Deep Sleep (Slow-Wave Sleep)

Deep sleep, slow-wave sleep, and N3 are three names for the exact same thing. Sleep doctors call it N3. On a brain-wave readout it shows up as slow-wave sleep (SWS). "Deep sleep" is just the friendly word your watch puts on the screen.

So what counts as N3? A sleep lab looks at your brain waves and tags a chunk of the night as deep when at least 20 percent of it is filled with big, slow waves (2 Hz or slower, at least 75 µV tall, which is the standard cutoff for calling a 30-second window deep). Those slow waves are huge crowds of brain cells switching on and off in sync. While that happens, the memory part of your brain (the hippocampus) fires off quick bursts (150 to 250 Hz) that tuck neatly inside the slow waves, stitched together by a third rhythm in between. When those three rhythms line up, that is the moment your brain files away the day's memories.

When does deep sleep actually happen? In healthy young adults, N3 makes up roughly 13 to 23 percent of the whole night. And it is heavily front-loaded. Most of it lands in your first two sleep cycles. The last cycle of the night is usually dream-heavy (REM) and deep-sleep-poor. Here is why that matters. Cut the night short at the back end (5 hours instead of 8) and you mostly lose dream sleep. Wreck the front half (a late bedtime, or alcohol fragmenting your first cycles) and you mostly lose deep sleep.

N3 is not the only sleep that counts. N2 is the middle gear, and it is the biggest slice of your night at about 45 to 55 percent. It repairs you too. The little bursts inside N2, called sleep spindles, are themselves a window into how sharp your brain is. So treating N3 as the only "good" sleep is a shortcut that misses the point. You need every stage.

Then why does N3 get all the longevity hype? Three reasons. First, growth hormone. Your body releases it during deep sleep, and a 1997 study made the link plain: when researchers used a drug (gamma-hydroxybutyrate) to push more deep sleep in healthy young men, growth hormone rose right along with it ^[2]. About two-thirds of your daily growth hormone comes out during night sleep, mostly during that first deep stretch. Second, the brain's overnight cleaning crew. The fluid that rinses waste out of your brain pulses in time with deep-sleep waves, and it flows hardest during deep-sleep-rich periods ^{[5, 6]}. Third, N3 is the stage that fades earliest and fastest as you age (more on that next).

A quick word on sleep spindles. Those little N2 bursts track with how well you think, and they fade in Alzheimer's. Nudging them is a real research goal. One method, called closed-loop auditory stimulation (tiny pulses of pink noise played at just the right moment during deep sleep), reliably boosts slow-wave activity, and in some studies the memory that follows ^{[10, 11]}. The slow-wave boost shows up again and again. The memory payoff is shakier.

Three things tie deep sleep to a longer, healthier life. Growth hormone. Brain cleaning. Memory. Here is how each one works.

1. Growth hormone and keeping your body in good repair. Researchers spent two decades tracking how deep sleep connects to the system that releases growth hormone (the brain-to-pituitary-to-liver chain that pumps it out). The headline finding came in 2000. As deep sleep dropped from about 19 percent of the night in young men (ages 16 to 25) to about 3.4 percent by midlife (36 to 50), night-time growth hormone fell about 75 percent right alongside it ^[1]. And the link runs both ways. A 1997 study showed that pushing more deep sleep with a drug (gamma-hydroxybutyrate) raised growth hormone in proportion ^[2]. Adult growth hormone is not the same as the kind that makes kids grow taller, but it still drives tissue repair, fat burning, IGF-1 (a growth signal), and bone renewal. That is one of the best reasons to protect your deep sleep as you get older.

2. The brain's overnight rinse cycle. A 2013 study found that the gaps between brain cells in mice open up about 60 percent during sleep, which roughly doubles how fast the brain flushes out amyloid-beta (the gunky protein that clumps up in Alzheimer's) ^[5]. A 2019 human study caught the same thing live: every big deep-sleep wave was followed by a pulse of fluid washing into the brain ^[6]. The flow is strongest during deep-sleep-rich stretches. One honest caveat. A 2024 study used a different method and found the brain actually clears less during sleep, so just how much waste gets washed out is still up for debate ^[7]. What everyone agrees on: the fluid moves differently while you sleep, and the change is biggest in deep sleep.

3. Memory. Two big reviews laid out the model for how sleep files memories ^{[8, 9]}. During deep sleep, your memory hub replays the day's experiences ^[39] and locks them to two other brain rhythms that nest inside the slow waves. That lineup broadcasts each memory to the rest of your brain, over and over, slowly moving it into long-term storage. Lose deep sleep and your recall for facts takes a hit. The pink-noise trick (closed-loop auditory stimulation) can boost slow-wave activity, and in some studies the memory that follows. But the memory effect is fragile and does not always show up again.

The dementia link. A 2017 review gathered the evidence that age-related deep-sleep loss is tied to thinning in the front of the brain, fewer spindles, and steadily worse memory filing ^[4]. And one big study put numbers on it. The Whitehall II cohort (n=7,959) found that habitually sleeping 6 hours or less at age 50 came with a 22 percent higher risk of later dementia ^[12]. Staying a short sleeper through ages 50, 60, and 70 pushed that up to about 30 percent. The likely path: less deep sleep means worse brain cleaning plus worse memory filing.

What deep sleep does NOT do. It is not a substitute for total sleep. Eight hours of light sleep beats four hours with extra-deep N3. Every stage earns its keep. "Chase deep sleep, ignore the rest" is a marketing line, not science. Here is the honest version: for a given amount of total sleep, more deep sleep is better. But never trade away total sleep just to bump up your deep-sleep percentage.

Deep sleep is the stage that disappears earliest and fastest. Roughly speaking, you lose about a quarter of your deep sleep per decade from young adulthood until the system flattens out by midlife: from about 19 percent of the night in your early 20s to about 3.4 percent by your late 40s. And it is not a slow trickle.

