Sleep is not a single, uniform state. Across a night, the brain moves through distinct stages, light sleep, deep sleep (also called slow wave sleep), and REM sleep, each serving different restorative functions. When people ask whether CBD helps with sleep, the more precise and revealing question is what it actually does to this underlying structure, called sleep architecture, and the answer is more nuanced than wellness marketing typically suggests.
What Sleep Architecture Actually Means
Sleep architecture refers to the pattern and proportion of different sleep stages across a night, measured objectively using polysomnography (PSG), which records brain electrical activity, eye movements, and other physiological signals. The key stages are N1 and N2 (lighter sleep stages), N3 (deep sleep or slow wave sleep, associated with physical restoration), and REM sleep (associated with dreaming and cognitive functions including memory processing). Total sleep time and sleep efficiency (the proportion of time in bed actually spent asleep) are also key architecture metrics. Subjective sleep quality, how rested someone feels, does not always correlate perfectly with these objective measures, which is part of why polysomnography research is valuable: it can reveal changes that subjective reports alone might miss, in either direction.
What a 2025 Cross-Sectional Study Found in Chronic Cannabis Users
A study published in the journal SLEEP in late 2025 examined polysomnography outcomes in a sleep-clinic cohort, comparing chronic cannabis users (defined as daily use for at least a year) to non-users, adjusting for 28 demographic, lifestyle, comorbidity, medication, and sleep-related covariates. The findings for chronic users were not uniformly positive: average sleep duration was around 6 hours, below the normal 7-9 hour range, with poor sleep efficiency, prolonged time awake after initially falling asleep, significantly prolonged REM latency (time to first enter REM sleep), elevated light sleep (N1), and significantly reduced deep sleep (N3) and REM sleep. The study authors specifically cautioned against generalising these findings to healthy sleepers broadly, since this was a sleep-clinic cohort, meaning participants may have had pre-existing sleep issues that interact with cannabis use in ways that wouldn’t necessarily apply to people without sleep disorders.
What a 2025 Meta-Analysis Found Across Multiple Trials
A broader systematic review and meta-analysis of cannabis and sleep architecture, also from 2025, looked across multiple studies using objective measures (EEG, polysomnography, polygraphy, actigraphy) and found a more mixed picture than the single cohort study above. Across these studies, cannabis administration showed no consistent effect on several sleep parameters, including total sleep time, sleep onset latency, time awake after sleep onset, light sleep, deep sleep, and sleep efficiency, meaning results varied across studies rather than pointing in one consistent direction. For REM sleep specifically, the picture was clearer historically but less so currently: high-dose studies from the 1970s consistently showed THC-related REM suppression, but in modern therapeutic-dose trials, REM suppression has been reported in some studies but findings are mixed, and the overall evidence base remains thin. The same review found that withdrawal from chronic cannabis use significantly disrupts sleep, reducing total sleep time and increasing sleep onset latency, though polysomnographic evidence specifically for withdrawal effects remains scarce.
What a Specific CBD/THC Trial in Insomnia Patients Found
A trial examining a CBD/THC combination in patients with insomnia, using overnight polysomnography with high-density EEG, found that compared to placebo, the CBD/THC treatment was associated with a statistically significant decrease in total sleep time (about 24.5 minutes less), no significant change in time awake after sleep onset, and significantly reduced time spent in REM sleep (a reduction of about 8.1 percentage points) along with significantly increased REM sleep latency (taking about 65.6 minutes longer to enter REM sleep). This is a notable finding because it suggests that, at least for this CBD/THC combination in this population, the treatment did not increase total sleep time, one of the most commonly assumed benefits, while it did suppress REM sleep, consistent with THC’s known historical association with REM suppression discussed above.
The CBN Research: A Different Cannabinoid, Different Findings
As discussed in detail in our piece on CBG and CBN, a 2025 animal study using polysomnography found that CBN, a different cannabinoid from CBD, increased total sleep time in rats with effects on non-REM sleep comparable to a common pharmaceutical sleep medication, while also increasing REM sleep (notably different from the THC-associated REM suppression discussed above), though with an initial biphasic pattern of sleep suppression before the larger increase. A separate human meta-analysis found that cannabinoid formulations containing THC and CBN, but not CBD alone, were associated with improved subjective sleep quality. Taken together, this body of research suggests that different cannabinoids may have genuinely different, sometimes opposite, effects on sleep architecture specifically, and that CBD in isolation may not be the cannabinoid most associated with sleep architecture changes, despite being the most heavily marketed for sleep among CBD-specific products.
