Cannabis and Insomnia: What the Research Actually Shows

Someone on my waiting list — I’ll call her Dana — had been using cannabis to sleep for four years. It worked beautifully at first. Fifteen minutes after a small vape hit, she was out. Then it took two hits. Then a higher-THC product. Then two products. By the time she reached me, she was using cannabis every night, sleeping five fragmented hours, and spending her days in what she described as ‘a fog that never quite lifts.’ She didn’t have a sleep problem anymore. She had two.

Dana is not unusual. Surveys consistently find that up to 80% of cannabis users in North America cite sleep improvement and relaxation as primary reasons for use. Cannabis is widely available, legal in British Columbia, and culturally positioned as a natural alternative to sleeping pills. For people who have already been through the pharmaceutical carousel — zopiclone, Ativan, melatonin at escalating doses — it can feel like a final option that actually works.

The neuroscience tells a more complicated story. Cannabis does something to sleep, but what it does depends on what compound you’re using, how often, at what dose, and for how long. Short-term use and long-term use produce measurably different effects. And the gap between what people report — feeling like they sleep better — and what their brains are actually doing is one of the more interesting puzzles in sleep medicine right now.

This article is an overview of what the research shows. It won’t tell you whether to use cannabis. That depends on your situation, your history, and your goals — and those are questions worth discussing with whoever is treating your sleep. What it will do is give you a clearer picture of the mechanism, so that whatever choice you make, you’re making it with accurate information.

How Cannabis Interacts With Your Sleep Biology

Your brain has its own internal cannabis system. It’s called the endocannabinoid system (ECS), and it plays a direct role in regulating when you fall asleep, how deeply you sleep, and when you wake up. The two main chemicals your brain produces — anandamide and 2-AG — interact with the same receptors that THC and CBD target. Phytocannabinoids, the compounds in cannabis, essentially exploit this existing architecture, but through mechanisms that differ meaningfully from what the brain’s own system does.

Anandamide is the endocannabinoid most closely studied for its role in sleep. When administered exogenously in animal studies, it increases slow-wave sleep and NREM sleep while inhibiting wakefulness — partly by raising adenosine levels in the basal forebrain, which builds sleep pressure. Its circadian profile is more complex: levels are actually higher after waking than before sleep, which suggests anandamide is doing more than simply signalling time for bed. The sleep-promoting effects appear to depend on where and when it acts in the brain, not just its presence. This is one reason why the relationship between cannabis and sleep is harder to read than it might initially seem.

THC, the compound responsible for the psychoactive effects of cannabis, acts as a partial agonist at CB1 receptors. At low to moderate doses it reduces the time it takes to fall asleep and can increase total sleep time — partly by suppressing orexin, a neuropeptide that promotes wakefulness. That’s the mechanism behind the familiar cannabis-induced drowsiness.

CBD works quite differently. It has very low binding affinity for CB1 receptors and appears to modulate sleep through other pathways, including the serotonin system. At high doses — around 150 mg in clinical trials — CBD shows therapeutic potential for people with moderate-to-severe insomnia. The dose-response relationship is not linear, and the direction of effect appears to depend on dose and timing; the common assumption that CBD is straightforwardly sedating is not well supported by current evidence.

So far, the mechanism sounds straightforwardly helpful. Here’s where it gets more complicated.

What’s Actually Happening to Sleep Architecture

People who use cannabis before bed reliably report sleeping better. What researchers see when they measure the actual structure of that sleep is a different picture.

A 2025 systematic review and meta-analysis pooled data from eighteen clinical trials and found that cannabis administration did not consistently alter total sleep time, sleep onset latency, time awake after initially falling asleep, sleep efficiency, or the proportional percentages of REM and deep slow-wave sleep. The authors concluded that the subjective sense of sleeping better may largely reflect the avoidance of withdrawal symptoms in regular users — not an objective improvement in sleep architecture.

A separate home sleep test study compared 89 adults who used cannabis close to bedtime with 88 who didn’t. The cannabis users fell asleep faster — 8.3 minutes versus 12.7 minutes on average. But they also spent significantly more time awake after initially falling asleep (60.5 minutes versus 45.8 minutes) and spent more of their night in light, non-restorative sleep. Total sleep time was effectively the same between groups. The cannabis users weren’t sleeping more — they were sleeping more shallowly, in shorter consolidated stretches, with a faster transition to unconsciousness at the start.

