Sleep and Longevity

Sleep matters for longevity because it is when your body does its core maintenance: clearing brain waste, filing memories, repairing DNA, resetting hormones, and arming the immune system. Both too little (6 hours or less) and too much (9 hours or more) track with dying earlier, so 7 to 8 hours is the longevity sweet spot.

While you lie there, five maintenance jobs run in parallel.

Your brain takes out the trash. At night a kind of cleaning crew gets to work: the glymphatic system, a drainage network around your brain's blood vessels, first mapped by Iliff and Nedergaard. It flushes spinal fluid through your brain tissue and washes out waste. That waste includes beta-amyloid, the gunk tied to Alzheimer's. The original mouse study in 2013 found the gaps between brain cells widen by about 60 percent during sleep, which roughly doubled how fast injected tracers got cleared ^[2]. Popular headlines stretched this into "the brain is 10x more active at clearing waste," which oversells the actual finding. Human brain scans show the fluid does move during sleep, in a 2019 study ^[3] and in the imaging series by Eide and Ringstad. None of them directly measured how much of the Alzheimer's-related proteins got cleared. And in 2024, a different lab using a different method found the opposite: less clearance during sleep, not more ^[4]. So the field is still arguing about it. What holds up: the fluid moves differently while you sleep, but how big the cleaning effect is depends on what you measure.

Your memories get filed. What you learned during the day moves from short-term to long-term storage. Deep sleep (also called slow-wave sleep, or N3, the deepest stage where your brain rolls out big slow waves) locks in facts and knowledge. REM sleep (rapid eye movement, the stage where most vivid dreams happen) handles skills and emotional memories.

Your hormones reset. Growth hormone, the stuff that repairs tissue, surges during deep sleep. Roughly two-thirds of your daily growth hormone gets released at night, tied to that first deep-sleep block. Sleep also sets your hunger hormones (leptin and ghrelin), your stress hormone cortisol (which climbs right before you wake, a normal morning spike), and prolactin.

Your immune system gets to work. While you sleep, your body builds and releases the proteins and antibodies that fight off infection. Skimp on sleep long enough and your vaccine response and your resistance to viruses both measurably drop.

Your cells repair their DNA. Repair work ramps up during sleep in nearly every type of cell studied.

So what does this mean for how long you live? A 2010 meta-analysis pooled 1.3 million people ^[1]. Short sleepers (6 hours or less) had a 12 percent higher risk of dying early. Long sleepers (9 hours or more) had a 30 percent higher risk. Picture a U-shaped curve, with the sweet spot in the middle. The long-sleep side may partly be people who are already sick, rather than the sleep itself doing harm. And a few famous cause-and-effect claims from popular sleep books have been picked apart (a 2019 critique of Why We Sleep). Treat that mortality curve as the solid baseline.

A night of sleep is not one flat block. Your brain runs through four stages (N1, N2, N3, and REM) over and over, in cycles of about 90 minutes.

The four stages. N1 (about 2 to 5 percent of the night in young adults) is the light drift from awake into asleep, the bit where you can still be nudged awake easily. N2 (about 45 to 55 percent) is the workhorse, where your brain fires off little bursts of activity (in the 11 to 16 Hz range) that help filter out noise so you stay asleep. N3 is the deep stuff, the slow-wave sleep, where your brain produces big, slow rolling waves. Sleep scientists count a chunk of the night as deep sleep when at least 20 percent of it is filled with those slow waves (under 2 Hz and at least 75 microvolts tall). In healthy young adults it makes up roughly 13 to 23 percent of the night. REM (about 20 to 25 percent) is the dreaming stage. Your brain looks almost as busy as when you're awake, your eyes dart around, and your body goes briefly limp so you don't physically act out your dreams.

The shape of the night. Each cycle runs about 90 minutes (anywhere from 70 to 120) and repeats four to six times. The mix is lopsided, though. Deep sleep is front-loaded: it dominates the first couple of cycles and often vanishes by the last one. REM is back-loaded, getting longer as the night goes on, with your longest dream stretch usually just before you wake up. So if you cut the night short at the end, you mostly lose REM. Cut it short at the start and you mostly lose deep sleep. That popular "wake at the end of a 90-minute cycle" trick is shakier than it sounds, because cycle length swings by 30 minutes or more between people and even within one night.

Why you fall asleep at all. Back in 1982, Borbély described sleep as a tug-of-war between two forces ^[5]. One is sleep pressure, which builds up the longer you're awake (think of it as a balloon slowly filling all day). The other is your body clock, the alerting signal from your brain's master timer that keeps you awake during the day and lets you fade at night. You fall asleep when the pressure gets high enough to overpower the clock. The molecule behind that building pressure is adenosine, a tiredness signal that piles up the longer you stay awake. A 1997 study showed adenosine climbs in the brain during long waking hours and drains away during recovery sleep ^[6]. Caffeine doesn't actually give you energy. It just blocks adenosine from docking, so it muffles the tiredness signal.

Your sleep changes shape as you age. A 2004 meta-analysis pulled together 65 studies spanning ages 5 to 102 ^[7]. As adults age:

• Total sleep drops about 10 minutes per decade.
• It takes longer to fall asleep.
• You spend more time in the light stages.
• You get less deep sleep and less REM.
• You wake up more in the night.

A 2000 study tracked the deep-sleep collapse precisely: in men it fell from about 19 percent of sleep at ages 16 to 25 down to about 3.4 percent at ages 36 to 50 ^[8]. That's roughly 38 minutes of deep sleep lost per decade through midlife, after which it levels off. Nighttime growth hormone dropped about 75 percent right alongside it.

