CBD and vivid dreams: what 7 studies actually measured

According to the Inserm sleep dossier, a night is made of 3 to 6 cycles lasting 60 to 120 minutes each. Every cycle goes from light slow-wave sleep to deep slow-wave sleep, then to REM sleep. In REM, the eyes move rapidly and muscle tone almost disappears, so the body lies still while the brain is highly active.

For Inserm, REM is the stage that favours dreaming most, above all the dreams remembered on waking. The US National Institute of Neurological Disorders and Stroke (NINDS) puts the first REM episode at about 90 minutes after falling asleep.

Dreaming more and remembering more dreams are therefore two separate things. If the end of the night is more broken, with brief awakenings, more dreams can stay in memory while the amount of REM sleep stays the same. Many accounts about CBD report a remembered dream, with no measurement behind it.

Only a handful of papers deal with CBD and dreams, and the table below lists them by date. An effect can be attributed to the product only in the studies that used a placebo.

Polysomnography records the activity of the brain, eyes and muscles through the night, and no other method separates sleep stages precisely. A wrist device can only estimate them from movement and pulse.

The most direct study came out in 2018 in Frontiers in Pharmacology, and its title gives the result: "No Acute Effects of Cannabidiol on the Sleep-Wake Cycle of Healthy Subjects". In Brazil, Linares and colleagues recruited 27 volunteers. Of these, 26 finished the study: 12 men and 14 women, 29 years old on average.

Every participant spent two double-blind nights in the lab, one after 300 mg of CBD and one after a placebo, taken 30 minutes before 8 hours of recording. Sleep onset, sleep efficiency and the share of each stage, REM included, stayed the same. On waking, memory and mood tests showed no difference either.

According to the authors, CBD does not appear to disturb the sleep cycle, whereas benzodiazepines and serotonergic antidepressants do. A single night and a single dose in 26 people cannot tell us anything about repeated use. For cannabidiol on its own, this remains the cleanest controlled measurement available.

Two Brazilian pharmacologists, Carlini and Cunha, published the only data on dream recall under CBD, in the Journal of Clinical Pharmacology in 1981. Their paper reviews the lab's work since 1972. In it, 15 volunteers with insomnia received 40, 80 or 160 mg of CBD, a placebo or a reference sleeping pill.

Participants reported longer sleep only at 160 mg. At all three doses, though, they remembered dreams less often. The usual story online says the opposite, and no modern study has tried to repeat the finding.

The same caution applies in both directions. With 15 people, questionnaires and a method dating back forty-five years, the result cannot prove that CBD erases dreams. What it does show is that the opposite idea has no support.

At the end of 2025, Velzeboer and colleagues published a meta-analysis in Sleep Medicine Reviews. It pooled 18 studies on cannabis and sleep architecture, 9 of them usable for statistics. Cannabis use has no consistent effect on sleep duration, sleep onset, wake time, efficiency or stages. The large falls in REM sleep come mostly from small, older studies with high THC doses.

Withdrawal after regular use, by contrast, shows a clear pattern: shorter nights, longer time to fall asleep and a REM rebound. The authors also point to a gap between how well users believe they sleep and what the recordings show.

Most of the accounts fit this. A person who replaces joints with CBD flowers adds cannabidiol and drops THC on the same day. The intense dreams of the first nights most likely correspond to the rebound found in this meta-analysis. That transition is covered in our guide to switching from THC to CBD.

A trial by Suraev and colleagues, published in the Journal of Sleep Research in 2026, agrees. In 20 people with insomnia, 10 mg of THC taken with 200 mg of CBD reduced REM sleep by 33.9 minutes and pushed its onset back by 65.6 minutes. The authors put this down mainly to THC.

People with REM sleep behaviour disorder keep their muscle tone during REM and act out their dreams: they talk, lash out or fall out of bed. A series of 4 cases published in 2014 described fewer episodes in patients with Parkinson's disease who took CBD for six weeks. Three of them took 75 mg a day and the fourth 300 mg. Once they stopped, the symptoms went back to their initial level.

The same team then ran a trial that the online story almost always leaves out. In the trial by de Almeida and colleagues, published in Movement Disorders in 2021, 33 patients received either CBD at 75 to 300 mg a day or a placebo for 14 weeks. CBD did no better than placebo on the frequency of disturbed nights or on the clinical scales.

Four uncontrolled cases pointed one way, and a randomised trial against placebo found nothing: this is what a control group changes. If you have repeated nightmares, violent movements in your sleep or daytime sleepiness, see a doctor, as these are signs of disorders that can be diagnosed.

A seller may say nothing about sleep. The European Union has authorised no health claim for cannabidiol, and a product presented as a sleep aid would be treated as a medicine. Epidyolex, the only CBD-based medicine authorised, is prescribed for rare forms of epilepsy.

On 20 May 2026, the French Ministry of Agriculture stated that foods containing CBD are illegal and must be taken off the market, since they hold no novel food authorisation. On 9 February 2026, EFSA published a provisional safe level of about 2 mg of CBD a day for a 70 kg adult. Safety was not established for people under 25, during pregnancy or with medication. The 2018 study used 300 mg, 150 times that benchmark.

The calculation is set out in our article on EFSA's reference dose. Sleep onset and night length are covered in our review of CBD and sleep, and our update on the rules in 2026 lists what is still on sale.