
(Credit: Andrea Piacquadio from Pexels)
Dreams appear to allow the brain to break free creatively in ways the conscious mind can not
In A Nutshell
- Researchers successfully steered people’s dreams toward specific unsolved puzzles by playing associated sounds during REM sleep, proving dreams can be deliberately influenced.
- For participants whose dreams responded to the sound cues, problem-solving success doubled, jumping from 20% to 40% the next morning.
- Surprisingly, forgotten dreams worked best: puzzles that people worked on during sleep but couldn’t remember dreaming about had a 67% solving rate, compared to just 11% for lucid dreams.
- The technique kept working after the lab: participants continued dreaming about the puzzles on their own for a week, suggesting pre-sleep focus has lasting effects on dream content.
Everyone’s heard the advice to “sleep on it” when facing a tough problem. Turns out there’s real science behind that old wisdom.
Researchers at Northwestern University have shown that dreams themselves, not just a good night’s rest, actively help crack problems that stumped us while awake. Better yet, they figured out how to aim your dreams at specific challenges while you’re sleeping.
We’ve all encountered problems that feel insurmountable. Well, imagine drifting off to sleep, and while you’re dreaming, scientists play a soft melody you heard while working on that problem. Your dreaming mind picks up the cue and starts playing with the problem again, free from the mental ruts that trapped you while awake. The next morning, the solution suddenly feels obvious.
That’s what happened in this study. When people whose dreams responded to the cues dreamed about a puzzle, their odds of solving it the next morning increased from about 20% to 40%.
Steering Dreams With Sound
The study, published in Neuroscience of Consciousness, recruited 20 people who frequently experience lucid dreams (those strange moments when you realize you’re dreaming while still asleep). Each person tackled several brain teasers designed to need creative leaps rather than logical grinding. The puzzles ranged from matchstick challenges to visual riddles, all requiring that “aha!” moment instead of step-by-step calculation.
Each puzzle came with its own 15-second soundtrack: maybe a bit of music, or environmental sounds. Participants worked on each puzzle for three minutes while that sound played on repeat, wiring their brains to connect sound and problem.
Then came bedtime. Once participants fell into REM sleep (the stage where vivid dreaming happens) researchers quietly played some of those soundtracks through speakers. Not all of them, just a randomly selected half. The sounds started barely audible and gradually increased, but only if participants stayed soundly asleep.
Twenty frequent lucid dreamers participated, and 15 of them dreamed about their unsolved puzzles at least once. Sometimes the puzzle appeared directly in the dream. Other times the connection was more abstract: one person heard a cue for a balance-scale puzzle and dreamed about a seesaw. Another dreamed of catching fish with a net, then woke up and solved a puzzle about fishing.
Participants even communicated from inside their dreams. They’d been trained to signal by moving their eyes in a distinct left-right pattern when they realized they were dreaming, and to do rapid sniffing breaths if they heard a puzzle cue and started working on it. Brain monitors caught these signals in real-time while participants slept.

When Dreams Work Best, And When They Don’t
Not everyone’s dreams responded to the sound cues equally. About 12 participants were “targeted dreamers,” meaning their dream content clearly shifted toward the puzzles they heard cues for compared to ones they didn’t. For these people, the benefits were real: cued puzzles got solved at twice the rate of uncued ones.
The other eight participants? The sounds didn’t seem to steer their dreams much, and accordingly, they didn’t show the same problem-solving boost. For those whose dreams didn’t respond to cues, there was no corresponding benefit. This suggests the dreaming itself matters, not just the sleep. Why some people’s dreams are more open to outside influence remains a mystery, though it might relate to how different brains process sound during sleep.
Interestingly, lucid dreams weren’t actually the most helpful. Puzzles that showed up in regular dreams were solved 46% of the time the next morning. Puzzles in lucid dreams? Just 11%.
The best results came from something researchers didn’t expect. Six times, participants signaled with sniffing breaths that they’d heard a puzzle cue and were working on it, but when they woke up, they couldn’t remember dreaming about that puzzle at all. Four of those six puzzles got solved anyway, a 67% success rate.
This pattern hints that maybe the sleeping brain does its best creative work when we’re not consciously directing it. In regular dreams, your mind freely explores weird connections and wild possibilities. But when you’re lucid and deliberately trying to solve something, you might fall back into the same mental traps that blocked you while awake.