That picture of "eight solid hours stuffed with deep sleep" describes a young adult. By 60, broken-up nights are just normal. So protecting whatever deep sleep you have left becomes the goal that matters for aging. Our deep sleep guide covers the specific levers for that.

Bad sleep ages you faster, and it does it through several doors at once.

It frays your telomeres. Telomeres are the protective caps on the ends of your chromosomes that get a little shorter every time a cell divides. Observational studies keep linking chronic sleep loss to shorter ones. The size of the effect bounces around study to study, but the direction always points the same way.

It rewrites your epigenetic clock. Your body tags your DNA with chemical markers that shift as you age, and those same markers are exactly what biological-age tests read. A 2018 study found that a single night without sleep changes the tags on your body-clock genes and dials down the genes that run your cells' power plants in your muscles.

It turns up inflammation. Lose sleep and three of the main inflammation markers in your blood go up. That feeds "inflammaging," the slow simmer of chronic inflammation that drives a lot of age-related disease.

It pushes you toward diabetes. One week of sleeping just 5 hours dropped insulin sensitivity in healthy young men by 11 to 20 percent ^[10]. So your body has to work harder to keep blood sugar in check. Push it harder still (4 hours a night for 6 nights) and glucose tolerance falls 30 to 40 percent, which lands you in pre-diabetes territory ^[11].

It ages your brain. The Whitehall II study followed about 8,000 British civil servants for 25 years ^[12]. People who regularly slept 6 hours or less at age 50 had a 22 percent higher risk of dementia later on.

And it cuts both ways. Aging messes with your sleep on its own. Older adults get less deep sleep and wake up more. So bad sleep speeds up aging, and aging makes sleep worse. A loop that feeds itself.

None of this is locked in. Better habits, treating any sleep disorders, and the other levers that slow aging can all slow these effects down.

The short answer: 7 to 9 hours if you're an adult aged 18 to 64, and 7 to 8 hours from 65 on. That's the National Sleep Foundation's range ^[55], and the American sleep-medicine societies agree on at least 7 hours for adults ^[56], so anchor there.

Age	Recommended sleep [55]	Typical deep sleep (N3)
18 to 25	7 to 9 h	about 19% of the night (men aged 16 to 25) [8]
26 to 64	7 to 9 h	about 3.4% at ages 36 to 50 (men) [8]
65 and older	7 to 8 h	stays low, with more waking at night [7]

Where's the sweet spot? A 2022 Nature Aging analysis of about half a million UK Biobank adults aged 38 to 73 found that about 7 hours went with the sharpest thinking and best mental health ^[50]. The big mortality studies point the same way: both short and long sleep track with dying earlier ^[1]. For most people, 7 to 8 hours is the safe zone. Go much shorter or much longer and the health risk climbs.

Quality beats quantity. Sleep is not just hours logged in bed.

• Sleep efficiency: time asleep ÷ time in bed. Aim for ≥85 percent.
• Sleep stages: you need enough N3 for physical recovery and enough REM for memory and thinking.
• Sleep continuity: unbroken sleep restores you more than the same hours chopped into pieces.

Regularity counts as much as hours. A 2024 UK Biobank study tracked sleep timing with wrist sensors in 60,977 adults. Compared with the least regular group, the more regular sleepers had a 20 to 48 percent lower risk of dying during follow-up, and regularity predicted death better than sleep duration did ^[48]. It's observational, so it shows a link, not proof. A fixed wake time, weekends included, is the cheapest fix.

Signs your sleep is good enough:

• You wake without an alarm feeling rested
• You hold energy through the day without leaning on caffeine
• You fall asleep in 15 to 20 minutes
• You don't wake up often at night
• You don't feel wiped out during the day

Signs it isn't:

• You need an alarm to get up
• You rely on caffeine to function
• You crash in the afternoon
• You fall asleep the instant your head hits the pillow (often a sign of sleep debt)
• You feel drowsy driving or in meetings

Your 2-week sleep diary. Before changing anything, measure. For 14 days, track: bedtime, wake time, caffeine cutoff, alcohol units, perceived sleep quality 1 to 5. Any notebook works. A wearable with a sleep log works too.

Two numbers to compute at the end:

• Sleep efficiency: time asleep ÷ time in bed. Target ≥85 percent.
• Sleep latency: time to fall asleep. Healthy is 15 to 30 min. Under 5 min = likely sleep debt. Over 30 min = onset insomnia.

Your sleep timing is set by light and temperature, not by willpower. You can't just decide to be a morning person.

Meet your master clock. Deep in your brain sits a tiny timer (the SCN, a small cluster of cells in the hypothalamus) that runs your 24-hour rhythm. It reads the time of day from special cells in your eyes, ones that aren't for seeing at all but are tuned to detect blue light (peaking right around 480 nm, the short-wavelength blue that floods the sky in daylight). A 2002 study identified those cells. A 2003 study then mapped how light shifts the clock: a 6.7-hour blast of bright light can pull your clock earlier by up to 2.0 hours or push it later by up to 3.6 hours ^[14]. Notice the asymmetry. Your clock slides later much more easily than it jumps earlier. That's why staying up late is easy and getting up early is brutal.

Morning light is the single biggest dial. Outdoor light on a clear day hits your eyes at 10,000 to 100,000 lux. A bright office gives you 300 to 500. Your living room in the evening, maybe 50 to 200. The problem: your body clock doesn't respond to light in a straight line, and normal indoor lighting sits in the "barely registers" zone for setting the clock. A 2013 study sent volunteers camping for a week. Their daytime light went up fourfold, their biological night snapped into line with the natural light-dark cycle, and the whole group drifted earlier, closer to the sun. The expert consensus: aim for at least 250 lux of clock-relevant light at your eyes during the day ^[13].

