Effects
THC and Sleep: Dosing, Timing and the REM Trade-Off
You take something to fall asleep faster and it works. What it does to the other 7 hours is the part worth knowing before you make it a habit. THC pushes sleep in one direction at the start of the night and the opposite direction later, and the size of that trade depends on your dose, your delivery method and how many nights in a row you have done it.
The short version
- THC shortens the time it takes to fall asleep and suppresses REM, the stage that carries most of your dreaming and sits mainly in the last third of the night.
- Start at 2.5 mg of THC, or 1 mg if you have no tolerance at all. More is not more: past roughly 10 mg most people get grogginess rather than better sleep.
- Smoked or vaporized THC peaks around 10 minutes in and fades in 2 to 3 hours, so take it 15 minutes before bed. An edible needs 30 to 120 minutes to start, so take it 90 minutes before bed.
- Tolerance to the sleep effect builds in days to weeks. Stopping brings 1 to 2 weeks of worse sleep and a stretch of unusually vivid dreams as REM rebounds.
- On a single night at ordinary amounts, alcohol does more damage to the back half of your sleep than THC does. Over months, THC's problem is tolerance and dependence rather than fragmentation.
- Nothing here treats insomnia. If you have had trouble sleeping 3 or more nights a week for 3 months, that is chronic insomnia and it has actual treatments.
What THC does to each stage of sleep
Sleep is not one block. A healthy adult night runs 4 to 6 cycles of roughly 90 minutes each, and every cycle moves through 3 non-REM stages and then REM. The mix shifts across the night: deep sleep is front-loaded into the first 2 cycles, and REM periods get longer as morning approaches.
N1 is the drift-off stage, 1 to 7 minutes of light sleep you can be pulled out of by a text message. N2 is where you spend most of the night, marked by sleep spindles and a falling body temperature. N3 is slow-wave sleep, the hardest stage to wake from and the one tied to physical recovery. REM is the paralysed, dreaming stage, and it does the bulk of emotional and memory processing.
THC touches these unevenly. The effect on falling asleep is real and quick, the effect on REM is the most consistent finding in the literature, and the effect on deep sleep is genuinely unclear.
| Stage | Share of a typical adult night | What it does | What THC does to it |
|---|---|---|---|
| N1, light | 2–5% | Transition into sleep, easily interrupted | Shortens the run-up, so you cross over sooner |
| N2 | 45–55% | The bulk of the night, spindles, temperature drop | Little consistent change |
| N3, slow-wave | 13–23% | Physical recovery, front-loaded before 2am | Mixed, possibly increased acutely, reduced with regular use |
| REM | 20–25% | Dreaming, memory and emotional processing, most of it after 3am | Delayed and shortened, and the effect scales with dose |
Feinberg and colleagues measured this in a lab as far back as 1975 in Clinical Pharmacology and Therapeutics, giving high oral doses of THC and recording REM suppression on the electroencephalogram. The 50 years of work since have complicated the deep-sleep picture and left the REM finding standing.
The practical read: THC buys you the front door of the night and charges you at the back door.
Faster to sleep, worse in the second half
The first half of a THC night usually looks good. Sleep latency drops, often by 15 to 30 minutes at low doses, and the subjective experience is exactly what people describe when they say it works.
The second half is where the cost lands. REM is pushed later and cut shorter, so the last 2 cycles of the night carry less of it. At higher doses, more people report waking in the small hours and struggling to get back down, and morning grogginess is common enough that regular users treat it as normal.
The flip side of suppression is rebound. Cut REM for weeks and the pressure to have it builds, so when the THC stops the REM comes back all at once, longer and more intense than baseline. That is the mechanism behind the famously strange dreams people get on quitting.
None of this makes THC useless for sleep. It makes it a sedative with an architectural cost, which is a different thing from a treatment.
Numbers
- 48%: share of medical cannabis patients in survey data who report using it for insomnia.