The big age curve. A 2000 study tracked 149 healthy men aged 16 to 83 across two nights in a sleep lab ^[1]. Deep sleep dropped from about 19 percent of the night at ages 16 to 25 to about 3.4 percent at ages 36 to 50, a loss of roughly 38 minutes per decade up to midlife. Then it flattened out. Dream sleep held fairly steady until midlife, then fell about 10 minutes per decade after 50. So your deep sleep starts collapsing a full 25 years before your dream sleep does.

Night-time growth hormone fell about 75 percent from young adulthood to midlife, lockstep with the deep-sleep loss. Evening cortisol (the stress hormone) only crept up after age 50. This is the single most-cited dataset on aging and deep sleep.

The big meta-analysis. A 2004 review pooled 65 studies spanning ages 5 to 102 (n=3,577 healthy people) ^[3]. Same story: deep sleep drops steadily, sharpest between young adulthood and middle age. Women lose it more slowly than men.

Why does this happen? The 2017 review pins it largely on thinning of grey matter in the front of the brain, which is the very region that generates slow waves in the first place ^[4]. Throw in fewer spindles, an earlier body clock (you start nodding off sooner), and more nightly fragmentation, and you get the classic older-adult sleep pattern: shorter nights, less deep sleep, more light sleep, more waking up.

Men and women age differently here. Women hold onto more deep sleep than men through most of adulthood. But perimenopause and menopause bring their own disruption. Hot flashes and night sweats break up sleep, often worst in the first half of the night, which is exactly where deep sleep lives. And falling estrogen reshapes sleep on its own, separate from the hot flashes.

What this means for you. At 25, your deep sleep is plentiful and tough to dent. At 50, you are running on roughly a third of what you had. So protecting what is left (with the levers below) matters more the older you get. The flip side is sobering: the older you are, the harder any intervention works. You can hold steady or claw back a little. You cannot rebuild what your brain no longer makes.

When illness speeds it up. A few conditions accelerate the normal decline. Obstructive sleep apnea (your airway repeatedly collapses while you sleep) shreds the night and blocks solid deep sleep. Chronic insomnia and untreated depression cut it too. And plenty of drugs (Z-drugs, benzodiazepines, fat-soluble beta-blockers, SSRIs) suppress deep sleep or scramble its brain-wave signature. Treat the underlying cause and deep sleep often bounces back measurably.

At home, the most accurate option you can buy is a forehead brain-wave band (the Muse S Athena, around 75 to 85 percent stage-level accuracy); a ring or watch like the Oura Gen3 only catches about 64 percent of deep sleep and is noisy on any single night, so read it as a 7-day trend, never a one-night verdict. You cannot improve what you cannot measure, and most of the "deep sleep" numbers on your wearable are noisy guesses. Here is what is actually real.

The gold standard. A full in-lab sleep study (polysomnography, or PSG) reads your brain waves directly and scores deep sleep by the official rules. But here is the surprise. Even two trained human experts only agree on about 83 percent of the night, window by window (that's a Cohen's kappa around 0.80, a standard agreement score) ^[35]. That is the ceiling. No gadget can be more accurate than the experts the rules are built on.

How accurate are wearables at deep sleep? A 2025 head-to-head against PSG (n=62) gives clean numbers for catching deep sleep:

• Apple Watch Series 8: ~51 percent
• Fitbit Sense / Charge 5: ~51 percent
• Whoop 4.0: ~70 percent (best wrist device for deep sleep in that group)
• Withings Scanwatch: ~67 percent
• Garmin Vivosmart 4: ~48 percent

The best of the bunch is the Oura Gen3 with OSSA 2.0. A 2024 study ran it against several nights of at-home PSG (421,045 short windows): overall accuracy 91.7 percent, kappa (PABAK) 0.83, which matches the agreement between two human experts ^[36]. For deep sleep specifically it hit about 64 percent. Worth noting: Oura Health paid for the study.

How to read all of this without fooling yourself. Wearables do not see your brain waves. They guess at deep sleep from your heart-rate patterns and how much you move. That guess holds up well across a group of people. It is noisy on any single night. So an Oura or Whoop saying "only 30 minutes of deep sleep last night" could easily be 30 minutes off in either direction. Watch the 7-day rolling average, not one night. Use it for the trend.

When you want real brain-wave data at home. The Muse S Athena (about $475) is the best consumer option right now for staging sleep off actual brain waves. It is a 4-sensor forehead band, with overall agreement around 0.76 in vendor-backed tests. Deep sleep is harder to read from a forehead-only setup than from a full lab hookup, but this is a genuine brain-wave reader, not a heart-rate-and-motion guess.

The Dreem 2/3 was the research-grade home brain-wave headband. A 2020 study clocked its deep-sleep agreement around 0.74, on par with two human experts ^[35]. Dreem stopped selling to consumers in 2021; the tech now lives on under Beacon Biosignals for research. If you can find a used Dreem 2 in good shape, it is still the best home brain-wave reader for deep sleep.

The Z-Machine Insight+ nails sleep-versus-wake but is weaker at scoring the individual stages than Dreem or Muse.

The pink-noise trick (closed-loop auditory stimulation, or CLAS). Tiny pulses of pink noise played at the right moment during deep sleep can boost slow-wave activity in research labs ^{[10, 11]}. The Philips SmartSleep DeepSleep Headband was the consumer version; Philips killed it in 2023. As of May 2026, no FDA-cleared, peer-reviewed consumer CLAS device exists. Real CLAS needs to read your brain waves live and fire in time with them, not just loop pink noise on a timer. Anything sold as "AI-powered deep-sleep enhancement" without brain-wave reading and timing is hot air.

