cooling (Eight Sleep, ChiliPad) for hot sleepers; weighted blankets (a real trial backs them for anxiety-linked insomnia); better bedroom air / ventilation (lowering overnight CO₂ measurably improved sleep — cheap and badly underrated); and the maximalist routines (Bryan Johnson's 'Blueprint,' Andrew Huberman's toolkit) work because they do the boring fundamentals rigorously. Evidence ●●●○○

The most scientifically grounded gadget frontier is EEG headbands that time stimulation to your brainwaves — Elemind (2024) sped sleep onset by ~29% in its trial by nudging brain rhythms with sound. Real mechanism, real (but small, company-funded, early) data: interesting for early adopters, not yet proven enough to bank on. Cranial electrical stimulation (CES) and DIY brain stimulation are weaker still and best left to research. Evidence ●●○○○

Fun but unproven: red-light-as-therapy, rocking beds (real in the lab, barely available as a product), and elaborate supplement stacks. Outright skip (no plausible mechanism or evidence): grounding/earthing sheets, PEMF mats, and magnets — that's the snake-oil end. And mouth taping stays a skip unless a clinician has cleared you for apnea. Evidence ○○○○○ for the snake-oil tier

Every load-bearing claim comes from a named guideline, trial, meta-analysis, or validation study, with the evidence dots (●●●●● = strong and replicated, down to ○○○○○ = essentially unproven) reflecting study design, size, and replication. Figures are group averages; individuals vary widely. This is educational, not a diagnosis or a prescription — and where a popular book (like Why We Sleep) overstates a scary number, the guide uses the underlying literature instead.

Foundations & metrics: the AASM/Sleep Research Society duration consensus, the Sleep Regularity Index mortality studies, and Van Dongen's sleep-debt work. Environment: the Zeitzer, Chang & Mason-Zee light studies, the Haghayegh warm-bath meta-analysis, the Drake caffeine study, and the Stutz exercise-timing meta-analysis. Tracking: the Chinoy multi-device validation and the orthosomnia literature. Breathing: the AASM apnea guideline, WatchPAT validation, SURMOUNT-OSA (tirzepatide), and the mouth-taping safety review. Supplements & drugs: the AASM pharmacologic guideline, the low-dose-melatonin work and the melatonin-quality analyses, and the DORA/doxepin trials. Jet lag & mind: the Cochrane melatonin jet-lag review, the creatine sleep-deprivation study (graded as early), the ACP recommendation of CBT-I as first-line, and the Balban breathwork trial. The frontier: the Ngo and Papalambros closed-loop pink-noise work, the Elemind trial, the Ekholm weighted-blanket RCT, and the bedroom-CO₂ study.