- 77% of men and 64% of women who use cannabis say they have tried it specifically for sleep.
- 90 minutes: the length of one sleep cycle, so a 7.5-hour night is roughly 5 of them.
- 20% to 25%: how much of a healthy adult night is REM, with the longest REM period usually arriving after 4am.
- Close to 90% of people who drink in the evening report at least 1 sleep problem.
- 15 to 30 minutes: typical reduction in how long it takes to fall asleep at a low THC dose.
How much THC to take before bed
Start at 2.5 mg. If you have never used cannabis, or you are over 60, or you are taking anything else sedating, start at 1 mg and hold there for 3 nights before you change anything.
The relationship between dose and sleep is not a straight line. Low doses sedate. Higher doses in the same person produce more anxiety, more heart rate, more middle-of-the-night waking and more next-day flatness, and the sleep does not get better in exchange. If 5 mg is not working, the answer is usually a different product or a tolerance break, not 15 mg.
| Nightly THC dose | Who it fits | What usually happens |
|---|---|---|
| 1–2.5 mg | No tolerance, first time, over 60, on other sedatives | Mild drowsiness, minimal next-day cost |
| 2.5–5 mg | Occasional users, 1 to 3 nights a week | Reliable drowsiness for most people |
| 5–10 mg | Regular users and most medical patients | The common maintenance range |
| 10–20 mg | Daily high-tolerance users | Diminishing returns, morning grogginess is normal |
| 20 mg+ | Very high tolerance | Rarely sleeps you better than 10 mg did before tolerance |
CBD dosing runs on a different scale entirely. Doses in the 5 to 25 mg range are commonly used alongside THC to soften the head effect, while the sleep research that found anything useful used far higher amounts, in the region of 160 mg in a single dose. A 10 mg CBD gummy is not a sleeping pill and was never tested as one.
Anchor everything to a single variable at a time. Change the dose or the product or the timing, never 2 at once, and give each version 3 nights before you judge it.
When to take it: 15 minutes for a joint, 90 for a gummy
Timing is the part people get wrong most often, and it is the cheapest thing to fix. Inhaled THC is in your blood within minutes. Swallowed THC has to pass through your stomach and your liver first.
| Method | Onset | Peak | Useful duration | Take it |
|---|---|---|---|---|
| Smoked flower | 1–5 min | ~10 min | 2–3 hours | 15 minutes before bed |
| Vaporized flower or cartridge | 1–5 min | ~10 min | 2–3 hours | 15 minutes before bed |
| Tincture held under the tongue 60–90 sec | 15–30 min | 60–90 min | 4–6 hours | 45 minutes before bed |
| Gummy, chocolate or capsule | 30–120 min | 2–4 hours | 6–8 hours | 90 minutes before bed |
That duration column is the other half of the timing decision. A joint at 11pm has largely stopped working by 2am, which is fine if your problem is falling asleep and useless if your problem is waking at 3am. A 10 mg gummy at 9:30pm is still active at 4am, which is what you want for maintenance and exactly what causes the morning fog.
Eat something small with an edible. Taking it on a completely empty stomach makes the onset faster and less predictable, and a fatty snack pushes absorption toward the reliable end.
Worked example
Target bedtime 11:00pm, alarm 6:30am. Occasional user, 3 nights a week, problem is a racing head at lights-out.
- 9:30pm: 1 gummy, 2.5 mg THC and 2.5 mg CBD. Phone goes on the charger across the room.
- 10:05pm: nothing yet. This is the exact moment people take a second one, and it is the mistake.
- 10:40pm: first real drowsiness, reading gets hard.
- 11:00pm: lights out. Asleep around 11:15 instead of the usual 11:50.
- 3:50am: brief wake, back down inside 10 minutes.
- 6:30am: alarm. Slightly flat for 30 minutes, no dream recall from the night.
