How to Fall Asleep When Your Mind Won’t Stop

Published 2026-05-20 · Updated 2026-05-20 · ~12 min read

By Drew Perdichizzi, Five Element Acupuncturist – research scientist turned acupuncturist, founder of RESTorative Co. & creator of Whisperliminals

You finally got to bed. The second your head hits the pillow, the meeting you have at 10 a.m. starts running on a loop, and underneath that loop is the conversation you should have had three hours ago, and underneath that is the unrelated thing you saw at lunch that your brain has now decided to relitigate. An hour passes. You check the clock. Now you’re calculating how few hours of sleep you’ll get if you fall asleep right now, which makes falling asleep right now even less likely.

Your mind isn’t doing anything wrong here. It’s doing exactly what it’s trained to do – keeping you sharp, anticipating, problem-solving – at the precise hour you need it to wind down. The cognitive horsepower that runs the workday is the same engine that won’t shut off at 11:47 p.m.

What follows is dense, because the problem has a real research base and no single hack. The structure front-loads the highest-leverage moves so you can start with them tonight.

TL;DR – the short version

• Racing thoughts at bedtime are a measurable phenomenon, not a character flaw. Researchers call it pre-sleep cognitive arousal. The standard instrument for measuring it – the Pre-Sleep Arousal Scale, developed by Nicassio and colleagues in 1985 – has been used in roughly 800 studies since.
• Trying harder to stop thinking makes it worse. Harvey’s (2002) cognitive model identifies five processes that keep insomnia going, and four of them are amplified by effortful thought suppression. The harder you push, the more aroused your brain gets.
• Two evidence-based moves for tonight: constructive worry (90 minutes before bed) and paced breathing (in bed). Take 15 minutes around 8 p.m. to write each worry in one column and the next concrete action in another. Then in bed, breathe at about six breaths per minute for 10 minutes to drop physiological arousal.
• If you’re still awake 20 minutes after lights-out, get up. Stimulus control – Bootzin’s 1972 protocol, supported by 2024 systematic reviews – is the most validated technique in the CBT-I toolkit for breaking the bed-as-struggle association.
• For chronic insomnia, the long-term answer is CBT-I. Cognitive Behavioral Therapy for Insomnia is the gold-standard treatment recommended by the American College of Physicians and the American Psychological Association, with response rates of 70–80% across more than 30 randomized controlled trials.

Why your mind races at night – the actual science

What you’re experiencing has a name. Researchers call it pre-sleep cognitive arousal, and they measure it with the Pre-Sleep Arousal Scale (PSAS), a 16-item self-report instrument that distinguishes cognitive arousal (racing thoughts, mental alertness, replaying conversations) from somatic arousal (heart racing, muscle tension, gut symptoms). Nicassio and colleagues developed the scale in 1985, and cognitive arousal has since been shown to predict sleep difficulty more strongly than the somatic kind.

A 2021 paper in Comprehensive Psychiatry argued that racing thoughts at bedtime are a distinct construct from the worry and rumination that researchers had previously conflated with them. Racing thoughts move fast, jump topics, and feel uncontrollable in a way that single anxious thoughts do not.

There is a neuroimaging story underneath this. In good sleepers, the brain’s default-mode network – a constellation of regions including the medial prefrontal cortex, the posterior cingulate cortex, and the precuneus – quiets down as the body transitions toward sleep. In people with insomnia, that quieting doesn’t happen reliably. A 2023 study published in Sleep found that greater waking connectivity between the retrosplenial cortex, the hippocampus, and other default-mode-network hubs predicted lower polysomnographically measured sleep efficiency that same night. The self-referential thought machinery stays online when it should be powering down.

For high-achievers, this matters because the default-mode network is the same circuit that handles planning, anticipating, replaying scenarios, and self-evaluation – the cognitive work that produces results during the day. The brain that builds your career is the brain that won’t stop building at midnight.

Why “just stop thinking” makes it worse

The intuitive move when racing thoughts hit is to try to stop them. This is almost universally what people try first, and it almost universally backfires.

Allison Harvey’s 2002 cognitive model of insomnia, published in Behaviour Research and Therapy, identified five maintaining processes that turn a few rough nights into a chronic problem:

1. Worry about not sleeping triggers physiological arousal. “I have to be up at 6, I cannot lose another night” – the thought itself elevates cortisol and heart rate.
2. Selective attention to sleep-related threats. Once you’re worried, you start monitoring: how tense your shoulders feel, how warm the room is, what time the clock says. The monitoring itself increases arousal.
3. Overestimation of the deficit. People with insomnia consistently underestimate how much they actually slept and overestimate how much daytime impairment they suffered.
4. Dysfunctional beliefs about sleep. “I need eight hours or I can’t function.” “I’ll be useless tomorrow if I don’t fall asleep in the next 15 minutes.” These beliefs are themselves arousing.
5. Safety behaviors. Going to bed early to “catch up,” canceling plans, avoiding caffeine after 10 a.m., napping defensively. Each one reinforces the underlying belief that sleep is fragile and you are at risk.

