Dreams and Creativity: The Science of Dream Incubation
Dreams and creativity have a specific, tested relationship that goes beyond the general idea that sleep helps you think more clearly. It's not just that sleep matters — it's that REM sleep specifically, the stage where vivid dreaming happens, changes the way your brain forms associations in a way no other sleep stage does. That distinction comes from a controlled study, not folklore, and it points to a technique you can use directly: deliberately steering a dream toward a problem you're stuck on.
The Study That Isolated REM's Effect
In 2009, Sara Mednick's lab at UC San Diego published a study in the Proceedings of the National Academy of Sciences that settled a question researchers had been circling for decades: does sleep improve creative thinking because of rest generally, or because of something specific to REM sleep? Denise Cai and colleagues had participants attempt Remote Associates Test problems, then split them into three groups — a nap with REM sleep, a nap without REM sleep, and quiet rest with no sleep at all — before retesting them on new RAT items, some of which were related to problems they'd been exposed to before the break.
Only the REM-sleep group showed a meaningful improvement, roughly 40% better than their pre-nap baseline on problems related to material they'd encountered earlier. The non-REM nap group and the quiet-rest group showed no comparable gain. This is the finding that separates "sleep helps creativity" from the more precise and more useful claim: REM dreaming specifically primes the kind of loose associative search that the Remote Associates Test measures, and quiet incubation without REM doesn't do the same thing.
Why REM Sleep Works Differently
During REM sleep, the brain's noradrenergic system — the neurochemical system associated with focused attention and tight, categorical thinking — goes largely offline, while acetylcholine activity rises. The practical effect is a temporary loosening of the same associative constraints that keep your waking thought organized and on-topic. Ideas and images that would normally be filtered out as irrelevant get activated and combined more freely.
This is a more specific mechanism than the general hypnagogic loosening that happens in the drowsy transition into sleep, which is covered in more depth in sleep and creativity. Hypnagogia is a brief transitional state; REM dreaming is a sustained one that recurs multiple times a night and increases in duration toward morning, which is part of why dream recall is more common — and dream content often more vivid — right before waking.
The Historical Record, Read Carefully
Several well-known creative breakthroughs are credited to dreams, and it's worth being accurate about how solid each story actually is rather than repeating them uncritically.
Mary Shelley's Frankenstein has the strongest documentation. In her 1831 introduction to the novel, Shelley directly describes the "waking dream" — a vivid hypnagogic vision of a scientist recoiling from his creation — that gave her the novel's central image during a ghost-story writing contest at Villa Diodati in 1816.
Paul McCartney's "Yesterday" is well documented in McCartney's own repeated retelling: the melody came to him fully formed in a dream, and he was so convinced he'd unconsciously plagiarized it that he played it for other musicians for weeks, asking if they recognized it, before accepting it was original.
Kekulé's benzene ring — the story that a snake-eating-its-tail dream image revealed benzene's ring structure to August Kekulé — is the most cited dream-discovery story in chemistry, but historians of science, notably Alan Rocke, have raised real questions about how literally to take Kekulé's own account, given it was delivered decades after the fact in a self-mythologizing speech. It's a useful illustration of the concept, but it shouldn't be cited as settled history the way the Mednick lab's controlled study can be.
The pattern across the well-documented cases is consistent with the incubation effect research more broadly: the dreamer had already done substantial conscious work on the problem before the dream occurred. Dreams don't manufacture expertise. They recombine material that's already been loaded into the system.
Deirdre Barrett's Dream Incubation Research
Harvard psychologist Deirdre Barrett has run the most direct experimental work on deliberately using dreams to solve problems, described in her book The Committee of Sleep (2001). In her studies, Barrett asked college students to select a real personal or academic problem and follow a specific incubation protocol before sleep for a week. Roughly half of participants reported a dream that related to their chosen problem within that week, and about a quarter reported a dream they felt contained an actual solution or useful insight — a substantially higher rate than would be expected from unguided dreaming.
The protocol she used is specific enough to replicate:
- Write the problem down as a single, concrete sentence. Vague problems produce vague dreams. "How do I improve team communication" incubates worse than "why does the Tuesday standup keep running long."
- Review the written problem right before turning off the light, as the last conscious activity before sleep.
- Form a visual image connected to the problem. Barrett's participants who created a specific mental picture related to their problem had higher incubation success rates than those who only thought about it verbally.
- Keep a notebook or phone by the bed. Dream recall decays within minutes of waking — often before you've fully sat up. Record anything you remember immediately, even fragments that don't seem to make sense yet.
- Repeat nightly. Barrett's data showed incubation success accumulating over several nights, not reliably on the first attempt.
Why This Connects to Waking Creative Practice
Dream incubation isn't a separate creative skill from the rest of the material on this site — it's an extension of the same incubation effect that makes stepping away from a creative block useful in the first place. What dream research adds is precision about which specific biological state does the work. Passive waking rest helps somewhat. REM sleep, per the Mednick lab's data, appears to help more, for the specific class of problems that require finding a non-obvious connection between things you already know.
The practical implication is straightforward: if you're stuck on a problem that has a "connect the dots" quality — the kind remote associates problems are built to simulate — doing the conscious preparatory work, then deliberately loading the problem right before sleep using Barrett's protocol, is a tested method, not a superstition. It won't help with problems that are missing information you don't yet have. It's specifically suited to problems where the pieces are already in your head and haven't yet been combined.
Ready to train your creativity? Try science-backed exercises that measure and improve your creative thinking. Start a Free Exercise