Cost of the night: roughly $2 of product from a $20 10-pack. If night 3 looks the same, that is the working dose. If it stops working by night 10, that is tolerance, not a reason to move to 5 mg.
Smoked, vaped, edible or tincture changes the drug
Delivery is not a preference, it is a pharmacological difference. How much THC actually reaches your bloodstream swings by an order of magnitude between methods, and swallowed THC arrives as a different molecule.
Huestis reported in Chemistry and Biodiversity in 2007 that smoked THC bioavailability ranges from about 2% to 56% depending on how a person inhales, while oral bioavailability sits at roughly 4% to 20%. That range explains why 10 mg swallowed and 10 mg smoked are not the same experience for anyone.
| Method | THC reaching your bloodstream | Onset | Duration | What it suits |
|---|---|---|---|---|
| Smoked | 2–56%, depending on inhalation depth and hold | Minutes | 2–3 hours | Falling asleep |
| Vaporized | Similar range, usually at the higher end | Minutes | 2–3 hours | Falling asleep with less throat irritation |
| Oral, gummy or capsule | 4–20% | 30–120 min | 6–8 hours | Staying asleep |
| Sublingual tincture | Between the two | 15–30 min | 4–6 hours | Dose control, because you can measure it |
When you swallow THC, your liver converts a large share of it to 11-hydroxy-THC before it ever reaches your brain. That metabolite is more potent than THC itself and lasts longer, which is why edibles feel heavier, hit later and outstay their welcome.
Tinctures are underrated for sleep specifically. A 1 ml dropper marked in quarters lets you run 2.5 mg, then 3.75 mg, then 5 mg over successive weeks, which is a resolution that gummies cannot give you without cutting sticky things in half with a knife.
THC, CBD and CBN: the ratio is the lever
THC is the sedating and intoxicating compound. CBD is neither, and at the doses in most retail products it mostly rounds off the edges of the THC rather than doing much on its own. Adding CBD to a THC dose tends to make the experience less anxious and less foggy without removing the drowsiness.
The ratio is how a menu expresses that balance. A 1:1 product carries equal milligrams of each, typically 5 mg and 5 mg per piece, and it is the standard recommendation for anyone who finds straight THC too much in the head. A 20:1 product is 20 parts CBD to 1 part THC, so a 100 mg piece carries 5 mg of THC, and it barely registers as intoxicating.
Read the milligrams, not the ratio. Some states print CBD first and some print THC first, and a 20:1 label tells you nothing until you find the 2 numbers it refers to.
CBN is the third name on sleep products. It forms as THC ages and oxidises, and it has a long-standing reputation as the sedating cannabinoid. That reputation rests on very thin evidence: a handful of small studies, mostly from the 1970s, mostly in combination with THC rather than alone. Retail CBN blends usually pair 2 to 5 mg of CBN with 2 to 5 mg of THC, and whatever sedation you feel from one is likely coming from the THC.
At the counter
Lead with the problem, not the product. "I fall asleep fine but I wake at 3am" and "I lie there for an hour" get you 2 completely different recommendations, and most budtenders will ask if you do not say.
Then work through these:
- "What do you have at 5 mg of THC per piece or lower?"
- "Anything 1:1 with CBD, in a gummy or a tincture?"
- "Do you carry a CBN blend, and what is the actual THC number on it?"
- "Is there a tincture with a marked dropper so I can take a quarter of a dose?"
If the answer to the first question is a shrug toward the 100 mg package, ask a second person. Every mature menu carries low-dose stock.
Reading a sleep product label, and what it costs
The biggest number on the front of the package is almost always the total for the whole package, and it is there for compliance rather than for you. The number that matters is the per-serving line, usually printed smaller on the back next to the piece count.