Trying to suppress the thoughts triggers nearly all five at once. The effortful suppression itself counts as cognitive work, which raises arousal. The closer you watch for the next intrusive thought, the more you find. The harder you fail, the stronger the belief that something is wrong with your sleep.

The way out is not effort. It is a different approach altogether – one that gives the brain something neutral to do, breaks the bed-as-struggle association, and over time changes what plays at bedtime in the first place.

The pre-bed protocol – what to do 60 to 90 minutes before lights out

The leverage in falling asleep faster is mostly not in bed. It’s in the hour or two before.

Constructive worry, around 90 minutes before bed. This is the single most-studied bedtime cognitive technique that does not require a therapist. The protocol – refined by Carney in CBT-I practice and supported by work from McGowan and Behar – takes 15 minutes. Sit down with a piece of paper. Draw two columns. In the left column, write everything that is on your mind: every worry, every open loop, every conversation you didn’t finish. In the right column, next to each item, write the next concrete action. Not the solution. The next step. “Pay invoice” becomes “open inbox at 9 a.m., find invoice email, mark for action.” A 2024 trial on worry postponement, a closely related variant, showed measurable reductions in pre-sleep cognitive arousal compared to no-intervention controls.

The mechanism is straightforward. Worry runs as long as the brain perceives that the worry is doing useful work. The moment a specific action exists, the worry circuit gets a place to discharge.

Wind down for real. The hour before bed should be low-stimulation. The most common mistake is using “wind down” as a label for activity that is anything but – checking work email, doom-scrolling, watching prestige TV that requires you to track six plotlines. The activity itself is cognitively engaging, and that engagement is what keeps the default-mode network online when you climb into bed. Try reading something familiar, walking slowly, or listening to music you’ve heard a thousand times. The aim is to give your brain nothing new to chew on, so the default-mode network doesn’t ride into bed still active.

Paced breathing primer. Slow, paced breathing at roughly six breaths per minute has been shown in multiple studies to increase parasympathetic (vagal) tone and reduce sleep onset latency. A 2019 study in the Journal of Clinical Medicine by Laborde and colleagues found that 15 minutes of slow-paced breathing each evening, sustained over 30 days, significantly improved both subjective sleep quality and cardiac vagal activity compared with a control group using social media before bed. If you have never done paced breathing, try it for five minutes before you get in bed. Inhale for five seconds, exhale for five seconds. That’s the whole protocol; complexity isn’t required for the technique to work.

The 20-minute rule – what to do when you’re in bed but not asleep

If you have been in bed roughly 20 minutes and you are clearly not falling asleep – racing thoughts active, frustration rising, body restless – get up.

This is stimulus control, the technique Richard Bootzin developed in 1972 and refined over the following two decades. The mechanism Bootzin originally proposed was classical conditioning: every minute spent awake and frustrated in bed strengthens the association between bed and struggle, and weakens the association between bed and sleep. A 2024 systematic review in the Journal of Sleep Research found the original conditioning hypothesis is harder to support than once thought, but the technique itself remains one of the most validated components of CBT-I. Newer thinking suggests stimulus control works by disrupting the cognitive loop – getting up breaks the runaway thought spiral, and re-approaching bed when sleepy reattaches the room to its proper function.

The protocol is specific:

• Get out of bed. Move to a different room if possible.
• Keep the lights dim. No overhead lights, no screens.
• Do something boring and predictable – read a book you’ve already read, fold laundry, sit quietly.
• Return to bed when you feel genuinely sleepy. The cue is sleep itself approaching, not the clock hitting 20 minutes.
• If you climb back in and the same loop starts, get up again.

The hardest part is psychological. Getting out of bed at 12:40 a.m. feels like giving up. The reverse is true: lying in bed wide awake for an hour is what gives up. Stimulus control is the active move. Patients in CBT-I trials report that the first three nights of stimulus control are brutal, the next three are tolerable, and by the second week, the bedroom starts to feel like a place where sleep happens again.

In-bed techniques – what to do when getting up isn’t an option

Sometimes you cannot get up. You are traveling, your partner is asleep, the room is freezing, you have a 5 a.m. flight and getting out of bed feels worse than the alternative. Several in-bed techniques have research support, and they all share a common mechanism: they give the verbal-thinking brain a low-stakes, neutral task that crowds out the racing-thought loop without engaging it.