That pattern — easier in, worse throughout — is worth sitting with. For someone with sleep onset insomnia, the appeal of cannabis is understandable. The number that matters most to them is how long it takes to fall asleep, and cannabis improves that number. But if sleep maintenance is also a problem, cannabis may be making that worse while the person’s subjective sense of the night focuses on the part that improved.

There’s also the REM question. Long-standing clinical lore held that cannabis suppresses REM sleep. The 2025 meta-analysis found mixed results — the REM suppression effect is less consistent than previously believed, concentrated mainly in older studies that used high THC doses in controlled laboratory conditions. More recent data from a 60-person in-home EEG study tracking 339 nights found that regular cannabis use was associated with less REM sleep (roughly 8.5 percentage points less) and reduced total sleep time. The same study found increased slow-wave sleep in the same group — which sounds like a benefit, but the overall picture was of less total sleep, less dreaming, and shorter consolidated sleep periods.

Why It Stops Working: Tolerance and Receptor Downregulation

Dana’s escalating pattern — one hit, then two, then a stronger product — has a neurological explanation that follows directly from how THC works at the receptor level.

When THC binds to CB1 receptors repeatedly, the brain mounts a two-stage adaptive response. First, it chemically modifies the receptors so they no longer transmit the signal as strongly — a process called desensitization. If exposure continues, the brain physically pulls those receptors off the cell surface and stores them inside the cell, reducing their overall number. This is called downregulation.

PET imaging studies have quantified this. Under daily cannabis use, CB1 receptor density falls by roughly 20–30% within two weeks. The regions most affected are the prefrontal cortex, cingulate cortex, and hippocampus — areas involved in executive function, emotional regulation, and memory. The basal ganglia are relatively spared. The cognitive fog that many regular cannabis users describe has a measurable neurological basis.

Clinically, this plays out as escalating dose. The person finds they need more cannabis to get the same effect. At some point, they’re using enough that the pharmacological cost — morning grogginess, daytime impairment, fragmented sleep architecture — starts to exceed the benefit. Some people reach this point within weeks. Others take months. The trajectory varies, but the direction doesn’t.

The downregulation reverses after stopping. The hippocampus recovers to near-baseline within about two weeks of abstinence. The prefrontal cortex takes longer — closer to four weeks. Receptor density begins to recover within 48 hours of cessation, with the most significant gains in the first three to seven days.

The recovery window matters, because it frames the withdrawal problem.

What Happens When You Stop

Cannabis withdrawal syndrome is formally recognized in the DSM-5. Among its features, sleep-related disturbances are the most common, most severe, and most likely to drive relapse.

When someone who has been using cannabis nightly stops, the endocannabinoid system — whose receptors have been downregulated — is suddenly underactivated. The brain overshoots in the direction of arousal. Sleep onset latency increases sharply, time awake in the night increases sharply, and REM sleep surges — what sleep researchers call REM rebound. The vivid, often disturbing dreams that accompany cannabis cessation are a direct consequence of that rebound.

Polysomnography during acute abstinence shows reduced total sleep time, lower sleep efficiency, elevated wake after sleep onset, and a substantial increase in periodic limb movements — the last of which correlates with both the severity and the duration of prior use.

This is the trap. Someone uses cannabis because they can’t sleep. Over time, their sleep depends on cannabis to remain even functional. When they try to stop, their sleep gets dramatically worse — often worse than before they started. The natural interpretation is that they need the cannabis, that their insomnia is simply that severe. What’s actually happening is that their nervous system is going through withdrawal. The insomnia is real, but its cause is now the treatment.

Dana recognized this immediately when I described it. She’d tried stopping twice and both times her sleep fell apart within three days. She’d concluded her insomnia was too severe to manage without cannabis. It probably wasn’t. It had just become, in part, a cannabis withdrawal problem.

Where Cannabis Does Have Clinical Evidence

The picture isn’t only negative. There are populations and situations where cannabinoids appear to offer genuine benefit, and where the clinical calculus looks different.

Chronic pain and sleep. A McMaster University meta-analysis pooled 39 randomized controlled trials involving 5,100 patients and found that non-inhaled medical cannabis produced a small but statistically significant improvement in sleep quality for people with chronic pain. The number needed to treat was 13 — meaning one in every 13 patients achieved a clinically meaningful improvement. For chronic non-cancer pain specifically, one in five patients experienced reduced sleep disturbance. These are modest numbers, but chronic pain is a situation where insomnia has a specific, persistent driver that behavioural interventions alone may not fully address.