What to actually buy if deep sleep is your target:

• Best independent option you can buy today: Muse S Athena. Expect 75 to 85 percent accuracy at the stage level. Watch trends.
• Best ring or wrist guess: Oura Gen3 with OSSA 2.0 firmware. Use it for 7-day rolling averages, never one night.
• A used Dreem 2/3 if you find one: the cleanest home deep-sleep measurement going, but no vendor support left.
• Don't pay for "deep sleep enhancement" gadgets with no published, peer-reviewed PSG validation behind them.

The strongest deep-sleep levers are things you do, not things you swallow. Here are the ones with real evidence.

The strange rebound from cutting sleep short. This is the most powerful lever, and the most backwards-feeling. Your deep-sleep system runs on pressure. Stay awake longer and the "need" for slow waves builds up; sleep again and deep sleep bounces back both in minutes and in intensity. That is the exact trick behind the sleep-restriction part of insomnia therapy (CBT-I). You deliberately spend less time in bed than you think you need, and what you do get is dense, efficient, deep-sleep-rich. A 1982 paper laid out the math for this: the slow waves per hour of sleep climb as sleep pressure builds up.

For a healthy adult without insomnia, one night of 6-hour sleep gives you a noticeably deep-sleep-rich rebound the next night. But chronic short sleep is not a plan. It does not give you more deep sleep. It gives you a deeper percentage inside a smaller total, and you pay for it with lost dream sleep and lost hours. For someone with chronic insomnia, structured sleep-restriction therapy (CBT-I) pulls a fragmented night back together and brings deep sleep back.

Aerobic exercise. A 2019 meta-analysis of 23 evening-exercise studies found that working out in the evening raised deep sleep by +1.3 percentage points (p=0.041) versus doing nothing ^[18]. Aerobic training (the kind that gets your heart pumping) gave the cleanest signal across the wider research. The reason is heat: aerobic exercise pushes your core temperature up 1 to 2 degrees C, and the overnight cool-down deepens your non-dream sleep. Heavy lifting makes you rate your sleep better too, but it does not move measured deep sleep as cleanly.

In practice: 3 to 4 aerobic sessions a week (30 to 60 min, moderate to hard), ideally finishing more than 60 minutes before bed. Evening exercise that wraps up at least an hour before bed is a net plus for deep sleep.

Keep your timing locked. A scattered or shifted schedule gives you scattered deep sleep. Same bedtime and wake time every day, weekends included, anchors your body clock and supports solid deep sleep. Social jet lag (your weekend sleep landing an hour or more later than your weekday sleep) measurably costs you deep sleep the next weekday. The research links social jet lag to metabolic disease, and the deep-sleep angle is a plausible part of why.

Get a handle on stress and racing thoughts. When your nervous system is stuck in fight-or-flight at bedtime, driven by stress hormones like adrenaline and cortisol, it suppresses deep sleep. CBT-I tools (stimulus control, reframing anxious thoughts) target this for insomnia. For everyone else, an evening wind-down that dials down that arousal (a warm bath, slow breathing, dim light) helps deep sleep by quieting the system before bed.

Fix what's fixable. Untreated sleep apnea is the most common cause of severe deep-sleep loss in adults. Each apnea event jolts you awake just enough to kill deep sleep. Effective CPAP often brings a dramatic deep-sleep rebound in the first weeks. Restless legs syndrome breaks up deep sleep the same way; topping up iron (when ferritin is below 75 ng/mL) and certain nerve medications restore it. If you have moderate-to-severe apnea and don't know it, nothing in this guide, behavioral or supplement, will recover your deep sleep until the apnea is treated.

Temperature is the best-proven, no-drug lever for deep sleep, and almost nobody uses it.

How it works. You fall asleep when your core temperature drops, and you drop it by dumping heat out through your hands and feet (your blood vessels there widen to release warmth). A 2000 study found that how much warmer your hands and feet are than your torso predicts how fast you fall asleep better than anything else ^[13]. Better than core temperature. Better than melatonin. Better than how sleepy you feel. That heat loss carries on all night, and your core hits its low point around 04:00 to 05:00.

Deep sleep responds to skin temperature in particular. A 2008 study put older adults in a temperature-controlled suit and nudged their skin warmth in tiny 0.4 degrees C steps during a sleep study. Warming the skin by just 0.4 degrees C nearly doubled deep sleep (about 8% to 14%) and cut night-time awakenings (p<0.001) ^[14]. That is one of the biggest single-move deep-sleep effects in all the human sleep-lab research. The temperature window for deep sleep is genuinely tight, which makes skin warmth a precise dial.

Cool room, warm feet: the winning combo. A 2012 review of the temperature-and-sleep research landed on 16 to 19 degrees C (60 to 67 degrees F) under normal bedding as the sweet spot for adults ^[16]. Push past 26 degrees C with humidity and both deep sleep and dream sleep fall apart. So the move is to keep the air cool (so your skin has somewhere to send heat) while your hands and feet stay warm (a warm bath before bed, socks if you run cold).

The hot-bath protocol. A 2019 meta-analysis pooled 13 trials of warming the body before sleep. Water at 40 to 42.5 degrees C for 10 or more minutes, 1 to 2 hours before bed, helped people fall asleep 36 percent faster on average (Cohen's d about 1.01) ^[15]. Why it works: warming your skin opens up those blood vessels; stepping into a cool room then dumps heat fast; and your core temperature slides down over the next 60 to 90 minutes. That drop is the sleep signal. The deep-sleep effects in the meta-analysis lean positive, but most of those studies were not measuring deep sleep as their main goal.

Cooling mattresses and heat-soaking bedding. An independent 2024 study across three centers (n=72, where everyone tried both options) tested a high-heat-capacity mattress against a regular one and found +7.5 minutes of deep sleep over a 7.5-hour night (p=0.0038) plus a lower heart rate (-2.4 bpm) ^[17]. Modest but real, and not funded by a mattress company.