A standard retail gummy is 10 mg of THC in a 100 mg package of 10 pieces, which is a legal serving convention rather than a dose recommendation. Low-dose products are a shelf category rather than a special order now, and a sleep shelf in Grand Rapids will usually hold 2.5 mg and 5 mg pieces next to the 10 mg standard.
| Product | Typical strength | Typical shelf price |
|---|---|---|
| 10-pack gummies, 10 mg each | 100 mg total THC | $15–$35 |
| Low-dose gummies, 2.5–5 mg each | 50–100 mg total | $18–$35 |
| CBN sleep blend, 20 pieces | 5 mg THC + 5 mg CBN each | $25–$45 |
| 1:1 tincture, 30 ml | 300 mg THC + 300 mg CBD | $35–$70 |
| Eighth of flower | 18–28% THC | $25–$60 |
Tax moves the real number more than the shelf tag does. A $30 tincture in a Massachusetts store lands near $36 once the 10.75% excise, the 6.25% sales tax and a local option of up to 3% are added at the register, so ask for the out-the-door price before you commit.
Selection is not even across states. A menu in Kansas City may list 2 or 3 CBN products where a long-running market lists 20, so the ratio you can actually buy is often set by the shelf rather than by what you decided you wanted.
Weed vs alcohol for sleep, and the night you have both
Both are sedatives, both get you to sleep faster, and both are worse for the night than doing nothing. They are not equivalent, and the differences are specific.
| THC | Alcohol | |
|---|---|---|
| Time to fall asleep | Shortened | Shortened |
| Deep sleep, N3 | Mixed, falls with regular use | Rises in the first half |
| REM | Suppressed across the night | Suppressed early, rebounds hard after about 4 hours |
| Second half of the night | Fragmented at higher doses | Reliably fragmented, with early final waking |
| Wakes you to urinate | No | Yes, alcohol is a diuretic |
| Breathing and snoring | Roughly neutral | Worse, it relaxes the upper airway |
| Tolerance to the sleep effect | Days to weeks | About 3 nights |
| On stopping | 1–2 weeks of rebound insomnia and vivid dreams | A few disrupted nights |
| Next morning | Grogginess for some | Grogginess, dehydration, measurable performance loss |
| Cut-off before bed | 15 min smoked, 90 min for an edible | At least 3 hours, and 1 hour per standard drink |
For one night at ordinary amounts, alcohol is the worse choice. It creates the sharpest second-half fragmentation, it adds bathroom trips, and it makes snoring and breathing interruptions worse rather than better. A standard US drink is 14 g of pure alcohol, which is 12 oz of 5% beer, 5 oz of 12% wine or 1.5 oz of spirits, and your body clears roughly 1 of those per hour.
Over months, the comparison flips direction. Nightly THC builds tolerance quickly, and the sleep you are chasing becomes the sleep you cannot get without it. Alcohol has its own dependence problem, but the sleep-specific tolerance arrives inside a week and the drinking dose usually escalates for reasons that have nothing to do with bedtime.
The night you have both is worse than either. Lukas and Orozco showed in Drug and Alcohol Dependence in 2001 that drinking before smoking substantially raises peak plasma THC levels, so the same joint hits harder after 2 beers than it does sober. Add alcohol's airway relaxation and its second-half fragmentation to THC's REM suppression and you get the least restorative version of both drugs. If you are drinking anyway, stop 3 hours before bed and skip the THC, rather than stacking them.
Tolerance, quitting, and why the dreams come back
Tolerance to the sleep effect of THC arrives faster than tolerance to almost anything else it does. Nightly users commonly report the same dose stops working within 2 to 4 weeks, and the honest fix is a break rather than an increase.