Putting This Together: What Does This Mean for CBD and Sleep?
The honest synthesis of this research is more complicated than simple marketing claims suggest. CBD alone, based on the meta-analysis discussed above, did not show a statistically significant effect on subjective sleep quality, in contrast to formulations containing THC and/or CBN. Where CBD is combined with THC (as in full spectrum products or the specific trial discussed above), the combination may not increase total sleep time and may suppress REM sleep, an effect attributable more to THC’s known mechanism than to CBD specifically. For chronic, heavy cannabis users generally (not isolating CBD specifically), polysomnography research has found architecture changes including reduced deep sleep and REM sleep, though this comes from a sleep-clinic cohort with the caveats noted. None of this means CBD has no role in sleep-related wellness; many people report subjective improvements, and CBD’s potential role in anxiety reduction (discussed in our CBD anxiety piece) could plausibly support sleep indirectly by addressing a common cause of sleep disruption, even without CBD directly altering sleep architecture in the way some marketing implies. The distinction between an indirect benefit (less anxiety, easier to fall asleep) and a direct architectural change (more deep sleep, more REM sleep) is one that current research has not clearly established for CBD specifically, in contrast to the more direct architectural findings for THC and CBN.
Frequently Asked Questions
Does CBD increase deep sleep?
Current research does not provide strong evidence that CBD specifically increases deep sleep (N3/slow wave sleep). A 2025 meta-analysis found no consistent effect of cannabis on deep sleep across studies, and a separate human meta-analysis found CBD alone did not show a statistically significant effect on subjective sleep quality, in contrast to formulations containing THC and/or CBN. The research on cannabinoids specifically increasing deep sleep is stronger for CBN (based on animal polysomnography research) than for CBD.
Does CBD suppress REM sleep like THC does?
This is less clear for CBD specifically than for THC. Historical high-dose THC research consistently showed REM suppression, and a specific CBD/THC combination trial in insomnia patients found significantly reduced REM sleep and increased REM latency, but this combination included THC, making it difficult to isolate CBD’s specific contribution. Research on CBD in isolation regarding REM sleep is more limited, and the broader 2025 meta-analysis found mixed results for REM suppression in modern, lower-dose trials generally.
Why do some people feel CBD helps them sleep if the research is mixed?
Several possibilities exist, and they’re not mutually exclusive. CBD’s potential effects on anxiety (discussed in our CBD anxiety piece) could indirectly improve sleep onset for people whose sleep difficulties are anxiety-related, without CBD directly changing sleep architecture. Subjective sleep quality and objective sleep architecture, as measured by polysomnography, don’t always align; someone could report feeling like they slept better without polysomnography showing significant architecture changes, or vice versa. Individual variation is also significant; research findings represent average effects across study populations and don’t preclude individual responses that differ from the average.
Is CBN better than CBD for sleep based on this research?
Based on the specific research discussed in this piece and in our CBG/CBN article, CBN has more directly relevant polysomnography research showing increased total sleep time with effects comparable to a pharmaceutical sleep aid (in animal studies), and a human meta-analysis found CBN (typically alongside THC) was associated with improved subjective sleep quality where CBD alone was not. This doesn’t mean CBD has no role, but the cannabinoid-specific research to date appears to favour CBN over CBD specifically for sleep-architecture-related outcomes, based on currently available studies.
Does taking CBD before bed change how long I sleep?
The CBD/THC combination trial in insomnia patients discussed in this piece found a statistically significant decrease in total sleep time compared to placebo, not an increase, though this involved a THC-containing combination rather than CBD alone. The broader meta-analysis found no consistent effect of cannabis (across various formulations) on total sleep time across studies. Based on current research, an increase in total sleep time is not a well-established effect of CBD specifically.
Should I stop taking CBD for sleep based on this research?
This research doesn’t establish that CBD is harmful for sleep; it establishes that the specific mechanism by which CBD might help (architectural changes like increased deep sleep or REM sleep) is not well-supported by current polysomnography research, in contrast to other cannabinoids like CBN. If you find CBD subjectively helpful for sleep, this research doesn’t necessarily contradict your experience, but it does suggest the mechanism may be indirect (such as through anxiety reduction) rather than a direct effect on sleep architecture itself, which is a useful distinction for setting realistic expectations rather than a reason to discontinue something you find helpful.