Studies paid for by the makers of the Eight Sleep Pod claim bigger wins (+14 minutes of deep sleep), but they carry obvious conflicts of interest and methodology gaps. The honest read: a heat-soaking cooling cover plausibly adds 5 to 15 minutes of deep sleep over a normal night, mostly for people who tend to sleep hot. The marketing numbers from Eight Sleep and chiliPad run well ahead of what the independent data show.

Sauna. Direct sleep-trial evidence is thin. A tiny 1976 Finnish study (n=5) reported +70 percent deep sleep in the first 2 hours after an evening sauna, but no modern trial has repeated it. The logic is the same as a hot bath: a sauna raises your core temperature about 1 to 1.5 degrees C and forces a strong cool-down afterward. Its heart benefits are well documented in a large Finnish study. Think of sauna as a hot bath with a stronger upside for your heart and weaker direct evidence for sleep.

Your practical temperature protocol for deep sleep:

• Room air: 17 to 20 degrees C (62 to 68 degrees F). Cool enough that you want a duvet.
• Warm hands and feet: a warm bath or shower 60 to 120 minutes before bed, or bed socks if your feet run cold.
• Hot bath before bed: 40 to 43 degrees C, 10 to 15 minutes, 60 to 120 minutes before lights-out, then into the cool dark bedroom.
• Optional: a heat-soaking mattress topper or cooling cover. Expect about 5 to 15 minutes more deep sleep over your baseline, not the marketing figure.

Here is the uncomfortable truth. Most "sleep supplements" do not deepen your deep sleep at all. They help you fall asleep faster, take the edge off anxiety, make you rate your sleep higher, or shift your body clock. Those are real wins. They are just not the same as adding deep sleep on a brain-wave readout. And that gap is the central lie in the whole sleep-supplement aisle.

Here's how to sort them, by what they actually do, judged against real brain-wave evidence in humans.

Tier 1: The closest thing to a real deep-sleep lever (and even here it's thinner than the marketing claims):

• Magnesium bisglycinate plus glycine (together or separate). Glycine at 3 g has brain-wave evidence of getting you to deep sleep faster, with no change in how much you get ^[19]. Bisglycinate delivers magnesium plus glycine, and how much glycine depends on your dose. More in the next section.
• Lemon verbena extract plus zinc, the only consumer combo with any human trial measuring a thing called "deep sleep." That measurement is a Fitbit, not a sleep lab. It gets its own section.

Tier 2: Faster sleep and a nicer feeling, not deeper sleep:

• L-theanine 200 to 400 mg: calms you without knocking you out, nudges relaxed brain waves, helps you drift off. No brain-wave evidence for deepening sleep.
• Saffron extract (affron) 14 to 28 mg/day for 4 to 6 weeks: better self-rated sleep and mood. The effect builds over weeks. No deep-sleep evidence.
• Melatonin at low dose (0.3 to 0.5 mg), timed to when your own melatonin starts rising in the evening: this shifts your body clock, handy for jet lag and a delayed sleep schedule. It gets you to sleep about 7 minutes faster in the meta-analysis. It is NOT a depth booster. High doses (5 to 10 mg) don't improve sleep quality and leave you groggy in the morning.

Tier 3: A plausible mechanism but no human sleep trial:

• Taurine on its own: the receptor story makes sense, but no human sleep trial with sleep as the main outcome exists.
• Most multi-ingredient "sleep formulas": 8 to 12 ingredients at doses too small to do anything, basically pricey multivitamins.

Tier 4: Actively bad for deep sleep:

• Diphenhydramine, doxylamine (the antihistamine sleep aids at the drugstore): they carry an anticholinergic load, raise dementia risk in older adults (the AGS Beers list says avoid them past age 65), and you build tolerance within days.
• Z-drugs and benzodiazepines: they cut deep sleep and distort its brain-wave signature; plus tolerance, dependence, falls in older adults, and mortality signals from observational studies.
• Alcohol: it cuts deep sleep in the first half of the night and fragments the second half ^[32].
• High-dose melatonin (5 to 10 mg): no better than 0.5 to 1 mg, and it brings a morning hangover and vivid dreams.

The one big honest line. No pill has been shown, in published sleep-lab trials, to meaningfully deepen deep sleep in healthy adults. Glycine slightly shortens the time it takes to reach deep sleep without changing the amount. Magnesium shortens how long it takes to fall asleep in people who are deficient. Lemon-verbena trials use a Fitbit, not brain waves. Melatonin doesn't move deep sleep in any real way. L-theanine works on relaxed brain waves and anxiety, not on slow waves.

So if a product promises to "increase your deep sleep," ask the seller for the sleep-lab study. The answer is almost always that the measurement came from a wearable (Oura, Fitbit, Eight Sleep Pod) or a questionnaire (PSQI, LSEQ). Those aren't worthless. They are just not the same as deep sleep measured on brain waves.

The real place to deepen your sleep is everything you do and where you sleep, the levers in the last two sections. Supplements help you fall asleep and feel better, which are worthwhile goals. But on today's human evidence, they do not meaningfully deepen your deep sleep.

Of all the sleep supplements, this pairing has the most defensible evidence. Still modest, but defensible.

Glycine has the cleaner story. A 2007 trial gave 3 g of glycine 30 minutes before bed to people who slept badly. The sleep lab showed they fell asleep faster AND reached deep sleep faster than on placebo, plus they rated their sleep better and felt less sleepy by day ^[19]. But here is the catch that matters most: the shape of their sleep did not change. Glycine got them to deep sleep sooner. It did not give them more of it.