What to actually do: get outside within 30 to 60 minutes of waking, for 10 to 30 minutes, ideally without sunglasses. Sitting by a window cuts the dose roughly in half compared to stepping outside. Even a cloudy day beats indoor lighting by 10 to 100 times.

Evening light is the flip side. A 2000 study showed that light at night switches off melatonin, the hormone that tells your body "it's nighttime," and it does it on a sliding scale ^[15]. Room light around 100 lux already cuts melatonin roughly in half, and you can see an effect below 30 lux. A 2015 study found that 4 hours of reading on a glowing eReader at night pushed melatonin back by about 1.5 hours and left people groggier the next morning, hours after the screen was off ^[16]. The 2022 consensus: keep clock-relevant light at 10 lux or less in the 3 hours before bed, and 1 lux or less while you sleep ^[13]. Most modern living rooms blow past that 10-lux mark from the ceiling LEDs alone. Blue-blocker glasses cut the impact by about half at typical screen brightness, but dimming the whole room works better.

Temperature opens the door to sleep. Across the night your core body temperature drops about 0.5 to 1.0 °C, bottoming out around 4 or 5 in the morning. That drop happens because your body dumps heat through your hands and feet, where the blood vessels open up near the surface to let warmth escape. A 2000 study found that the temperature gap between your hands and feet and your trunk (basically, how warm your extremities are versus your core) predicts how fast you fall asleep better than anything else ^[17]. Better than core temperature, better than melatonin, better than how tired you say you feel. Warm hands and feet are the signal that sleep is on the way.

How cold should the bedroom be? A 2012 review found that under normal bedding, room temperatures from about 13 to 23 °C barely changed sleep stages, while heat, especially humid heat, cut both deep sleep and REM ^[18]. So lean cool. Around 16 to 20 °C (61 to 68 °F) is a sensible rule of thumb, not a lab-proven number. Cool the room, but warm the extremities (socks help if you tend to have cold feet).

Your chronotype is real, and mostly in your genes. Twin studies peg it at 40 to 50 percent heritable. A 2012 study looked at social jetlag, the gap between when you sleep on your days off and when your alarm drags you up for work, basically how badly your schedule fights your biology. Beyond sleep duration, it was linked to a higher BMI ^[52]. In the heavier subgroup, every hour of social jetlag came with about 33 percent higher odds of being overweight. You can nudge your chronotype 1 to 2 hours with disciplined light and dark timing, but you can't fully override it.

Caffeine, alcohol, nicotine, cannabis and late heavy meals all make sleep worse. One rule covers most of them: stop each one at least 3 hours before bed. Caffeine needs 6 hours or more, and 8 or more if you're sensitive.

They are also exactly what most people reach for in the evening.

Caffeine: the one you're underrating. Caffeine sticks around in your body anywhere from 2 to 10 hours, depending on how fast your liver clears it. Some people are genetically fast at this, others slow. A 2013 study found that 400 mg of caffeine taken 6 hours before bed cut total sleep by more than an hour compared to a placebo ^[19]. The old "stop at 2 PM" rule only holds if you go to bed at 10 and process caffeine at the average speed. If you're a slow clearer, you'll do better cutting off at noon, or sticking to mornings only. And decaf isn't zero: a 2006 lab analysis found 3.0 to 15.8 mg of caffeine in a single shot of decaf espresso.

Alcohol: knocks you out, then wakes you up. A 2013 review summed it up ^[20]. Booze helps you drop off fast and deepens the first half of the night. But in the second half, as the alcohol clears your blood, your sleep shatters into fragments and you keep half-waking. REM gets squashed early and then rebounds late, often with vivid dreams. Wearables routinely show a high resting heart rate and a low HRV (heart-rate variability, a quick read on how well your body is recovering overnight) for 24 to 48 hours after just one or two drinks. And it makes sleep apnea worse, because it slackens the muscles in your airway. Rule: finish at least 3 hours before bed, cap it at 1 drink on sleep nights, and skip it entirely if you have apnea.

Nicotine: a stimulant in a relaxant costume. It's a stimulant, full stop, hitting the same alertness switches in your brain as your natural wake-up chemistry. It makes you take longer to fall asleep, breaks the night up, cuts both your total sleep and your deep sleep, and squashes REM ^[21]. It also stirs up restless legs and twitchy limbs at night. The vaping research is thinner but points the same way as smoking. Rule: last nicotine at least 3 hours before bed. Quitting is the only real path back to normal sleep, and expect 2 to 4 weeks of rough nights while you do.

Late, heavy meals. A 2019 study of 296 apnea patients found that people who habitually ate late had a slightly higher apnea score, took longer to fall asleep, lay awake longer in the night, and got less REM ^[22]. It's an observational link and the effects were small, but they all point the same way. Why? Digesting a meal makes heat, and that heat fights the natural cool-down your body needs to fall asleep. Lying down on a full stomach also worsens reflux and apnea, and overnight blood-sugar spikes rev up your fight-or-flight system. Rule: last meal at least 3 hours before bed, and try to keep your eating to a 10 to 12 hour window.

Cannabis: a trade now, a bill later. A 2017 review laid it out plainly ^[23]. A hit of THC helps you fall asleep faster and, at low doses, might briefly bump up deep sleep, but it squashes REM, your dreaming sleep. Worse, you build tolerance fast, so the sleep benefit fades. And when you stop, you get insomnia and a flood of vivid dreams (REM bouncing back hard after being held down). A recent placebo-controlled brainwave study of oral THC/CBD found it cut REM by 8 percent and pushed back the start of REM by 66 minutes. CBD on its own, at calming doses, doesn't disrupt sleep structure. Rule: nightly THC for sleep is a tolerance trap. If you use it, keep the dose low and don't do it every night. Brace for 1 to 2 weeks of bad sleep when you stop.