PTSD and nightmares. The REM-suppressing properties of THC — which are a liability in general insomnia — are a potential asset in PTSD, where the problem is specifically the content and intensity of REM sleep. Synthetic CB1 agonists like nabilone have shown meaningful results in controlled trials: in one double-blind crossover study, 44% of participants on nabilone achieved complete resolution of recurring nightmares at seven weeks, compared to 0% on placebo. Whole-plant cannabis trials have been less conclusive, with high placebo response rates complicating interpretation.

Specific pharmaceutical formulations. Several pharmaceutical-grade cannabinoid preparations have now completed randomized controlled trials for insomnia specifically. ZTL-101, a sublingual extract combining THC, CBD, and CBN, showed significant improvements in insomnia severity, sleep onset, time awake in the night, and total sleep time in a double-blind crossover trial. Entoura 10:15, an oral oil combining 10 mg THC with 15 mg CBD, found that 60% of participants no longer met clinical criteria for insomnia after two weeks of use. These are small trials, but they represent a different category of evidence than most of the self-treatment literature.

The pattern across all of these is specificity: cannabis-based interventions show the clearest benefit when they’re formulated with documented cannabinoid profiles, used at low doses, and matched to the specific insomnia phenotype being treated — sleep onset versus sleep maintenance — rather than used as a general sedative.

The Access Gap 

One finding from the epidemiological literature deserves more attention than it gets.

A study examining self-treatment patterns among people who had recently experienced involuntary job loss found that marijuana was the single most common self-administered treatment for sleep and relaxation problems. Its prevalence of use was comparable to all evidence-based treatments combined. Only 0.1% of participants reported receiving CBT-I — cognitive behavioural therapy for insomnia, which is the first-line clinical recommendation for chronic insomnia in guidelines across North America.

Read that number again. One in a thousand.

People aren’t choosing cannabis over CBT-I because they’ve tried both and prefer cannabis. They’re choosing cannabis because CBT-I is largely inaccessible — underfunded, underprescribed, and almost entirely absent from standard primary care pathways. Cannabis is at the dispensary. CBT-I requires knowing it exists, finding a provider, navigating waitlists, and often paying out of pocket.

This matters for how we think about cannabis use in insomnia. It’s partly a treatment choice. It’s partly a treatment gap filling itself with whatever’s available.

What This Means Practically

If you’re currently using cannabis to sleep, the research suggests a few things worth knowing.

Daily use is where the risk concentrates. The tolerance and withdrawal dynamics described above are heavily dose- and frequency-dependent. The harm reduction literature suggests that intermittent use — rather than nightly — may help preserve receptor sensitivity over time, though this guidance reflects clinical reasoning rather than controlled trial evidence. Regular periods of abstinence appear to support receptor recovery, with meaningful gains beginning within the first three to seven days.

The route of delivery matters for the insomnia type. Inhaled cannabis reaches peak levels in 5–10 minutes but wears off in 2–4 hours, making it more suited to sleep onset difficulties. Oral formulations take 45–90 minutes to peak but last 6–8 hours, which is a better match for sleep maintenance problems — waking in the night and being unable to return to sleep.

High-THC products carry the most risk for the tolerance and REM disruption patterns described above. The controlled trials with the most promising outcomes have used low-THC formulations combined with CBD — which appears to moderate some of THC’s receptor effects — rather than high-THC products used alone.

If you think your sleep has become dependent on cannabis in the way Dana’s had, that’s worth treating directly. The withdrawal-induced insomnia in the early days of stopping is real, but it’s time-limited. Addressing the underlying sleep patterns — the conditioned arousal, the hypervigilance around sleep, the behaviours that perpetuate insomnia independent of any substance — is where the longer-term work happens.

A Final Note

The question I get most often about cannabis and sleep isn’t ‘does it work?’ It’s ‘why did it stop working?’ The mechanism in this article answers that question directly. The brain adapted to the thing that was regulating it. That’s not a failure of the person. It’s a predictable neurological response to a daily pharmacological input.

Dana’s sleep is considerably better now. Not because she found the right cannabis product, but because we addressed the underlying arousal patterns that had been driving her insomnia since long before she started using cannabis. The cannabis had been managing symptoms. CBT-I addressed what was generating them.

That’s not a pitch for CBT-I over cannabis. It’s an observation about what each one is actually doing. Understanding the difference is where better sleep decisions begin.

If you’re trying to understand whether your insomnia is driven by hyperarousal, sleep behaviour patterns, or something else, the free consultation is a good starting point. We can usually get a clear picture of what’s going on in a single conversation.