A 2012 study went further: 3 g of glycine plus 25 percent sleep restriction over three nights cut fatigue scores (p=0.022) and sped up reaction time ^[20]. A 2015 study worked out how glycine does it. It nudges your master body clock (a tiny brain region) to widen the blood vessels in your skin and drop your core temperature, which is what tips you into sleep faster ^[21]. (The researchers ruled out glycine's own receptor and pinned it to a different one, and removing the clock in animals erased the effect entirely.)

Two caveats on the glycine trial: a glycine maker (Ajinomoto) funded it, and the group was small. But the mechanism is clean and repeatable. Effective dose: 3 g, 30 minutes before bed. On safety, your body already takes in and makes more than 11 g of glycine a day (per a Norwegian risk assessment), so an extra 3 g sits comfortably inside normal levels. It has been tested up to 9 g/day in healthy adults with no trouble.

Magnesium has a thinner story. A 2021 meta-analysis pooled three trials (n=151) of magnesium in older adults with insomnia: it cut the time to fall asleep by 17.4 minutes (95% CI -27.27 to -7.44, p=0.0006); total sleep time rose 16 min, but that wasn't statistically meaningful ^[22]. One single trial in there reported about a 6-minute bump in deep sleep over placebo, but that was inside the measurement noise, not pooled, and not meaningful on its own. The overall evidence quality was rated low to very low. The authors put it bluntly: "the quality of literature is substandard for physicians to make well-informed recommendations."

The most-cited magnesium-insomnia trial, from 2012, gave 46 elderly insomniacs 500 mg of elemental magnesium (as oxide) twice a day or placebo for 8 weeks ^[23]. Self-rated sleep improved (Insomnia Severity Index p=0.006, time to fall asleep p=0.02). Their hormones moved in a good direction (melatonin up, cortisol down). But there was no sleep lab. And the group was already low on magnesium to start, so you can't stretch the result to healthy adults who have plenty.

The "magnesium bisglycinate is two actives in one" idea. Neat on paper, only half-true in the numbers. Magnesium bisglycinate is one magnesium atom bonded to two glycine molecules (molecular weight about 172), so by weight it is about 14 percent magnesium and about 86 percent glycine. At a 400 mg elemental magnesium dose, you swallow about 2,860 mg of the whole compound, which works out to roughly 2.5 g of glycine. That is 83 percent of the studied 3 g sleep dose.

At a more typical 200 mg elemental magnesium dose, you only get about 1.24 g of glycine, under half the studied dose. So the two-in-one idea only fully holds at the high end. If your label says "1,000 mg providing 140 mg elemental magnesium," the glycine that rides along is about 860 mg, far below the dose used in the glycine trial.

What to actually do for deep sleep:

• Take magnesium bisglycinate at 200 to 400 mg of elemental magnesium in the evening (gentle on your gut, absorbs well; the EU supplemental limit is 250 mg, so at 400 mg you are over the EU limit and nearing the NIH limit of 350 mg).
• Take glycine on its own at 3 g about 30 minutes before bed if you specifically want its sleep-onset effect.
• Don't assume the glycine riding along in your magnesium covers the sleep dose at normal magnesium amounts.

Magnesium L-threonate (Magtein). Sold as the superior sleep magnesium on the back of one rat study (better brain uptake) and one industry-funded Oura-ring human trial (AIDP-funded). Premium price, no sleep-lab evidence, and a premium the independent data don't justify. Skip it.

Magnesium taurate is a defined magnesium salt (about 9 percent elemental magnesium, bound to taurine). Both halves have a plausible effect on the brain (taurine acts on calming receptors; magnesium blocks an excitatory one). For someone focused on heart health it is reasonable; for deep sleep, bisglycinate is the cleaner pick.

The honest scorecard:

• Glycine 3 g gets you to deep sleep faster but doesn't add more: Moderate.
• Magnesium cuts time-to-fall-asleep in deficient or elderly people: Moderate.
• Magnesium raises measured deep sleep in healthy adults who already have enough: Weak / Unproven.
• Bisglycinate delivers a useful glycine dose at your daily magnesium amount: Only at the top end (400 mg elemental magnesium).

Lemon verbena plus zinc is the only consumer combo marketed specifically for "deep sleep" that can point to any human trial measuring sleep stages at all. But the honest reading is blunt: that data comes from a Fitbit, not a sleep lab, and the combination itself has never been tested on its own in a published, peer-reviewed sleep-lab trial.

What's actually in the bottle. The lemon verbena extract behind nearly all these "deep sleep" claims is RelaxPLX (also sold as PLX) from Monteloeder S.L. in Alicante, Spain, a subsidiary of SUANNUTRA. It is standardized to 28 percent or more total phenylpropanoids and 24 percent or more verbascoside (its key active), at a clinical dose of 400 mg/day. It is not the Frutarom/IFF or PLT Health Solutions ingredient, which are different lemon verbena products. And Robuvit (Horphag) is unrelated entirely: that is oak wood, not lemon verbena.

The two key human trials, both of lemon verbena alone:

The 2022 trial. RelaxPLX 400 mg/day, no zinc, n=40, 8 weeks. It measured sleep quality scores, perceived stress, blood cortisol, and a Fitbit Charge 2. No sleep lab. Sleep-quality scores dropped 12.2 percent at 8 weeks (p<0.05); cortisol fell 15.6 percent; perceived stress fell 10.7 percent. The widely-quoted "increase in deep sleep" is a Fitbit estimate from movement and heart rate, not deep sleep measured on brain waves. And a Fitbit only catches roughly 40 to 60 percent of deep sleep versus a sleep lab in independent tests. Industry-funded (Monteloeder sponsored it) ^[24].

The 2024 trial. Aloysia citrodora extract, 90 days, n=71. It measured sleep-quality scores, a visual scale, a movement tracker, and night-time melatonin. No sleep lab. The self-rated measures improved; night-time melatonin rose 14 percent. Three of the authors work for Monteloeder ^[25].