Sleeping pills: Z-drugs out, the newer class in. The 2017 AASM guideline gave only weak backing to every sleeping-pill class and pushed CBT-I (the talk-therapy protocol) as first choice ^[24]. The old Z-drugs (zolpidem, zaleplon, eszopiclone) and benzodiazepines come with a list of problems:

• They squash your deep sleep.
• You build tolerance.
• They cause falls and weird half-asleep behaviors.
• Observational data show a worrying mortality signal: a 2012 analysis found people on them died at 3.6 to 5.3 times the rate, rising with dose ^[53]. That's an observational link, not proof the pills caused the deaths.

In April 2019 the FDA put a boxed warning, its strongest, on zolpidem, zaleplon and eszopiclone. The trigger was reports of serious injuries and deaths from sleepwalking, sleep-driving and other things people did while not really awake.

The newer class, DORAs (suvorexant, lemborexant, and daridorexant), works differently: instead of sedating you, it just blocks orexin, your brain's stay-awake signal, so it lets sleep happen while leaving your REM and deep sleep intact. The big trials of daridorexant (1,854 patients across two trials at 1 and 3 months) showed it genuinely cut the time spent awake in the night and the time to fall asleep, improved next-day functioning, and looked safe ^[25]. Rule: CBT-I first. If you do need a pill, pick a DORA over a Z-drug, especially for long-term use or if you're older.

Everyday meds that mess with sleep. Beta-blockers (especially the fat-soluble ones like propranolol and metoprolol that cross into the brain) shut down your nighttime melatonin and bring vivid dreams and insomnia. A small 2012 trial found that 2.5 mg of melatonin restored sleep quality in patients on beta-blockers. SSRIs (a common antidepressant class) roughly double the time it takes to reach REM and cut your total REM. Steroids cause insomnia, especially taken in the evening. ADHD stimulants delay sleep if you take them after noon. Rule: if insomnia shows up within weeks of starting a new medication, suspect the medication. Where it makes sense medically, move the dose to the morning.

Four habits all pull the same lever, your core body temperature, and they're behind most of the drug-free sleep gains people chase. The evidence ranks them pretty clearly.

Exercise helps, modestly but reliably. A 2019 meta-analysis of 23 evening-exercise studies found that working out in the evening actually added 1.3 percentage points of deep sleep (p=0.041), pushed back the start of REM by 7.7 minutes, and trimmed light sleep by 0.9 points ^[26]. So the old "never exercise at night" rule is dead. The only thing that hurts is hard exercise that ends within an hour of bed. Anything you wrap up at least an hour before lights-out is neutral to helpful, deep sleep included.

What kind of exercise? Steady cardio gives the most reliable boost to deep sleep, because it heats you up and the cool-down afterward triggers your sleep signal (our exercise guide covers how to build that base). Lifting weights improves how good your sleep feels (measured by a standard sleep-quality questionnaire) more than it adds measured deep sleep. Interval training is the one where timing really matters: a moderate session 90 minutes or more before bed is fine, but going all-out too close to bedtime will keep you up.

Timing. Morning exercise tugs your body clock earlier by about 0.6 hours per session (from a 2019 study mapping how exercise shifts the clock). That's handy if you're a night owl trying to shift earlier, fighting eastward jet lag, or just want an earlier bedtime. Evening exercise (finished at least an hour before bed) won't disturb your sleep and gives deep sleep a small lift.

The hot bath is the most powerful trick of all. A 2019 meta-analysis pooled 13 trials ^[27]. Water at 40 to 42.5 °C (104 to 108.5 °F) for at least 10 minutes, 1 to 2 hours before bed, cut the time to fall asleep by about 36 percent (a large effect, Cohen's d around 1.01) and improved sleep efficiency. The mechanism is backwards from what you'd guess: warming up your skin makes the blood vessels near the surface open wide. Then when you step out into a cool room, all that opened-up circulation dumps heat fast, and your core temperature drops sharply over the next 60 to 90 minutes. That falling core temperature is what tells your body to sleep.

Hot-bath recipe: 40 to 43 °C water, 10 to 15 minutes, finish 60 to 120 minutes before lights-out, then step into a cool, dim room.

Sauna. The heart benefits are well established (the Finnish KIHD study, 2015, in 2,315 middle-aged men: 4 to 7 sessions a week linked to about 50 percent lower heart-disease death versus once a week ^[51]), but the direct sleep evidence is thin. Mostly it's people reporting they slept better, plus one tiny 1976 study (just 5 people) that showed 70 percent more deep sleep in the first 2 hours. The mechanism overlaps with the hot bath. A sensible evening session: 15 to 25 minutes at 80 to 90 °C, 1 to 2 hours before bed, with a cool (not cold) shower after.

Cold plunges. In a 2000 study, an hour in 14 °C water raised norepinephrine, an alertness chemical, about sixfold (up 530 percent) ^[49]. Even a short plunge pushes it up. Great at 7 AM, a disaster at 9 PM. Morning or midday cold is fine. Cold within an hour of bed will wire you up and delay sleep. The Finnish sequence (sauna, then a cool shower, then bed) is reasonable, but keep the cold to a quick 30 to 60 seconds rather than a long plunge.

Cooling mattresses and the bedroom itself. A 2008 study showed that gently warming the skin of the hands and feet by just 0.4 °C doubled deep sleep in older adults, from 8 to 14 percent ^[28]. An independent 2024 crossover study (72 people) found a heat-absorbing mattress added 7.5 minutes of deep sleep and lowered heart rate by about 2.4 bpm. The Eight Sleep and chiliPad studies funded by the makers report bigger numbers, but nobody independent has reproduced them. A cool bedroom (16 to 20 °C) gets you most of the benefit for free.