Stopping is not comfortable. Sleep difficulty and disturbing dreams are both recognised features of cannabis withdrawal, and for daily users the worst of it lands in the first week. Sleep onset can stretch past an hour on the first 2 nights, and irritability, appetite loss and sweating often ride along with it.
| Days off | What most daily users report |
|---|---|
| 1–2 | Hardest nights, long sleep onset, restlessness |
| 2–6 | Peak withdrawal, irritability, low appetite |
| 7–14 | Sleep starting to normalise, dreams becoming intense |
| 14–21 | Vivid and often unpleasant dreams as REM rebounds |
| 21+ | Sleep architecture largely back to baseline for most people |
The dreams are the clearest evidence of what the THC was doing. Weeks of REM suppression build pressure, and when the suppression lifts the REM comes back longer and denser than normal. Unpleasant is common. It is also temporary, and it is a sign the system is re-regulating rather than a sign something is wrong.
If a 2-day break sounds impossible, that is information about dependence rather than about your sleep. A 48-hour break is usually enough to notice a difference in dose response; 2 weeks resets most of it.
Numbers
- 2 to 4 weeks: how long nightly use typically takes to blunt the sleep effect at a fixed dose.
- 48 hours: the shortest break that noticeably restores dose sensitivity.
- 2 weeks: what most people need for a full reset before restarting at a lower dose.
- 3 nights: how long to hold any new dose before deciding whether it works.
- 1 to 2 weeks: the typical span of rebound insomnia after stopping daily use.
Conditions where the evidence is strongest and weakest
Most of the research on cannabis and sleep is in people who have something else wrong. That matters, because the strongest results come from conditions where sleep improves as a side effect of something else improving.
| Condition | Strength of the evidence | What it looks like in practice |
|---|---|---|
| Chronic pain | Moderate, and mostly indirect | Sleep improves because pain drops, not because sleep itself changes |
| PTSD nightmares | Small controlled trials, promising | Nabilone, a synthetic THC analogue, reduced nightmare frequency in military trials |
| Primary insomnia | Weak, short-term only | Faster sleep onset for a few weeks, then tolerance |
| Restless legs syndrome | Case reports only | Small case series report relief, with no controlled trials behind them |
| Bedtime anxiety | Split, dose dependent | Low doses calm, higher doses provoke anxiety in the same person |
| Obstructive sleep apnea | Evidence points against it | Not recommended, and it can mask a condition that needs treating |
Sleep apnea is the one to be careful about. The American Academy of Sleep Medicine issued a position statement in 2018 in the Journal of Clinical Sleep Medicine advising against medical cannabis and synthetic cannabinoids for obstructive sleep apnea, on the grounds of unreliable delivery and unproven safety and effectiveness. Sedating yourself through a breathing disorder is the wrong direction, and loud snoring plus daytime sleepiness is a reason to get tested rather than a reason to buy a gummy.
Anxiety deserves its own note because it cuts both ways in the same person. The dose that quiets a racing head at 2.5 mg can produce a racing heart and a racing head at 10 mg, and someone whose bedtime problem is anxiety has a narrower useful window than someone whose problem is pain.
Who should not take THC for sleep
Some of these are conditional and some are not.
- Anyone with a personal or family history of psychosis or schizophrenia. THC can precipitate psychotic symptoms, and the risk is dose dependent. This one is not a maybe.
- People taking sleeping pills or benzodiazepines. Zolpidem, temazepam and their relatives are already central nervous system depressants, and stacking THC on top compounds the sedation and the next-day impairment. Talk to whoever prescribed them first.
- Anyone who has been drinking that evening. Alcohol raises the THC level you actually get from the same amount, so the dose you know becomes a dose you do not.
- People who are pregnant or breastfeeding. THC crosses the placenta and appears in breast milk.
- Anyone with untreated or suspected sleep apnea. Sedatives of any kind work against you here.
- People with unstable heart conditions. THC raises heart rate acutely, sometimes by 20 to 50 beats per minute, and that spike arrives right when you are lying down.
- Anyone driving in the early morning. An edible taken at 10pm can still be measurably impairing at 6am, and driving impaired is illegal in every state regardless of where you bought the product.