The forgotten dreams are perhaps the most intriguing piece of all. They suggest the most powerful problem-solving might happen in parts of dreaming we never remember, and that trying to hold onto every detail might actually interfere with the process.
Why REM Sleep Is a Creativity Goldmine
This study adds hard evidence to stories that have floated around for centuries. Paul McCartney said the melody for “Yesterday” came to him in a dream. The chemist who figured out the ring structure of benzene credited a dream about a snake biting its own tail. Mary Shelley dreamed the central scenes of Frankenstein.
But anecdotes prove nothing on their own. Maybe those people would have had the same insights while awake. Maybe the dreams just reflected creative work already happening beneath conscious awareness. By controlling which specific puzzles entered people’s dreams, these researchers could test whether the dreaming itself mattered.
It does appear to matter. And REM sleep seems particularly suited for creative breakthroughs. During this stage, the sleeping brain makes unusual connections between ideas that seem unrelated while awake. Studies show people woken from REM sleep come up with more unexpected word associations than people woken from other sleep stages. The brain’s wiring during REM appears to favor the kind of loose, associative thinking that sparks creative solutions.
After leaving the lab, participants kept dream journals for about a week. A dozen of them reported dreaming about the study’s puzzles on their own, even without any sounds being played. That persistence suggests focusing on problems before sleep keeps them bubbling in dreams for days.
Of course, questions remain. When participants woke up and remembered dreaming about a specific puzzle, did they then think about it more consciously before being tested? Are the benefits happening in the dream itself, or in the hazy moments of remembering it, or in the hours between? The study can’t fully separate these possibilities.
Still, the core finding holds up: dreams can actively contribute to solving problems, not just reflect solutions the sleeping brain arrived at through other means. And now we can deliberately provoke dreams about specific challenges.
Future sleep technology may one day allow people to target their dreams toward real-life problems: preparing for a difficult conversation, working through a creative block, or finding new angles on work challenges. For now, at least, we know that sleeping on it isn’t just rest. Your dreaming mind is working, making connections your waking brain misses, quietly solving problems while you’re off in some other world entirely.
Paper Notes
Limitations
The within-subjects design meant each participant juggled multiple puzzles simultaneously, potentially causing interference between problems. Binary outcomes (solved or unsolved) reduced statistical power. Sample size—20 participants with roughly 8 unsolved puzzles each—was likely too small to detect all effects reliably.
Participants couldn’t be fully blinded once they heard puzzle cues in their dreams. Researchers couldn’t rule out demand characteristics, where participants unconsciously try to confirm what they think the study expects. The design couldn’t separate effects of dreaming from cognitive activity during dream reporting or unconscious processing between dreams and testing.
The study lacked a control night without cues to establish baseline rates. Cues for one puzzle might have activated memories of related puzzles, diluting observed effects. Because participants were told to focus on puzzles before sleep and that they might hear cues, the study couldn’t isolate whether benefits came purely from dreaming versus the combination of pre-sleep intention and dream content.
Funding and Disclosures
Karen R. Konkoly was supported by the US National Institutes of Health (T32-HL007909). The study received funding from National Science Foundation grant BCS-1921678, US National Institutes of Health grant DP1-HL179370, and the Mind Science Foundation 2022 BrainStorm Award to Karen R. Konkoly. The authors declared no conflicts of interest.
Publication Details
Title: Creative problem-solving after experimentally provoking dreams of unsolved puzzles during REM sleep | Authors: Karen R. Konkoly (Psychology Department, Northwestern University, and Dust Systems), Daniel J. Morris (Psychology Department, Northwestern University), Kaitlyn Hurka (Psychology Department, Northwestern University), Alysiana M. Martinez (Psychology Department, Northwestern University), Kristin E.G. Sanders (Psychology Department, University of Notre Dame), Ken A. Paller (Psychology Department, Northwestern University) | Journal: Neuroscience of Consciousness, 2026, Volume 2026(1), niaf067 | DOI: 10.1093/nc/niaf067 | Published: February 5, 2026 | Corresponding Author: Karen R. Konkoly, Psychology Department, Northwestern University, Swift Hall, 2029 Sheridan Road, Evanston, IL 60208, United States. Email: [email protected] | The study was approved by the Institutional Review Board at Northwestern University and all participants provided informed consent.