For long-term insomnia, sleeping pills are not the first step. Every major sleep society (the AASM, the European Sleep Research Society, the UK's NICE, the American College of Physicians) now recommends a talk-therapy protocol called CBT-I (Cognitive Behavioral Therapy for Insomnia) first. Pills come second, are meant to be short-term, and stop working fast.

Does it actually work? A 2015 meta-analysis pooled 20 trials (1,162 patients): CBT-I cut the time to fall asleep by 19 minutes, cut time awake in the night by 26 minutes, and raised sleep efficiency by 9.9 percentage points ^[29]. And the gains stuck around at follow-up. A 2012 review found it matches sleeping pills in the short term and beats them over the long haul. A 2017 trial followed 160 patients for two years and saw 44 to 63 percent go into remission ^[54]. The people who weaned off zolpidem while sticking with CBT-I did better than the ones who kept taking it as needed.

What's actually in it. CBT-I is a structured, short course, usually 4 to 8 sessions, built from five pieces:

1. Stimulus control (a 1972 method). Retrain your brain to link the bed with sleep, not with lying there frustrated.
2. Sleep restriction (a 1987 method). On purpose, spend less time in bed so your sleep packs together instead of scattering.
3. Cognitive restructuring. Catch and rewrite the catastrophic thoughts about sleep ("I'll be useless tomorrow").
4. Relaxation training. Progressive muscle relaxation, slow paced breathing.
5. Sleep hygiene. The basics: caffeine, alcohol, light, exercise, temperature.

Sleep hygiene is the weakest piece. The 2021 AASM guideline actually recommends against using sleep hygiene on its own ^[30]. The real engine is sleep restriction plus stimulus control.

Sleep restriction does the heavy lifting. It feels backwards, but it's the most powerful part. The protocol:

1. Keep a sleep diary for 7 to 14 days and work out your average actual sleep time.
2. Set your time in bed to that average plus 30 minutes, but never below 5 hours. Pick a fixed wake-up time first, then count backward to find your bedtime.
3. Hold that window for 7 days. You'll feel sleepy. That's the whole point: it cranks up your sleep pressure.
4. Adjust weekly. If your sleep efficiency hits 85 percent or more, go to bed 15 to 30 minutes earlier. Between 80 and 84 percent, hold steady. Below 80 percent, trim 15 minutes.
5. Keep going until you settle into a sustainable window (usually 6.5 to 7.5 hours) where you sleep efficiently.

A 2011 trial tested a stripped-down 4-session version in older adults with chronic insomnia ^[31]. 67 percent responded and 55 percent went into remission. The number needed to treat was just 2.4: treat two to three people and you get one extra success.

Stimulus control rules.

• Get into bed only when you're genuinely sleepy (heavy eyelids, not just tired).
• Use the bed for sleep and sex, nothing else.
• If you're still awake after about 20 minutes, get up. Do something quiet in dim light, then go back when you're sleepy again.
• Wake up at the same time every day, weekends included.
• No daytime naps while you're doing the protocol.

The app version works too. A 2020 trial randomized 1,721 Norwegian adults: digital CBT-I beat plain patient education by a large margin on insomnia severity (Cohen's d of −1.21), with 58 percent showing a clinically meaningful improvement ^[32]. Germany's DiGA program covers the somnio app on a statutory-insurance (GKV) prescription. And in the UK, NICE guidance from 2023 (TA922) only backs the drug daridorexant once CBT-I has been tried and didn't work, or isn't available or suitable, and wants it reviewed within 3 months.

When to bring in a specialist: you've done 6 to 8 weeks of structured CBT-I at home with no luck; your insomnia-severity score is still above 15 at week 8; your days are badly impaired; or you also have apnea, restless legs, REM behavior disorder, major depression with suicidal thoughts, shift work, or you recently used Z-drugs and need help tapering off.

Not every sleep problem yields to good habits and CBT-I. Some need a doctor. Here's how to tell, and what they'll do.

Sleep apnea (OSA, where your airway keeps collapsing shut during the night). It's more common than people think. In adults with a moderate-or-worse case ^[33]:

• Men 30 to 49: about 10 percent
• Men 50 to 70: about 17 percent
• Women 30 to 49: about 3 percent
• Women 50 to 70: about 9 percent

What raises your odds: extra weight, a thick neck, being male, getting older, and a recessed lower jaw.

Screen yourself with STOP-BANG, a quick checklist: loud snoring, daytime tiredness, anyone observed you stop breathing, high blood pressure, a BMI over 35, age over 50, neck over 40 cm, and male sex. Three or more puts you at middling risk; five or more is high risk. To actually diagnose it, you get either an overnight lab sleep study (with sensors for brainwaves, breathing, and movement) or a portable take-home version for straightforward cases.

The treatment ladder: CPAP, the mask that splints your airway open with a gentle stream of pressurized air, is the gold standard. If you can't tolerate it, or your case is mild to moderate, there are jaw-positioning mouthguards. Then positional therapy, weight loss, and a small implant that nudges your tongue nerve to keep the airway open (Inspire; the 2014 STAR trial dropped the apnea score from 29 down to 9 at 12 months). The newest tool is a drug: a 2024 trial found tirzepatide cut the apnea score by 20 to 24 breathing events per hour beyond placebo in weight-related apnea over 52 weeks ^[34]. On 20 December 2024 the FDA approved it (as Zepbound) for moderate-to-severe OSA in adults with obesity. It is the first drug approved for OSA itself, not just for its symptoms.