Cognitive shuffling (Serial Diverse Imagining). Developed by cognitive scientist Luc Beaudoin and validated in a 2016 randomized trial, the technique works like this. Pick a neutral seed word – blanket, for example. Take the first letter, B. Imagine an object that starts with B. A balloon. Hold the image for a few seconds, then move to another B-word. A bicycle. Hold. Then another. A banana. When you exhaust your associations, move to the next letter – L for blanket. Lamp, lemon, lighthouse. Then A, then N, and so on. The brain enters a state structurally similar to the fragmented, associative thinking that characterizes early sleep stages. Beaudoin’s theory is that the technique mimics the brain’s own pre-sleep cognitive pattern, which signals the sleep-onset system that it is safe to switch over.

Body scan. Start at the top of your head and move slowly downward. Forehead, eyes, jaw, throat, shoulders, chest, abdomen, and so on down to your feet. As you notice each region, deliberately release any tension you find. The technique is structurally similar to progressive muscle relaxation, but without the active tensing step that some people find activating. The body scan works because it shifts attention from verbal-narrative thought to body-mapped attention, which uses a different cortical circuit.

Paced breathing again, this time in bed. Inhale four seconds, hold one second, exhale six seconds, pause one second. Don’t aim for perfection. The slow exhale is the active ingredient – it is the part that engages the vagus nerve and downshifts the autonomic nervous system. If counting feels like one more thing to do, drop the counting and just breathe slowly and steadily.

The mechanism is the same across all three. Verbal thought – especially the self-referential kind – is the cognitive activity most tightly linked to default-mode-network engagement and pre-sleep arousal. Anything that gives the verbal-thinking brain neutral, low-stakes content reduces that engagement.

The bigger lever – rewriting the script that plays at bedtime

The techniques above are tactical. They help tonight. The deeper observation is that what runs at bedtime is rarely random. It is the same handful of scripts, replaying.

For most people, the bedtime mind is dominated by a few looping themes – the replay of an interaction at work where you wish you had said something different; the anticipation of tomorrow and a vague low-grade dread about its demands; some version of background self-talk that gets loud at night because the daytime distractions have stopped. In the research literature, this is sometimes called autobiographical self-referential processing. The default-mode network handles it. And here is the part that matters for what comes next: that script is rewritable.

A 2024 fMRI study by Jo and colleagues, published in Brain Sciences, examined what happens in the brain when people hear themselves – their own recorded voice – deliver self-referential content. The result was that 13 brain regions activated significantly (Z = 3.86 to 5.50), including bilateral medial prefrontal cortex, precuneus, and parahippocampal cortex. These are the self-referential hubs of the default-mode network – the same regions that won’t quiet down at bedtime in people with insomnia. The main effect was statistically robust (F(2,61) = 14.72, p < 0.001).

The practical implication: the bedtime brain isn’t a passive recipient of whatever script happens to be running. The script can be authored. People in coaching, therapy, and high-performance contexts have long known this – what they may not have known is that delivering the new script in your own voice routes it through the brain regions most tightly coupled to identity and belief.

This is the longer game. Whisperliminals exists to make it tractable: you record self-coaching scripts in your own voice, the audio plays through the hypnagogic window and into the night, and the brain works on the rewrite while you sleep. Over weeks, the script that plays at bedtime starts to shift – because you’ve been deliberately authoring it, not inheriting whatever runs by default.

When to escalate – and what to expect over weeks

If racing thoughts at bedtime are a frequent, weeks-long problem – not just an occasional bad night – the answer is CBT-I. Cognitive Behavioral Therapy for Insomnia is recommended by the American College of Physicians as the first-line treatment for chronic insomnia. Meta-analyses of more than 30 randomized controlled trials show response rates of 70–80%. A 2024 trial in Frontiers in Psychiatry found internet-delivered CBT-I non-inferior to face-to-face treatment in routine clinical practice, which makes the treatment newly accessible for people who cannot find a CBT-I-trained provider locally.

CBT-I isn’t relaxation training. The protocol typically runs six to eight weeks, and includes sleep restriction (deliberately consolidating your sleep window to rebuild sleep pressure), stimulus control (the 20-minute rule above), cognitive restructuring of the dysfunctional beliefs Harvey identified, and sleep hygiene as a smaller component than most people expect. Apps like SleepioRx (FDA-cleared in 2025) and CBT-i Coach (free, developed by the U.S. Department of Veterans Affairs) deliver the protocol digitally.

Sleep medication is rarely the right long-term answer. Most prescription sleep aids produce sedation without the brain-state changes that constitute real sleep, and most are recommended only for short-term use. Melatonin helps shift circadian timing but does not address the cognitive-arousal mechanism behind racing thoughts.