The threshold for seeing a doctor is concrete. Trouble falling asleep or staying asleep 3 or more nights a week, for 3 months or longer, is chronic insomnia by clinical definition. It has a first-line treatment, cognitive behavioural therapy for insomnia, which outperforms sedatives over the long run and is not something a dispensary can sell you. Loud snoring, witnessed breathing pauses or falling asleep during the day are separate signals and warrant a sleep study rather than a stronger product.
At the counter
When the product you have been using stops working:
- Do not double the milligrams. Ask instead for the same THC number in a different format, because a tincture at 5 mg behaves differently from a gummy at 5 mg.
- Say plainly: "I have been taking 10 mg nightly for 2 months and it stopped working." A good budtender will tell you to take a break. That is the correct answer, and it costs them the sale.
- If you are taking a break, buy nothing for 2 weeks. Buying a smaller package to ration is how a break turns into a taper that never ends.
- Coming back, restart at half of what you were taking. Tolerance drops further than people expect, and the old dose will feel like too much.
FAQ
Does weed affect REM sleep?
Yes, and it is the most consistent finding in the research. THC delays the first REM period and shortens total REM time, with the effect scaling with dose, which is why regular users often report they have stopped dreaming. Suppressing REM night after night builds pressure that releases as unusually vivid dreams when you stop.
How much THC should I take before bed?
Start at 2.5 mg, or 1 mg if you have no tolerance, are over 60, or take anything else sedating. Hold the dose for 3 nights before judging it, because the first night of anything is not representative. Past roughly 10 mg most people get grogginess and middle-of-the-night waking rather than better sleep.
How long before bed should you take an edible?
90 minutes. Edibles start working somewhere between 30 and 120 minutes depending on your stomach contents and metabolism, and they peak at 2 to 4 hours. Taking one at bedtime means it peaks at 1am, which is the version people describe as sleeping badly on an edible.
Why do I get vivid dreams when I stop smoking weed?
Because REM rebounds. THC suppresses REM sleep while you are using it, so the need for it accumulates, and when the suppression lifts your brain runs longer and denser REM periods than baseline. The effect usually peaks in the second and third weeks off and settles within about a month.
Is weed or alcohol worse for sleep?
For a single night at ordinary amounts, alcohol. It fragments the back half of the night more reliably, it wakes you up to urinate, and it relaxes the upper airway so snoring and breathing interruptions get worse. THC's costs are more about REM suppression and how quickly tolerance builds over weeks.
Is CBD or THC better for sleep?
THC is the sedating one, and CBD at retail doses mostly softens the anxiety and fog that THC can bring rather than causing sleep itself. The research that found CBD useful for sleep used doses around 160 mg, far above the 10 to 25 mg in most products. A 1:1 combination is the practical middle ground for people who find straight THC too heady.
Can cannabis cure insomnia?
No. It can shorten how long you take to fall asleep for a few weeks, and then tolerance erodes that benefit while the REM suppression stays. Chronic insomnia, meaning trouble sleeping 3 or more nights a week for 3 months, responds better to cognitive behavioural therapy for insomnia than to any sedative, cannabis included.
Sources
- Feinberg I, Jones R, Walker JM, Cavness C, March J. "Effects of high dosage delta-9-tetrahydrocannabinol on sleep patterns in man." Clinical Pharmacology and Therapeutics 17(4): 458-466, 1975.
- Huestis MA. "Human cannabinoid pharmacokinetics." Chemistry and Biodiversity 4(8): 1770-1804, 2007.
- Ramar K, Rosen IM, Kirsch DB, et al. "Medical cannabis and the treatment of obstructive sleep apnea: an American Academy of Sleep Medicine position statement." Journal of Clinical Sleep Medicine 14(4): 679-681, 2018.
- Lukas SE, Orozco S. "Ethanol increases plasma delta-9-tetrahydrocannabinol (THC) levels and subjective effects after marihuana smoking in human volunteers." Drug and Alcohol Dependence 64(2): 143-149, 2001.