Two big trials, SAVE ^[35] and ISAACC ^[46], failed to show CPAP cutting heart attacks and strokes. But people in both wore the mask only 2.8 to 3.3 hours a night on average, and both studies enrolled patients who weren't even sleepy, the group least likely to feel a benefit. So CPAP is still first-line if your apnea has symptoms.

Insomnia disorder. Covered in the CBT-I section. The clinical bar: trouble sleeping at least 3 nights a week, for at least 3 months, that wrecks your days, despite having enough time to sleep.

Restless legs (RLS, also called Willis-Ekbom disease). An itch-like urge to move your legs, worse when you're at rest, eased by moving, and worst in the evening and night. Always get your iron checked first (ferritin and transferrin saturation, the two key markers for how much iron you've stored and how it's moving). The 2024 AASM guideline strongly recommends IV iron for people with the right iron status ^[37]. It starts iron top-ups when ferritin is under 75 ng/mL or transferrin saturation is under 20 percent. When ferritin sits between 75 and 100 ng/mL, it points specifically to IV iron. A class of nerve-calming drugs (gabapentin enacarbil, gabapentin, pregabalin) is the preferred first medication. The old first choice, dopamine drugs (pramipexole, ropinirole), is now advised against for standard long-term use (a conditional recommendation), because they tend to backfire over time and make the symptoms worse (about 7 percent of patients per year, adding up).

REM sleep behavior disorder (RBD). Normally your body is paralyzed during dreams. In RBD that switch fails, so people physically act out their dreams: kicking, punching, shouting, leaping out of bed. This is the most urgent red flag in this whole guide. A 2019 study (an international group of 1,280 patients across 24 centers) found that 6.3 percent of them per year went on to develop Parkinson's, Lewy body dementia, or multiple system atrophy, reaching 73.5 percent within 12 years ^[36]. RBD is the single strongest early warning sign for that family of brain diseases (the ones tied to a misfolding protein, which includes Parkinson's). Get referred to neurology, not just for symptom relief (a doctor may prescribe clonazepam or melatonin at 3 to 12 mg at night) but for ongoing monitoring and a possible spot in trials of brain-protective treatments.

Body-clock disorders. The common one is delayed sleep-wake phase disorder, the genuine night owls (5 to 16 percent of teens and young adults). This is a diagnosed condition, so see a sleep clinician before you start. The standard fix combines bright morning light with a tiny dose of melatonin (0.3 to 0.5 mg) taken 5 to 7 hours before you usually fall asleep, timed to when your body naturally starts releasing melatonin in the evening. The mirror-image early-bird disorder, shift-work disorder, and the non-24-hour rhythm seen in fully blind people each have their own protocols.

Narcolepsy. Type 1 is a shortage of orexin (your brain's stay-awake chemical) plus cataplexy, sudden muscle weakness set off by emotion. Type 2 has no cataplexy and normal orexin. Diagnosis takes an overnight study plus a daytime nap test (you fall asleep in 8 minutes or less on average and drop into REM unusually fast in at least 2 naps). A specific immune-system gene variant (HLA-DQB1*06:02) shows up in more than 95 percent of type 1.

Periodic limb movements in sleep: these are leg twitches that show up on a sleep study. You only treat them if they're actually disturbing you. Same iron-first approach as restless legs.

Red flags: when to get referred.

• Loud snoring + witnessed pauses in breathing + daytime sleepiness → apnea workup
• Acting out your dreams → RBD → see neurology
• Insomnia lasting 3 months or more → CBT-I, not a nightly Z-drug
• Sleepy all day (Epworth score 10 or more) despite enough time in bed → sleep clinic
• Bad evening leg discomfort → check ferritin, try a nerve-calming drug
• Nodding off behind the wheel → urgent apnea or narcolepsy workup
• Snoring + morning headaches + blood pressure that won't respond to treatment → strongly suggests apnea

Where to go in Germany, Austria, and Switzerland. Germany: the DGSM (Deutsche Gesellschaft für Schlafforschung und Schlafmedizin) keeps the list of accredited sleep centers. The overnight study is covered by statutory insurance (GKV) through a stepwise process, and CPAP machines and supplies are reimbursed as medical aids, with your usage checked over time. The somnio app is prescribable on insurance (as a DiGA) for CBT-I. Austria: the ÖGSM, with sleep labs at AKH Wien, Graz, and Innsbruck. Switzerland: the SGSSC, where the Fähigkeitsausweis Schlafmedizin is the formal sleep-medicine qualification.

Sleep tracking is a real tool now, just a limited one. The lab sleep study (polysomnography, or PSG) is the reference standard. Yet even two trained humans scoring the same study only agree about 83 percent of the time (an agreement score around 0.80, where 1.0 is perfect) ^[39]. That's the ceiling. No gadget's algorithm can be more accurate than the experts it's trained against.

Wrist and ring trackers: good at total sleep, rough at stages. A 2025 study tested them against the lab (62 people) and measured how far off they were on total sleep time ^[47]:

• Apple Watch Series 8: off by about 28 minutes
• Fitbit Charge 5: about 31 minutes
• Whoop 4.0: about 50 minutes
• Garmin Vivosmart 4: about 54 minutes

They're great at spotting when you're asleep (91 to 96 percent of the time). But they're bad at spotting when you're awake (29 to 52 percent), so they tend to score restless time as sleep ^{[38, 47]}. If your nights are broken up, or you have insomnia, your device is probably making things look better than they are.

