Cave team · June 25, 2026 · 12 min read

How to stop racing thoughts at night: the methods, scored

TL;DR: The most evidence-backed way to stop racing thoughts at night is a brain dump: five minutes of getting every open loop out of your head and into something that keeps it — paper, a notes app, or an AI companion like Cave that remembers for you. In a randomized sleep-lab study, people who wrote tomorrow's to-do list before bed fell asleep 37% faster than people who wrote about tasks they'd already finished.

A brain dump is not journaling, and it is not "thinking positive." It works because of three specific findings about how brains handle unfinished business, and it fails when you do it vaguely. Below: why thoughts race, each method scored by its measured effect, and how to run the one that wins.

Why won't my brain shut up at night?

Your brain treats unfinished tasks and unresolved worries as open loops, and re-raises them whenever the day's noise stops — which is exactly when you lie down.

Psychologists have measured this. In a series of four experiments, Roy Baumeister and E.J. Masicampo found that unfinished goals intrude on unrelated thinking — participants with an open task performed worse on a reading task because the goal kept surfacing (Journal of Personality and Social Psychology, 2011). The nagging is a feature: your mind rehearses what's undone so you won't forget it. At 2 a.m., the feature has no off switch.

Trying to force the thoughts away makes it worse. Daniel Wegner's famous experiments asked people not to think of a white bear; they thought of it more than once per minute, and after the suppression ended, the bear came back harder than for people who'd been allowed to think about it freely (Journal of Personality and Social Psychology, 1987). "Stop thinking and go to sleep" is a suppression instruction. Suppression rebounds.

And when the loops are emotional rather than practical, the circling has a name: rumination. Susan Nolen-Hoeksema, who spent her career studying it, defined rumination as repetitively focusing on distress and its causes without moving to action. Her research found it doesn't solve the problem — it deepens negative thinking, impairs problem-solving, and in prospective studies predicted who went on to develop depression. Replaying the argument for the ninth time is not analysis. It's a loop.

Does a brain dump actually work? The bedtime study

Yes — and the best evidence is unusually direct. In 2018, Michael Scullin's team at Baylor put 57 adults in a sleep lab, wired for polysomnography, and randomized them: five minutes writing tomorrow's to-do list, or five minutes writing about tasks already completed.

The to-do-list group fell asleep in 15.8 minutes on average; the completed-tasks group took 25.1 (Journal of Experimental Psychology: General, 2018). Nine minutes, from five minutes of writing — measured with electrodes, not questionnaires.

Minutes to fall asleep after five minutes of bedtime writing Bar chart from Scullin et al. 2018. Participants who wrote a to-do list fell asleep in 15.8 minutes on average. Participants who wrote about completed tasks took 25.1 minutes. Nine point three minutes faster, effect size d equals 0.63. Five minutes of writing, nine minutes of sleep sooner Minutes to fall asleep (polysomnography), N = 57 · Scullin et al. 2018 10 min 20 min 30 min Wrote tomorrow's to-do list 15.8 min Wrote about finished tasks 25.1 min 9.3 min faster (d = 0.63)
Offloading tomorrow beat reflecting on yesterday — writing about finished tasks left people awake longer. Scullin, Krueger, Ballard & Bliwise, J. Experimental Psychology: General, 2018.

Two details in the data matter for how you do it. First, direction: writing about completed work didn't help — it was the slower group. The relief comes from parking what's open, not from reviewing what's done. Second, specificity: within the to-do group, the more specific the list, the faster the writer fell asleep (r = −.39). Vague items like "sort out the move" don't close the loop. "Email the landlord about the deposit before lunch" does.

Why does getting thoughts out of your head work?

Because your head is working memory, not storage — and working memory never clocks out. Cognitive scientists call the fix "cognitive offloading": using physical action to cut the processing load, the same move as setting an alarm instead of trusting yourself to wake up (Risko & Gilbert, Trends in Cognitive Sciences, 2016).

Three mechanisms stack up when you dump properly:

  1. A written plan counts as handled. Masicampo and Baumeister's key finding: the intrusive thoughts stopped once participants wrote a specific plan — the goal didn't have to be completed, just planned. Your brain stands down when it trusts the loop is held somewhere.
  2. Naming a feeling turns its volume down. In fMRI work by Matthew Lieberman's lab, putting a feeling into words reduced amygdala response and engaged the prefrontal cortex instead (Psychological Science, 2007). "I'm dreading Thursday's review" does something in the brain that silent dread does not. This is also why saying it out loud works — venting is sorting, not complaining.
  3. Worry accepts an appointment. A 2023 meta-analysis of seven randomized trials (999 participants) found that postponing worry to a scheduled slot reduced how long and how often people worried (Dippel et al., International Journal of Cognitive Therapy). A bedtime dump is that appointment, kept before the 2 a.m. spiral instead of during it.

How do you do a brain dump that actually works?

Five minutes, one container, four moves. The method below is assembled from the studies above rather than from productivity folklore:

  1. Empty everything, ugly and unsorted. Tasks, worries, the thing you said in 2019. No editing, no order. The goal is zero loops left inside.
  2. Make the vague ones specific. Rewrite "deal with insurance" as the next physical action: "call Dana about the claim at 9." Specificity was what separated fast sleepers in the Scullin data.
  3. Give each worry its appointment. Next to anything that isn't a task ("what if the biopsy is bad"), write when it gets attention: "talking to mom after the results Friday." A plan, not a solution — that's all the loop needs to close.
  4. Name the feeling under the noise. One line: "mostly I'm angry about the meeting." That's the affect-labeling step, and it's the one a to-do list alone skips.