What about zinc on its own? A 2017 review walks through how zinc tweaks several brain receptors involved in sleep ^[26]. A 2015 study found that a zinc-yeast extract raised non-dream sleep in mice by 20 to 30 minutes per hour for about 6 hours after a dose, without distorting the brain-wave signature (so it looked like natural sleep, not drugged sleep) ^[27]. Mice, not people. Human evidence for zinc alone comes down to self-rated sleep quality in elderly folks taking multi-ingredient blends.

The combination itself. A focused search of the literature turns up no peer-reviewed trial testing lemon verbena plus zinc on its own against placebo with sleep-lab endpoints. The products that pair RelaxPLX with zinc bisglycinate (and often other herbs) have never been published as a tested formula.

Some names get dropped in marketing (Buguet, Carrillo-Vico, Funes, Marhuenda). They show up in the broader sleep and supplement literature, but not on a published lemon-verbena-plus-zinc deep-sleep trial. If a seller cites those names as deep-sleep proof for this combo, the citation is misleading.

Mouse sleep-lab data for lemon verbena alone. A 2025 study tested the extract in mice at 40 to 160 mg/kg (which scales to roughly 230 to 900 mg in a human, overlapping the 400 mg trial dose) ^[28]. Non-dream sleep rose 110 to 194 percent; dream sleep rose 73 percent. The mechanism: it ramped up an adenosine receptor and reshaped a calming receptor in a way that's distinct from how benzodiazepines work. Real sleep-lab data, in mice. Not the same as a human trial.

The honest verdict. Lemon verbena plus zinc is a reasonable, low-risk, modestly-studied stack for self-rated sleep, with mouse sleep-lab support for the lemon verbena half. It is NOT a proven "deep sleep booster." Calling it "the only combination on the list with direct deep-sleep human data" overstates the evidence by at least one rung.

Where it sits: about level with magnesium glycinate, ashwagandha, or saffron for sleep quality. Plausible mechanism, modest self-rated improvement, industry-funded human trials, no sleep-lab confirmation of deeper sleep. Worth a look as part of a broader stack if you want help falling asleep and feeling rested. Not worth paying a premium for "deep sleep enhancement."

Safety. The zinc upper limit (EFSA) is 25 mg/day total (NIH puts it at 40 mg). Long-term doses above 25 mg risk copper deficiency, a type of anemia, low white cells, and nerve problems. Keep your total daily zinc (food plus supplement) at or below 25 mg unless a test shows you're deficient. Lemon verbena (per its EMA monograph) is a traditional herbal medicine with no notable safety flags at trial doses. Avoid it in pregnancy and while breastfeeding.

How to say it straight. "Lemon verbena standardized to verbascoside, with a low dose of zinc, has a plausible biological story: verbascoside acts on a calming receptor and an adenosine receptor, and zinc tweaks two more. Two small industry-funded trials of the lemon verbena alone show modest gains in self-rated sleep and cortisol. There is no human sleep-lab study of this combination, and the 'deep sleep' figure everyone quotes from the 2022 trial is a Fitbit estimate, not deep sleep measured on brain waves. The mouse sleep-lab data are real but rodent. A reasonable low-risk option, not a proven deep-sleep booster."

These three turn up in nearly every consumer "sleep formula." They work on falling asleep, anxiety, and how you rate your night. They do not work on the depth of your deep sleep. That difference is the whole point.

L-theanine (200 to 400 mg, 30 to 60 minutes before bed). A 2008 study nailed down the mechanism: a single 50 mg dose raised relaxed brain waves (alpha, 8 to 12 Hz) at the front of the brain. Those are the waves of being calm but awake, not of being asleep.

A 2019 crossover trial: n=30, 200 mg/day for 4 weeks in healthy stressed adults. Overall sleep-quality scores improved (p=0.013); time to fall asleep, disturbance, and medication use all got better. Anxiety scores p=0.006. Self-rated measures only, no sleep lab ^[29]. A 2011 trial of 98 boys aged 8 to 12 with ADHD, 400 mg/day for 6 weeks, used a movement tracker and showed better sleep efficiency, though time to fall asleep didn't budge. And a 2020 review of 9 trials concluded that 200 to 400 mg/day cuts stress and anxiety under pressure, which is probably the real reason any sleep benefit follows.

The honest take: L-theanine calms you down. It promotes relaxed brain waves. It makes you rate your sleep better. But no published L-theanine trial has measured deep sleep on brain waves and shown it getting deeper. Best fit: people kept awake by racing thoughts and bedtime tension. No EU health claim.

Saffron extract (affron, from the saffron crocus stigma). It works on serotonin, the mood chemical. Its active compounds act a bit like antidepressants in early research, and the antidepressant signal is saffron's strongest clinical finding. So its sleep benefit probably rides on better mood and less anxiety, rather than any direct knock-you-out effect.

A 2020 trial: n=55 adults with self-reported poor sleep, 28 mg/day of affron for 4 weeks. Insomnia scores improved from 15.75 to 11.74 (Cohen's d=1.07) versus placebo's 14.74 to 13.46 (d=0.31, group-by-time p=0.017). Restorative-sleep scores d=0.72 versus 0.10. Funded by Pharmactive Biotech, the maker of affron ^[30]. A 2021 trial of n=66 on 15.5 mg/day of saffron extract for 6 weeks saw improvements on a movement tracker and sleep-quality scores. Funded by Comercial Quimica Masso.

Every positive saffron sleep trial is industry-funded and uses self-rated scales or a movement tracker. The effect builds over 4 to 6 weeks. No sleep-lab evidence of deeper sleep. Best fit: self-rated sleep quality where mood plays a part. No EU health claim.

