Sleeping on it - how REM sleep boosts creative problem-solving

Blogging on Peer-Reviewed ResearchThe German chemist Friedrich Kekule claimed to have intuited the chemical structure of the benzene ring after falling asleep in his chair and dreaming of an ouroboros (a serpent biting its own tail). He's certainly not the only person to have discovered a flash insight after waking from a good sleep. In science alone, many breakthroughs were apparently borne of a decent snooze, including Mendeleyev's creation of the Periodic Table and Loewi's experiments on the transmission of nervous signals through chemical messengers.

Most of us have tried sleeping on a difficult problem before and using an elegant experiment, Denise Cai from the University of California in San Diego has shown that this old technique really does have merit to it. She found that our brains are better at integrating disparate pieces of information after a short bout of REM (rapid eye movement) sleep - a deep, dream-rich slumber that involves a rapid fluttering of the eyes.  Cai thinks that REM sleep catalyses the creative process by allowing the brain to form connections between unrelated ideas.

Cai is by no means the first person to link sleep or dreaming to creative revelations, but she is one of the few to test it directly through experiments. She asked 77 people to complete a task, where they were given a list of three words and had to find a fourth that was linked to all three. For example, 'cookie', 'heart' and 'sixteen' are all associated with 'sweet'. In each example of this 'Remote Associates Test' (RAT), the missing fourth word has a different relationship to each of the three targets.


Each volunteer was tested at 9am and 5pm on the same day and in between, they were brought into a special room at 1pm and monitored. They either rested quietly to the tune of classical music, or slept for a maximum of 90 minutes. Of the 40 who dozed off, 28 managed to get some REM sleep. Regardless of how they spent their time, all the recruits did better at the word puzzles in the second session. That's what you'd expect; the mere passage of time ought to give people a slight edge, as they mull the problems in their minds.

Some of the recruits also did a different set of tests after their morning puzzles, where they had to fill in a missing word to complete an analogy (e.g. FAST is to SLOW as HARD is to E...) Unbeknownst to them, some of the missing words were actually the answers to the earlier RAT task. This time, the recruits who got some REM sleep in the intervening time showed significant improvements in their afternoon test scores, while those who only achieved non-REM sleep or a quiet rest fared no better.


The REM naps were usually longer than the non-REM ones, but Cai found that the total length of sleep had no relation to the recruits' success at the second round of tests. It was not the quantity of sleep but its quality that made all the difference to their performance. Only REM sleep allowed them to effectively draw the link between the clues hidden in the analogy tasks and their memory of the morning RATs .

Of course, having a catnap in the day does improve a wide range of mental abilities from alertness to memory. That alone might have been behind the recruits' afternoon improvements, but Cai showed otherwise. She found that the REM-sleepers weren't any more likely than the others to remember the answers to the morning analogies, and in fact all three groups remembered about 90% of them. It was only the REM group that managed to use that information to their advantage.

Cai also found that the REM sleepers scored higher in the afternoon tests if they were given RATs that they had seen before - if they were given new sets of words, their scores were the same as the other two groups. That strongly suggests that in this case, the benefits of REM sleep lay not in boosting memory or general mental agility, but in specifically allowing the volunteers to create associations between existing ideas. Indeed, many thinkers have defined creativity as exactly that.

In 2004, Ullrich Wagner also studied the link between slumber and creativity, and showed that people are twice as likely to discover a novel solution to a mathematical problem after sleeping on it for a night. Wagner suggested that sleep might improve "cognitive flexibility" and Cai agrees.

She suggests that REM sleep (aided by falling levels of neural signalling molecules like norepinephrine and acetylcholine) helps us to incorporate new information into existing experiences. That creates a richer network of links for us to draw upon in the future and providing the fuel for flashes of insight.

Reference: Cai, D., Mednick, S., Harrison, E., Kanady, J., & Mednick, S. (2009). REM, not incubation, improves creativity by priming associative networks Proceedings of the National Academy of Sciences DOI: 10.1073/pnas.0900271106 Image of sleepy men by Bertlividet

More on sleep and creativity:


More like this

Over the weekend, I had a little essay in the Times on some new research on why dream at night. When I can't sleep, I think about what I'm missing. I glance over at my wife and watch her eyelids flutter. I listen to the steady rhythm of her breath. I wonder if she's dreaming and, if so, what story…
It's a shame that we stop encouraging naps once the preschool years are over. After all, there's a growing body of scientific evidence that the afternoon siesta is an important mental tool, which enhances productivity, learning and memory. (It's really much more effective than a cup of coffee.)…
In my final year of university, with exam deadlines looming and time increasingly fleeting, I considered recording some of my notes and playing them over while I was asleep. The concept of effectively gaining 6 extra hours of revision was appealing, but the idea didn't stick - it took too long to…
After Freud lost his scientific credibility, psychology became very dismissive of dreams. The leading scientific theory held that dreams consisted of mental detritus, the scraps and fragments of memories that your brain didn't want to remember. While Freud mined our nighttime thoughts for hidden…

Arvind, you're a genius. I will now take copies of this paper along to conferences so that I can justify it when I fall soundly asleep. "No, no, I'm just boosting my ability to form associations between your riveting talk and my existing knowledge."

So I can quote this study to my boss, then?

Isn't that interesting. Maybe that's why I have to start writing in the morning--I've been prepped during sleep. If I have to do something else I find it much harder to get going on my book even a couple of hours later.

wow...that is an impressive looking graph! (I've just been reading over my dissertation work and totally laughing at all my dubious correlations)

I am going to print out this paper and clip it to my lab coat when I take my thrice-weekly nap on the couch in our lobby - that way, everyone will know that I am actually hard at work from 3:30 to 4 pm!

Not to be a wet blanket, but (contrary to the advertising) this study doesn't tell us anything about how REM aids creative problem solving. Knowing that norepinephrine and acetylcholine levels differ in REM and in deep sleep is sorta' informative, but not really. It's sort of like knowing that your hard drive makes funny noises when you save documents; it's not false, but it's not an explanation either.

I have sleep apnea, moderate to severe. When I do "nap" in the afternoons, or mornings, it is unintentional, and there are no dreams. I almost died in a fire last November. I fell asleep on the couch. Literally, sat down, and was out cold, while stew was on the stove. 4 smoke detectors and one Carbon Monoxide detector, all on the same floor were screaming loudly, as my husband was coming home from work. I do not remember him trying to wake me up. This has happened more than once in the last few years. My CPAP is now at level 10, but I am breathing over that! Without dreaming, I am a walking zombie during the day, am depressed, tired, anxious, defensive and moody. I have been tested twice since October of 2008, and both show signs of little or no REM sleep. I can't survive like this much longer, and would like to know what to do? Any suggestions?