How well they read sleep stages (agreement with the lab, where 1.0 is perfect):

• Apple Watch S8: about 0.53
• Fitbit Sense: about 0.42
• Whoop 4.0: about 0.37
• Garmin Vivosmart 4: about 0.21

The Oura Gen3 ring looks strong in its own 2024 validation (96 people, 421,045 epochs against multi-night at-home lab recordings) ^[40]: about 92 percent accuracy for asleep versus awake, and stage accuracy from about 76 percent (light sleep) to about 91 percent (REM). Its 0.83 score is a different agreement measure (sleep vs wake only), so don't line it up against the wrist numbers above. One caveat: Oura funded that study, so read it with that in mind.

Deep-sleep detection runs 47 to 70 percent accurate across wrist devices, with the Oura Gen3 around 80 percent. None of them actually measure deep sleep. They guess at it from your heart rate and movement, proxies that hold up across a crowd but get noisy for any single person on any single night.

Home EEG headbands: the only ones that truly read your brainwaves. The Dreem 2/3 headband hit about 0.74 agreement with the lab on deep sleep, which matches what two humans manage ^[39]. The consumer version was discontinued in 2021; Beacon Biosignals keeps it going for research. The current consumer pick is the Muse S Athena (around $475), a 4-sensor forehead EEG that hit about 0.76 in vendor-backed testing. The Z-Machine Insight+ only does asleep-versus-awake.

Sound-based deep-sleep boosters. Some devices play soft pink-noise pulses timed to your slow brainwaves, and in small studies this nudged up slow-wave activity (around 8 percent) and helped lock in memories (a 2013 study and a 2017 study). The brainwave boost reliably repeats; the memory benefit is hit or miss. As of May 2026, there's no FDA-cleared, peer-reviewed consumer device for this. Philips SmartSleep was killed in 2023, Dreem in 2021.

Apnea screening. The Withings Sleep Analyzer, an under-mattress pad (validated in 2021 on 118 people), catches moderate-to-severe apnea about as well as a take-home test (its accuracy score landed at 0.93 to 0.95, where 1.0 would be flawless) ^[41]. That study was run largely by Withings staff, so treat it as promising, not settled. Recent Apple Watch models can also flag possible sleep apnea from breathing disturbances at night. Treat that as a nudge to get tested, not a diagnosis. A positive result still needs a proper take-home or lab test to diagnose, and you'll need that test for insurance to cover treatment.

HRV (heart-rate variability) overnight. Your overnight reading roughly tracks your morning one, but it gets muddied by the way your heart shifts across sleep stages and brief wake-ups. Morning HRV is the cleaner number. Take it lying down for 5 minutes, before caffeine, after you've peed. And watch the 7-day rolling average, not any single day.

What to buy:

• For total sleep time: any modern wearable. The Apple Watch is the most accurate wrist option, the Oura the best ring.
• For measuring deep sleep: the Muse S Athena. Expect 75 to 85 percent accuracy at the stage level.
• For apnea screening: the Withings Sleep Analyzer (about €130) at the moderate-apnea threshold, then off to a clinic.
• For recovery via HRV: a Polar H10 chest strap plus a morning HRV app (HRV4Training) is the cheapest gold-standard setup (about $90, no subscription).

Take maker-funded validations with a grain of salt (Whoop, Eight Sleep, and the Oura Gen4 all lack independent review). The cleanest references are the 2025 cohort ^[47], the 2024 head-to-head study, and the 2020 Dreem paper ^[39].

Most "sleep supplements" work on falling asleep, calming anxiety, or lifting mood, not on the depth of your slow-wave sleep, which is the part that matters most for aging. That gap between the marketing and what's actually proven is the big problem in this whole category. Our deep sleep guide digs into the deep-sleep specifics.

What can legally make a sleep claim in the EU:

• Melatonin gets exactly two: it "contributes to the alleviation of subjective feelings of jet lag" (at 0.5 mg or more, taken near bedtime on the travel day) and it "contributes to the reduction of time taken to fall asleep" (at 1 mg, near bedtime). That's the whole list. Bigger doses buy you little: a 2013 meta-analysis found higher doses shaved only a few extra minutes off falling asleep and added no benefit for how well you slept ^[42]. The overall effect is modest either way: about 7 minutes faster to fall asleep and 8 minutes more sleep. The 10 mg gummies you see in US stores are wildly overdosed and tend to leave you groggy the next morning with weird dreams and nothing to show for it. In Germany, regulators start to treat anything above 1 mg as a medicine. A statement from the Federal Institute for Risk Assessment (BfR) on 17 September 2024 points that way. In 2023 the Higher Regional Court of Koblenz (OLG Koblenz, case 9 U 1947/22) went further: advertising even a 1 mg melatonin supplement as sleep-promoting counts as a regulated health claim. That tightened, not loosened, what marketers may say.

What helps you fall asleep and feel better (but not deeper):

• L-theanine (an amino acid from tea, 200 mg, 30 to 60 minutes before bed). A 2019 trial showed sleep quality improving over 4 weeks ^[43]. It works by calming a busy, anxious brain. Best if your problem is lying there with your mind racing. No EU claim.
• Saffron extract (affron), 14 to 28 mg a day over 4 to 6 weeks, improved how good people's sleep felt (a 2020 trial and a 2021 trial). The effect runs through mood, not through deep sleep, and both trials were industry-funded. No EU claim.
• Magnesium (bisglycinate is the gentle-on-the-gut form), up to 250 mg of elemental magnesium in the evening, which is the EU's safe upper limit for supplements. A 2021 meta-analysis of 3 small trials in older adults found magnesium cut the time to fall asleep by about 17 minutes, but the evidence quality was low ^[44]. None of those trials tested bisglycinate specifically. The glycine it's bound to may add a small effect of its own (more in the deep sleep guide). Its EU claim: it "contributes to a reduction of tiredness and fatigue."
• Glycine, 3 g about 30 minutes before bed. A small 2007 trial shortened the time to reach deep sleep and made sleep feel better, though it did not add any actual deep sleep ^[45]. It seems to work by gently widening blood vessels in your hands and feet (a 2015 study), the same warm-extremities trick that signals sleep. No EU claim.