Taurine on its own as a sleep aid. The mechanism is lab-bench only: it acts on calming receptors. A focused literature search turns up no human trial of taurine by itself with sleep as the main outcome. The "taurine and sleep" data out there involve caffeine-plus-taurine energy-drink studies on simulated tasks, or liver and heart trials that never measured sleep. Taurine has a clean story for your cell powerhouses (see our mitochondria guide) and a phase III trial in a rare disease called MELAS. For sleep, the case is purely theoretical.

If a multi-ingredient "sleep formula" throws taurine in with glycine and magnesium, you can't credit any benefit to the taurine. Calling it a "deep sleep" ingredient stretches from receptor biology and animal data, not from human sleep-lab evidence. No EU health claim.

Melatonin is flat-out NOT a depth supplement. Worth saying loudly, because almost everyone assumes the opposite. Melatonin is a body-clock signal: it tells your brain it's dark and shifts your timing. A 2013 meta-analysis (19 studies, n=1,683) found it cut time-to-fall-asleep by about 7 minutes, added 8 minutes of total sleep, and nudged self-rated quality (SMD 0.22) ^[31]. No sleep-lab evidence that it adds any deep sleep. A 2024 review confirms the same: melatonin shifts your clock, it doesn't deepen your sleep.

High doses (5 to 10 mg) work no better for sleep quality (the meta-analysis found no extra quality benefit from going higher) ^[31]. They flood your blood with far more than your body makes, and it lingers into the morning as grogginess and vivid dreams. The 10 mg gummies sold across US shelves are pharmacologically way too strong.

Melatonin's real use is narrow: jet lag (especially flying east), a delayed sleep schedule, shift work, and circadian disorders in blind people. The EU allows two claims (jet lag at 0.5 mg or more; faster sleep onset at 1 mg). In Germany, products above 1 mg start drifting into drug territory (a BfR statement of 17 September 2024; an OLG Koblenz ruling, 9 U 1947/22, May 2023, eased the rules for 1 mg or below). And EMA's Circadin (2 mg slow-release) is approved for insomnia in adults 55 and over.

The pattern. Not one of L-theanine, saffron, taurine, or melatonin has sleep-lab evidence for deepening deep sleep in humans. They help you fall asleep, feel better, lift your mood, or fix your timing. All real and worth having. Just not the same as deepening your deep sleep.

Some things actively kill your deep sleep. If you want to protect it, these are the first to go.

Alcohol. A 2013 review of decades of sleep-lab studies sums it up: alcohol gets you to sleep faster and deepens the first half of the night at high doses ^[32]. Then the second half falls apart, with fragmentation and more waking, as your blood alcohol drops. Dream sleep is squashed in the first half and rebounds vivid and choppy in the second. Across most studies, your total dream sleep for the night goes down. Deep sleep often spikes early (that's the sedative kick), but the overall cost to your sleep is steep: a broken second half, lower heart-rate variability the next morning, and a stressed-out nervous system for 24 to 48 hours after even one or two drinks. Alcohol also worsens sleep apnea by relaxing the muscles that hold your airway open.

There is no "safe" amount of alcohol for protecting deep sleep. If you drink, finish at least 3 hours before bed, keep it to about 10 g (one drink) on a sleep night, and skip it entirely if you have diagnosed apnea. Your wearable's heart-rate-variability tracker will show the hit for days.

Z-drugs and benzodiazepines (zolpidem, zopiclone, eszopiclone, zaleplon, the classic benzodiazepines). These boost a calming brain receptor, so they get you to sleep faster and add total hours. But they also cut deep sleep in many sleep-lab studies and distort its brain-wave signature. You build tolerance within weeks. Next-day fog, falls, and bizarre sleep behaviors (sleepwalking, sleep-driving) earned the Z-drugs an FDA boxed warning in 2019. A 2012 observational study reported death-rate hazard ratios of 3.6 to 5.3 as doses climbed, versus matched non-users ^[37]. That study can't prove cause (it's observational and easy to confound), but the signal hasn't been retired.

If you're on a Z-drug or benzodiazepine and want off, taper under a doctor's care over 4 to 8 weeks or more. The rebound insomnia is brutal. Bring in CBT-I at the same time.

There's a better drug class: DORAs. Suvorexant, lemborexant, and daridorexant block the brain's wake-up signal (a neuropeptide called orexin) instead of cranking up the calming receptor. The phase 3 trials of daridorexant (1,854 patients across two trials) showed it worked on both waking-in-the-night and falling-asleep at 1 and 3 months ^[33]. And a pooled follow-up analysis confirmed it kept dream sleep and deep sleep intact versus baseline, unlike Z-drugs ^[40]. If a drug is needed for chronic insomnia, DORAs are the evidence-favored class. Daridorexant is recommended on the UK NHS (NICE TA922, 2023) after CBT-I.

High-dose melatonin (5 to 10 mg). As covered above: way too strong, no better than 0.3 to 1 mg for sleep quality, and it leaves you groggy with vivid dreams. In Germany, products above 1 mg start edging into drug territory.

Diphenhydramine and doxylamine (the drugstore "sleep aids"). Antihistamine sedatives. They carry an anticholinergic load and you build tolerance within a few nights. The AGS Beers list says avoid them past age 65, because of fall risk, confusion, and a link to dementia in long-term studies (a 2015 study found the heaviest anticholinergic users carried higher dementia risk over a 10-year follow-up) ^[34]. High cost, zero deep-sleep benefit.

Late nicotine and late caffeine. Both cut your deep sleep. Caffeine taken 6 hours before bed shaves more than an hour off total sleep ^[38]. Nicotine breaks up and shortens deep sleep. Cut both hours before bed.

Nightly cannabis. A hit of THC gets you to sleep faster and may briefly bump deep sleep at low doses. But tolerance to the sleep effect builds within weeks, and quitting brings rebound insomnia and vivid dreams (a dream-sleep rebound). One acute THC/CBD trial found a significant 8 percent drop in dream sleep and a 66-minute delay before dream sleep started. Nightly THC for sleep is a tolerance trap. CBD on its own, at calming doses, doesn't disrupt sleep in healthy adults.

