Why Your Dreams Go Crazy When You Stop Smoking Weed

By
Photo: Tegra Stone Nuess/Getty Images

Dreams are strange, misunderstood things. This week, Science of Us is exploring the latest research that helps explain what they are, what they might mean, and how they affect our waking lives.

It is well-known, among those who use a lot of marijuana, that to stop doing so is to invite a sudden torrent of crazy, vivid dreams. Just Google “weed and dreams” and you’ll see that most of the top results are about what happens when you stop smoking pot.

Poking around online, I had little trouble finding people who could provide first-hand accounts. And their stories emphasized that if we knew more about exactly what’s going on with weed and dreams, it could potentially help a lot of people looking to scale back on their weed use, get better sleep, or both.

Franny, the first name of a U.K. resident I met on Twitter and subsequently emailed with, said that she smoked pot heavily from the time she was 26 until her early 30s, at which point she stopped. She said that when she saw a tweet I posted asking people about the weed-cessation/crazy-dreams connection, “it definitely gave me a ‘lightbulb’ moment.” “My dreams are exhausting most days,” she wrote. “They’re so real that I sometimes wake up tired after eight hours sleep. They usually involve ‘hero’ activities. Saving people from aliens, zombies, etc., but also I can sort of see myself outside of them looking at them giving myself a monologue about the dream.” Franny said that while she’s always had “hero” dreams, she doesn’t remember them ever being nearly as vivid as they are now.

I also Twitter-met a 20-year-old guy named John who lives in suburban Georgia and who had a similar story. About a year after graduating high school, he said in a DM chat, he “got to the point where I was smoking several times a day every day,” which lasted for about six months before he quit — he said the weed made him neurotic, caused him to stutter, and led to other unpleasant side effects.

Right after he quit, John said, his dreams got intensely realistic and disturbing. “The one I remember is I dreamt my mother killed herself,” he said. “And I found her body lying in her closet. When I found her I was devastated. It felt so real. I was saying, No no no no, walking up to her. Once I got close I looked down and screamed. The scream woke me up and I was screaming in real life as I woke.” Luckily, dreams like this only lasted for about a week, and since then, John said, his dreams have been fairly normal.

Finally, Seth Smith, a Navy veteran who was a year below me in grad school, told a similar story, but in reverse. “I’ve been a crazy dreamer ever since childhood,” he said in an email. “I would sleepwalk often and even make phone calls and full conversations and leave the house.” Later he underwent the military’s notoriously intense SERE training. This didn’t help with his sleep problems. At one point, he said, “I fell asleep holding the girl I was seeing at the time from behind and dreamt that I was back in the snow and the forest and the mountains of SERE with enemies giving chase. I jumped out of the bed with her still in my arms and awoke to her screams of ‘You’re sleeping on the couch.’”

Smith had heard that marijuana could help, but smoking in the military wasn’t an option. After he was honorably discharged, though, “My friends who had been so good about not using cannabis around me during my military service began to tell me about its health benefits and how it wasn’t just about getting high anymore but about really looking to use it as a medicine to treat specific problems.” He started smoking, and he said it worked wonders. “I now use cannabis regularly, particularly to treat restlessness and disturbed sleep,” he said. “I swear by it. My dreams are far less of a nuisance than they once were. My sleep seems to be very regular and sleeping next to me is no longer a contact sport.”

What can explain these stories? There seems to be a widespread belief that the connection between marijuana and dreams has to do with REM sleep. In this story line, versions of which have appeared on highly Google-ranked articles addressing the weed/dreams question in Vice, Psychology Today, and LeafScience, smoking pot reduces REM sleep and therefore dreaming, and when you quit, there’s a rebound effect: more REM sleep, and more — and more vivid — dreams.

This sounds like a tidy, satisfying explanation, but according to Dr. Timothy Roehrs, a sleep expert at the Henry Ford Health System, it’s at the very least a bit premature, in light of the available evidence, and might be entirely wrong. “The literature on whether or not marijuana affects REM sleep is extremely weak and equivocal,” he said. “Some studies have shown it suppresses REM sleep, some studies have shown it doesn’t.”

The aforementioned articles do tend to point to some studies reporting such a link, but Roehrs said that this research tends to be older — from the 1970s and 1980s — and not particularly well-controlled. Plus, he said, “there are only six studies” on this subject in the literature, total. Overall, Roehrs said he doesn’t subscribe to this theory. “If you look carefully at it, [marijuana] doesn’t suppress REM sleep,” he said.

So if the crazy dreams can’t be chalked up to unsuppressed REM sleep, what is going on?

Roehrs believes he and a colleague, Leslie Lundahl of Wayne State University School of Medicine, have uncovered a pretty big hint about what’s going on. As part of a larger study that wasn’t specifically about marijuana and dreams, he and Lundahl brought a bunch of heavy marijuana users into a sleep lab. “In the design of the study that was being conducted, marijuana was being smoked in the morning and the afternoon, and on some days it was active marijuana, and on other days it was a placebo, 0.4 percent THC or something,” he said. “The active marijuana was 3 percent THC.” So the participants alternated, in other words: one day smoking real pot provided by the researchers, the next day smoking placebo pot (Roehrs said he wasn’t sure whether or not the participants could tell the difference). The researchers recorded the participants’ sleep and compared it to the members of an age-matched healthy control group whose members weren’t heavy pot smokers, and who didn’t smoke during the experiment.