Then close the notebook, or the app, and don't reopen it. The container has it now. If you dump in the same place every night, the ritual itself becomes a signal that the day is filed — and you build a record you can actually look back on, which is where a dump quietly turns into a journal you didn't have to force.

Every method, scored: how much do they really help?

More than nothing, less than a miracle — and it depends on which tool you mean. The effect sizes across this family of techniques are worth seeing side by side, because the pages that recommend them almost never show the numbers:

Effect sizes of writing and worry-offloading techniques Horizontal bar chart of Cohen's d values. Expressive writing across 146 trials: about 0.15. Worry postponement, frequency of worry: 0.19. Worry postponement, duration of worry: 0.31. Bedtime to-do list and faster sleep onset: 0.63. Worry postponement in diagnosed generalized anxiety: 0.82. Reference marks at 0.2 small, 0.5 medium, 0.8 large. What the trials actually found (effect size, Cohen's d) Bigger bar = stronger measured effect · dashed lines: 0.2 small / 0.5 medium / 0.8 large 0.2 0.5 0.8 Expressive writing (146 trials) ≈0.15 Worry postponement — frequency 0.19 Worry postponement — duration 0.31 Bedtime to-do list → sleep 0.63 Worry postponement in GAD 0.82
Sources: Frattaroli 2006 meta-analysis (r = .075, converted); Dippel et al. 2023 meta-analysis; Scullin et al. 2018; Krzikalla et al. 2024 RCT. Small effects are still real — and these techniques are free.

The honest reading: expressive writing, across 146 randomized studies, helps a little on average (Frattaroli, Psychological Bulletin, 2006). Worry postponement helps modestly in general and a lot in diagnosed anxiety — a 2024 trial found d = 0.82 and a 40% recovery rate for worry in GAD. The bedtime to-do list sits in the middle and targets the exact moment you care about. None of this is a cure. All of it is cheap, safe, and repeatable tonight.

One group should take the technique most seriously: ADHD brains. Racing thoughts at night are near-universal there — in a 2021 clinical study, 44% of adults with ADHD met criteria for full insomnia disorder, against 6–15% of the general population, with trouble falling asleep the most common form. Russell Barkley, the field's leading researcher, made externalization a core principle of ADHD management: information held in an ADHD head is a weak signal, so get it out of the head and into the world. For ADHD, a nightly dump isn't a hack. It's the strategy.

And a boundary: if your mind races every night for weeks, with sleep, appetite, or mood sliding, that pattern deserves a clinician — insomnia and anxiety are treatable, and a notebook is not treatment.

Where should the brain dump go?

Into whatever you will actually use at 11:47 p.m. — but the containers aren't equivalent, so choose on purpose:

  • Paper has the lowest friction and no notifications waiting to ambush you. Its weakness: last Tuesday's dump is functionally gone. Nobody rereads page 34.
  • A notes app is searchable and always with you. Its weakness: it's a junk drawer. The dump lands next to the grocery list, and nothing ever comes back to you unless you go digging.
  • An AI companion is the only container that talks back. Cave is an AI companion with real memory — you think out loud, and it keeps what you said, organized by topic, readable whenever you want. In practice that covers the steps a tired brain skips: it asks the small next question that turns "sort out the move" into an actual plan, it reflects the feeling back so it gets named, and tonight's dump lands connected to what you said about the same worry in March instead of evaporating. At 2 a.m. it's awake, and if typing feels like too much, you can just say it out loud on a voice call.

The fair caveat: if a paper notebook is the thing you'll reliably open, a paper notebook beats any app you'll abandon. The best container is the one that's still in use in November.

Sources

FAQ

Does writing before bed actually help you fall asleep?

Yes, if you write the right thing. In a randomized sleep-lab study (Scullin et al., 2018), people who spent five minutes writing tomorrow's to-do list fell asleep in 15.8 minutes versus 25.1 for people writing about completed tasks — 37% faster, measured by polysomnography. More specific lists worked better. Writing about the past did not help; offloading the future did.

What is the difference between a brain dump and journaling?

A brain dump is unloading: everything open in your head goes into a container, fast and unsorted, so your working memory can stand down. Journaling is reflecting: choosing one thread and thinking it through in writing. The dump helps you fall asleep tonight; the journal helps you understand your month. Dumps also make journaling easier to sustain, because the raw material is already captured.

Should I do a brain dump on paper or in an app?

Whichever you'll still be using in a month — consistency beats the tool. Paper is fastest and distraction-free but doesn't come back to you. A notes app is searchable but inert. An AI companion like Cave adds what the others can't: it asks the follow-up question that makes vague worries specific, and it remembers the dump — organized by topic, readable later, connected to what you said weeks ago.

What if the same worries come back every night?

Recurring worries respond to scheduling, not suppression. Across seven randomized trials, postponing worry to a fixed daily slot reduced both its duration and frequency, and a 2024 trial in diagnosed anxiety found a large effect (d = 0.82). Write the worry down with its appointment. If the loop persists for weeks and starts costing you sleep and mood anyway, bring it to a clinician — chronic insomnia and anxiety are treatable conditions.

Does a brain dump help with ADHD racing thoughts?

It's one of the most-recommended strategies there is. Externalizing — moving information out of your head and into the world — is a core principle of ADHD management in Russell Barkley's framework, and the need is real: about 44% of adults with ADHD meet criteria for insomnia disorder, with trouble falling asleep the most common type. A nightly dump into one consistent place is the minimum viable version.