Big late dinners, especially heavy or carb-loaded ones, blunt the overnight core-temperature drop that triggers deep sleep and spike your blood sugar overnight. A 2019 study in apnea patients found that a late dinner raised the apnea count (beta=1.28 events/h) and cut dream sleep. Eat 3 or more hours before bed, especially if you have or suspect apnea.

The honest order of operations for deep sleep. Cut alcohol, treat apnea, taper off Z-drugs and benzodiazepines, drop late caffeine and nicotine, eat earlier, ditch the anticholinergic sleep aids, and use melatonin only at low doses for body-clock reasons. None of those is a supplement. They're subtractions. And the subtractions do far more for your deep sleep than any supplement you could add.

Here's a 30-day deep-sleep plan, ordered by how much each step gives you for the effort. Start at the top and work down.

Days 1-7: Measure your baseline and remove the deep-sleep killers.

• Start a 7-day sleep diary. Track bedtime, wake time, alcohol, caffeine cutoff time, perceived quality 1 to 5.
• If you have a wearable, log nightly: total sleep time, sleep efficiency, deep-sleep estimate.
• Cut alcohol within 3 hours of bedtime. If you can stop drinking entirely for the 30-day protocol, do.
• Cut caffeine after noon if you are unsure of your CYP1A2 phenotype (the liver enzyme that decides whether you are a fast or slow caffeine metabolizer).
• Move the last meal to 3 hours or more before bed.
• Drop diphenhydramine, doxylamine, and OTC anti-histamine sleep aids.
• If you take Z-drugs or benzodiazepines nightly, talk to a clinician about tapering. Not for this protocol, but as a longer-term project.

Days 8-14: Add the temperature and light steps.

• Bedroom at 17 to 20 degrees C (62 to 68 degrees F). Cooler than you probably keep it. Get the room genuinely cold and use a duvet for warmth.
• Hot bath at 40 to 43 degrees C for 10 to 15 minutes, 60 to 120 minutes before lights-out. This is the warming protocol from the meta-analysis: about 36 percent faster sleep onset on average, as warming your skin triggers a core-temperature drop once you step out ^[15]. A hot shower works too, just less strongly.
• If your feet run cold, wear bed socks. Warm hands and feet are the signal that sleep is on its way.
• Morning outdoor light within 30 to 60 minutes of waking, for 10 to 30 minutes, no sunglasses. Aim for 1,000+ lux hitting your eyes. Cloudy days still count. This sets your body clock.
• Dim light in the evening, 10 melanopic EDI or below, for the 3 hours before bed. Dim the whole room. Phones face-down or in another room if you can.

Days 15-21: Add aerobic exercise.

• 3 to 4 aerobic sessions a week, 30 to 60 minutes each, at an effort of RPE 5 to 6 ("can talk but not sing"). Zone 2 is the practical target.
• Add 1 to 2 HIIT sessions a week (say, 10 x 1 minute at about 90 percent of your max heart rate, with 1 minute of recovery). HIIT fires off fast signals to your cell powerhouses and feeds your deep sleep.
• Finish all workouts 60 minutes or more before bed. Inside that hour, hard exercise pushes sleep back. Outside it, exercise is a net plus for deep sleep ^[18].
• Morning aerobic exercise also pulls your body clock earlier by about 0.6 h per session.

Days 22-28: Add supplements (only once the rest is in place). These give a modest boost. They don't replace the earlier steps.

• Magnesium bisglycinate at 200 to 400 mg of elemental magnesium (typical evening dose; the EU supplemental limit is 250 mg, NIH's is 350 mg). At 400 mg you're nearing the NIH limit. Don't go higher without medical advice.
• Glycine at 3 g about 30 minutes before bed (the dose from the glycine trial) ^[19]. It tastes sweet and mixes well into water or tea. Don't combine it with clozapine.
• Optional L-theanine 200 mg. Add it if racing thoughts at bedtime are your main problem.
• Skip high-dose melatonin (over 1 mg) unless you have a body-clock reason (jet lag, a delayed sleep schedule).

Talk to your doctor before starting any of these, especially if you take medication.

Days 29-30: Take stock.

• Compare week 1 (baseline) against week 4 on your sleep diary and wearable trends. Look at 7-day rolling averages, never single nights.
• Your sleep efficiency should be 85 percent or higher.
• Your self-rated sleep quality should have improved.
• If your wearable shows more deep sleep, take it as a direction, not a precise number.
• If nothing has moved, the problem is probably medical. Screen for apnea with STOP-BANG. Score 3 points or more and get a home sleep apnea test.

When to see a sleep doctor.

• Loud snoring plus witnessed breathing pauses plus daytime sleepiness, even with this plan in place, points to undiagnosed apnea. Get tested.
• Insomnia that drags on 3 months or more despite the plan: do structured CBT-I (somnio via DiGA in Germany, Sleepio in the UK, or a behavioral sleep medicine clinician).
• Acting out your dreams: see a neurologist about REM sleep behavior disorder. This is the single most urgent referral in all of sleep medicine.
• A restless-legs pattern: get your ferritin and transferrin saturation checked. The AASM 2024 guideline favors IV iron when ferritin is below 75.

The one principle to keep. Deep sleep is what fades fastest with age, and most of what you can do to protect it is about taking things away: no alcohol within 3 hours of bed, no supplements that suppress deep sleep, no warm bedroom, no blast of bright evening light, and aggressive treatment of apnea if you have it. The add-ons (cold room, warm bath, aerobic exercise, glycine, magnesium bisglycinate) give smaller but real gains on top of that foundation. Anyone selling you a one-step product for deep sleep is hyping the add-on side and ignoring the part that actually matters.