If smoking pot does something to REM sleep, that could show up in an experiment like this in at least one of two ways: First, the regular weed should affect REM sleep in a way the placebo weed shouldn’t. Second, the heavy pot smokers should have different REM-sleep patterns than the non-using control group.

On the potent-versus-impotent weed question, “smoking the 3 percent THC marijuana relative to smoking the placebo marijuana had no effect on REM sleep,” said Roehrs. And when the researchers compared the two groups, said Roehrs (that is, the night when they smoked the barely-weed weed), they found the two groups got the same amount of REM sleep.

So that’s 0 for 2. There’s an important but, though: “The sleep of the heavy marijuana users was different than the sleep of the healthy normals on the placebo night” in other ways, said Roehrs. On the placebo nights, members of the marijuana group “showed lower amounts of slow-wave sleep, which is that deep, restorative kind of sleep,” he explained, as compared to the control group and to their own sleep on nights when they smoked regular pot. “They also showed very poor sleep efficiency, meaning that in the 8 hours that they spent in bed, they slept about 80 percent of the time [sleeping].” By way of comparison, Roehrs said that “I have an entry criteria of a sleep efficiency of 85 percent or less” for an insomnia study for which he is currently recruiting volunteers. “So these people would have qualified for my study,” simply from having gone without smoking real marijuana for a day. Toking up fixed things, though. “When they got the active marijuana their sleep efficiency was normalized,” he said. “So now they were sleeping like a healthy normal” — that is, like someone in the control group.

This could explain a lot of the weird-dreams stuff. As Roehrs pointed out, “when you awaken abruptly from REM sleep,” you remember your dreams vividly. When you sleep through your REM cycles, you’re less likely to. So if, when you stop using marijuana, it makes you more likely to suddenly awaken during REM sleep — even briefly, even if you don’t remember doing so — that could leave you with some very intense dream-memories come morning.

That’s what Roehrs thinks is going on, and he pointed out that this could be part of a more general, well-established withdrawal pattern that goes on when people cut out other substances, too. “This is what happens to alcoholics,” he said. “When they discontinue alcohol, they have frequent awakenings and disruptions of sleep and they report vivid dreaming. So this might be very much a parallel.” Roehrs also pointed out that while marijuana didn’t seem to reduce the duration of REM sleep, it did seem to reduce the duration of slow-wave sleep, which could be part of the story here, too. “It’s a layperson’s perception that REM sleep is the deepest sleep,” he said. “But that’s actually incorrect: If you actually measure arousal threshold, the arousal threshold during REM sleep is less than that of deep slow-wave sleep and more akin to light sleep.” Meaning if you’re going to awaken suddenly in the night as part of that withdrawal process, it’s pretty likely that awakening will occur during REM sleep, raising the likelihood that you’ll remember whatever you were dreaming about vividly.

Again: This is as-yet unpublished research, so a grain of salt or two is appropriate. But Roehrs made strong cases about the inconsistency and poor design of the older studies that comprise the only published evidence for a marijuana-REM sleep link, and that his and Lundahl’s work had a more careful methodology behind it: Their study didn’t just compare smokers to nonsmokers, which can introduce various confounds, but also, through the use of the placebo weed, it was able to more intricately track the short-term sleep effects of smoking marijuana on heavy users.

As for my correspondents, the theory seemed to fit John’s story better than Franny’s (Seth’s is a different case, since he started smoking specifically to fight intense dreams, and hasn’t stopped). After all, John specifically said that the intense dreams stopped after a week, which could fit with some sort of withdrawal effect. Franny, on the other hand, said that 18 months after she quit smoking heavily — she still tokes up two or three times a month — the intense dreams persist (“they are basically 4K at the moment,” she said, meaning ultra-high-def). But she was quick to point out she also plays a lot of video games, and thinks they could be a factor, too.

So it can be hard to know exactly what’s going on with any individual. Maybe the key point here is less whether Roehrs’s theory fits these and other accounts exactly — anecdotes only take us so far — and more that there’s an urgent, unmet-thanks-to-the-feds need for more research into marijuana and its effects, especially as the decriminalization and legalization trains continue thundering onward. After all, Roehr and Lundahl were only able to do the study they did because Lundahl is one of a small handful of researchers licensed by the government to administer marijuana during studies. And when I asked Roehr whether the government’s long-standing, heavy-handed restrictions on marijuana research — restrictions which were just upheld by the DEA a couple months ago — have stymied expert’s progress toward better understanding the drug’s effects, he said yes, echoing a common belief among researchers.

Given marijuana’s popularity, in 2016 we should know a lot more about what it does to sleep, a rather important human activity, and to other aspects of the human experience. For now, Roehr and Lundahl’s research does offer up a theory that feels a bit more compelling and scientific than all that REM talk. But the only way to test whether it holds up is to conduct more studies, and researchers have to jump through way too many hoops to do so.