What's mostly hype:

• Taurine on its own: the brain-chemistry story is plausible, but no human trial has ever tested it with sleep as the main outcome.
• "Sleep formulas" that cram 8 to 12 ingredients in at tiny doses are basically expensive multivitamins.
• Diphenhydramine and doxylamine (the over-the-counter antihistamine sleep aids) come with a hidden cost: they blunt thinking and memory, especially in older people, and regular use is linked to a higher dementia risk in older adults. Geriatric guidelines (the AGS Beers criteria) advise against them for anyone 65 and up.

For deep sleep specifically. The real deep-sleep levers are mostly behavioral (sleep restriction paradoxically makes deep sleep rebound, exercise raises it, a cool room lets it happen) and avoiding the wrong substances (alcohol kills early-night deep sleep, late caffeine cuts it). Drugs that genuinely deepen sleep are still experimental or reserved for specific conditions. The supplements with the most defensible deep-sleep case are magnesium bisglycinate (where the glycine carrier is half the story) and a lemon verbena plus zinc combination. Neither has strong lab-grade human evidence. The deep sleep guide has the full breakdown.

The one question to ask. When a product promises to deepen your sleep, ask: was that measured with a proper brainwave study, with a wearable's guess, or just with a questionnaire? Nearly every "deep sleep" claim in the supplement aisle is wearable-derived or subjective. The mechanism stories might be real. The hard human evidence for deeper sleep usually isn't.

"I can't fall asleep." Start with the basics: bedroom at around 16 to 20 °C, last caffeine at least 6 hours before bed (8 or more if you're sensitive), no alcohol in the last 3 hours, no screens or warm lamps in the hour before bed, and a fixed wake-up time. If it's been dragging on for 3 months or more, do CBT-I (sleep restriction plus stimulus control). If it kicked off within weeks of a new medication, suspect the medication.

"I wake at 3 AM and can't get back to sleep." That's your sleep fragmenting. The usual causes:

• Alcohol: sleep breaks up as the alcohol leaves your blood.
• Apnea: snoring plus witnessed pauses plus a morning headache means run the STOP-BANG checklist and get tested.
• Depression or anxiety: the 3 AM wake-up is a textbook sign of depression.
• Getting up to pee: often linked to heart and metabolic health.
• Perimenopausal hot flashes.

Treat the root cause, not the symptom.

"I sleep 8 hours but wake up exhausted." Almost always a quality problem: undiagnosed apnea, poor sleep efficiency, sleep broken up by alcohol or late meals, or depression. Grab a wearable and check your overnight resting heart rate (it'll be high) and HRV (it'll be low) as a quick read on how hard your body worked. If you snore, screen for apnea. If your mood is off, see a clinician.

"I'm a night owl and can't function in the morning." Probably a delayed body clock. This is a diagnosed condition, so a sleep clinician can confirm it and supervise the fix. The standard approach pairs morning outdoor light within 30 minutes of your target wake-up with a tiny dose of melatonin (0.3 to 0.5 mg) taken 5 to 7 hours before your target bedtime, shifting your clock 30 to 60 minutes a week. It takes steady effort over 2 to 4 weeks. And if you're a genuine night owl by nature, sometimes the smarter move is to accept it and build your schedule around it instead of fighting it forever. The social-jetlag data suggest that a schedule which keeps fighting your biology comes at a metabolic cost ^[52].

"I get jet lag for a week." Flying east is harder than flying west, because your body clock slides later more easily than it jumps earlier ^[14]. What works: time your light (morning light at your destination going east, evening light going west), take a tiny dose of melatonin (0.3 to 0.5 mg) timed to your new evening melatonin window, and use a little caffeine in the morning. Shifting your schedule 2 to 3 days before you fly helps too. Our free jet lag calculator turns those rules into clock times for your flight.

"I snore loudly and my partner complains." Screen for apnea with STOP-BANG. If you score 3 or more, push for a take-home or lab test. Snoring on its own isn't apnea. But loud snoring plus daytime sleepiness plus witnessed pauses in breathing is the high-suspicion trio.

"I wake up screaming or thrashing." Acting out your dreams (REM behavior disorder) needs a neurology referral. A 2019 study found 73.5 percent of these patients went on to develop Parkinson's, Lewy body dementia, or multiple system atrophy within 12 years ^[36]. Sleepwalking and night terrors are different (they come out of deep sleep, not dreams) and are usually set off by sleep deprivation, alcohol, or fever. Treat the trigger and make the bedroom safe.

"My legs feel weird and I can't stop moving them." That's the restless-legs pattern (worse at rest, worse in the evening, eased by moving). Check your iron (ferritin and transferrin saturation). If ferritin is 100 ng/mL or lower, iron is worth discussing: tablets or IV below 75 ng/mL, IV specifically between 75 and 100 (AASM 2024) ^[37]. Nerve-calming drugs (gabapentin enacarbil, pregabalin) are first-line; steer clear of long-term dopamine drugs.

"I take sleeping pills every night and want to stop." Don't quit benzodiazepines or Z-drugs cold turkey. The rebound insomnia is brutal. Taper off under medical supervision over 4 to 8 weeks or more. Replace them with CBT-I, ideally running it alongside the taper rather than waiting until after. If you need a pharmacological bridge, daridorexant or another of the newer DORAs is safer than a Z-drug.