Can you choose what you dream about?

I was flipping back through my dream journal a few weeks ago, looking for something else entirely, and I noticed how many entries start with whatever I'd been thinking about before bed. Not always directly. But the thread was there more often than I expected.

It made me wonder whether that connection is just coincidence or something you can actually use. Whether you can load a question into your mind before sleep and have it show up in a dream. The research says: sort of. Not reliably, and not with any precision, but more often than chance. And people have been trying this for a very long time.

This isn't a new idea

The ancient Greeks built entire temple complexes around it. Over 300 Asclepieia operated across the Mediterranean, temples dedicated to Asclepius, the god of medicine. People would travel hundreds of miles, fast, undergo purification rituals, and then sleep in a sacred chamber hoping to receive a healing dream. The practice was called enkoimesis, and it was a formal part of Greek medicine for centuries.

Egypt had something similar. Patients at temples of the healing god Imhotep were placed in darkened chambers, sometimes after chanting and fasting, to await dreams that would guide their treatment.

We obviously don't attribute divine messages to dreams in a clinical setting anymore. But the underlying question, whether you can set an intention before sleep and have it influence what shows up, is something modern sleep researchers have been testing with real data.

What the research actually says

Deirdre Barrett, a psychologist at Harvard Medical School, has probably done more to study this than anyone. Her method is straightforward: she asks participants to choose a specific problem they're working on, spend 15 minutes thinking about it right before bed, and hold a mental image of the problem as the last thing in their mind before sleep. Then they do this every night for a week.

The results are consistent across her studies. About half the participants dream about their chosen problem. Of those, roughly half find something useful, whether that's a new angle on the problem or a direct solution. That's about 25% of people getting a dream that meaningfully addresses something they chose to think about.

Barrett describes dreaming as "thinking in a different biochemical state." Her view is that the brain doesn't shut off problem-solving during sleep. It just switches to a mode that favors imagery and loose associations over linear reasoning. Problems that benefit from visualization tend to get the most help from dreams.

An older study by William Dement makes a similar point. He had 500 students attempt brainteasers before sleep, and 87 of their dreams directly addressed the puzzle. Some dreamed the answer outright.

It gets weirder in the lab

The technology has caught up to the question. At MIT's Media Lab, researchers built a device called Dormio, a hand-worn sensor that detects when you're entering the earliest stage of sleep (N1, or hypnagogia) and plays a pre-recorded audio cue at that exact moment.

In a 2025 study using Dormio, 25 participants were cued with the word "tree" as they drifted off. 92% of them reported dreams that directly included the theme. Across all awakenings, 40.6% included the cued content. And a week later, 42% of participants were still reporting the theme in their home dream journals, without any further cues.

The study also found something I didn't expect. The experience of successfully influencing a dream changed how people thought about their dreams going forward. Their "dream self-efficacy" went up, basically how much they believed they could shape what happens in their dreams, and it stayed elevated a week later.

Then in February 2026, a team at Northwestern led by Karen Konkoly took it further. They used targeted memory reactivation — playing sounds linked to unsolved puzzles during REM sleep — and found that dream-related puzzles were solved at a 42% rate, compared to 17% for puzzles that weren't cued. The dreams didn't just reference the puzzles. They helped solve them.

How to try it without a sleep lab

The research-grade studies use EEG caps and specialized devices, but Barrett's basic technique doesn't require any equipment. I've been reading through the protocols, and the steps are simpler than I expected.

Pick something specific. Vague intentions produce vague results. "I want to dream about my project" is weaker than "I want to understand why that conversation went wrong." The more concrete and personal the question, the more likely it is to show up.

Then spend real time with it before bed. Barrett's protocol uses 15 minutes, not a passing thought. Some researchers recommend free-writing about the problem to get past surface-level thinking. The Dream Studies Portal suggests writing your intention down and repeating it as the last thing before sleep.

If the problem lends itself to an image, use one. Barrett found that holding a visual representation, a photo or a diagram or even a physical object on your nightstand, improves how often the topic shows up in dreams. Your brain during REM is running heavily on imagery. Give it something to work with.

Don't expect results on the first night. Barrett's studies run for a week. Some researchers suggest continuing for two weeks and writing down all your dreams during that period. The connection to your intention doesn't always appear in obvious ways. Sometimes it takes reading back through several nights of entries before you see it.

And write everything down immediately. This part isn't optional. You lose roughly 90% of dream content within ten minutes of waking. If your dream did address your problem, you need to catch it before it disappears.

What people actually report

Reading through forums and other dream journalers' experiences, the pattern is consistent: the first few nights, nothing obvious. Then a dream shows up that doesn't seem related at first but connects to the question sideways. A metaphor. An old memory resurfacing. Something that only makes sense when you sit with it for a while.

A 2026 study from IMT School for Advanced Studies Lucca helps explain why. Valentina Elce's team analyzed over 3,700 dream reports from 287 participants and found that dream content is shaped by personality, emotional priorities, and waking attention. People who value their dreams and engage with them tend to have richer dream experiences. What you pay attention to feeds forward.

This lines up with what Barrett and others have found. You're not programming a movie when you try dream incubation. You're loading a question into a system that already recombines your experiences overnight. It might deliver something useful, or something oblique that only clicks days later. Or it might not respond at all.

What I still don't know

I find this research genuinely interesting, and I want to try it properly. But it matters to be honest about the limits. The success rates in controlled studies sit around 25-50% for problem incorporation, depending on the study. That's well above chance, but it's not a guarantee. And "the dream addressed my problem" covers a wide range, from direct solutions to metaphorical connections that take some interpretation.

What the research does suggest is that dreams aren't sealed off from your waking intentions. What you think about before sleep shapes what shows up. A dream journal makes that visible.

I'm going to try Barrett's protocol for a week, pick a specific question, spend 15 minutes with it before bed, and write down whatever shows up in the morning. If nothing comes of it, the worst case is a week of paying closer attention to my dreams. That's never been a waste.