What to expect during a reset: most people who commit to a real protocol – constructive worry, stimulus control, paced breathing, consistent rise time – see meaningful change in two to four weeks, not two to four nights. Week one is often worse before it gets better, because stimulus control disrupts habits that, however dysfunctional, were stable. By week three the cognitive loop is usually quieter; by week six the new pattern is consolidating.

Tonight – a concrete plan

If you’re reading this at 11 p.m. and want one thing to do tonight: spend 10 minutes writing the worry list described above – worries in the left column, next concrete action in the right. Set the paper aside. In bed, do paced breathing for 10 minutes. If you’re still awake at the 20-minute mark, get up, sit in a different chair with dim light, and do nothing demanding until sleep is approaching. Then return to bed.

The science is dense, but the practice is short. The hardest part is giving yourself permission to stop fighting your own brain in bed, and to treat sleep as something built over weeks rather than caught night by night.

Frequently Asked Questions

How long is “too long” to lie awake before getting up? The standard CBT-I guideline is around 20 minutes, but most protocols suggest using your subjective sense rather than the clock – if you are wide awake and frustration is rising, get up, even if it has only been 10 minutes. Watching the clock to determine the right moment to get up is itself a form of the selective-attention monitoring that Harvey’s cognitive model identifies as arousing.

What’s the difference between worry and racing thoughts? Worry is typically focused on a specific concern and proceeds in a more linear (if repetitive) way. Racing thoughts move faster, jump from topic to topic, and feel uncontrollable. A 2021 Comprehensive Psychiatry paper argued for treating them as distinct constructs in insomnia research. Practically, the worry list technique handles the linear worry; in-bed techniques like cognitive shuffling handle the racing-thoughts pattern.

Does counting sheep actually work? The research suggests not – counting is too repetitive and monotonous to occupy the verbal-thinking brain effectively, which leaves the racing-thought channel free to keep running. Cognitive shuffling, which constantly switches to new associations, occupies the same channel that the racing thoughts use.

Can I do paced breathing in bed without making it worse? For most people, yes. A small minority of people experience increased anxiety with breathing-focused practices (interoceptive anxiety), in which case the body scan or cognitive shuffling are better alternatives. If paced breathing makes you more aware of your heartbeat or breathing in a way that feels worse, switch techniques.

What if I wake up at 3 a.m. with racing thoughts? Same protocol. Give yourself about 20 minutes. If sleep isn’t coming back and the thought loop is active, get up. Dim light, boring activity, return when sleepy. The bias to lie in bed and hope is the same trap as when you first tried to fall asleep.

Does alcohol help me fall asleep faster? Often yes for sleep onset, no for the night as a whole. Alcohol shortens sleep latency but fragments the second half of the night and suppresses REM sleep, when most memory consolidation happens. The next-day cognitive cost is usually higher than the falling-asleep benefit, especially for high-performers whose work depends on executive function.

Should I journal at bedtime? Evidence is mixed. Constructive worry done 90+ minutes before bed has good support. Journaling in bed can backfire because it activates the verbal-thinking circuits the bedtime brain needs to power down. If you want to journal, do it earlier.

Is it OK to read in bed? For some people, yes – boring or familiar reading can help. For people with chronic insomnia, CBT-I guidance is to keep the bed for sleep only: no reading, no working, no phone. The bed becomes a stronger sleep cue when it’s exclusively associated with sleep.

How long until CBT-I works? Most patients see measurable improvement in sleep efficiency by week three, and full protocol effects by weeks six to eight. The first week is often harder than baseline because stimulus control and sleep restriction deliberately consolidate sleep before the rebound takes hold.

Will Whisperliminals replace CBT-I? No, and that’s not the framing. CBT-I addresses the cognitive and behavioral architecture of the insomnia loop. Whisperliminals addresses what plays at bedtime – the script that runs through the hypnagogic window and into sleep, in your own voice. They are complementary. People in CBT-I who also use audio self-coaching often report that the cognitive restructuring lands deeper because the new content is replayed during the brain’s most receptive window.

If you want help with the script that plays at bedtime

Most of the techniques above work on the surface. They distract the verbal brain, slow the breath, break the bed-as-struggle association. The deeper move is changing what the verbal brain replays in the first place.

Whisperliminals is built for that deeper move. You record self-coaching scripts in your own voice – the content you want your brain to absorb instead of the looping replays it produces by default. The audio plays through the hypnagogic window and into the night while you sleep.

Start at app.whisperliminals.com/start.

References

• Bootzin, R. R. (1972). Stimulus control treatment for insomnia. Proceedings of the American Psychological Association, 7, 395–396.
• Nicassio, P. M., Mendlowitz, D. R., Fussell, J. J., & Petras, L. (1985). The phenomenology of the pre-sleep state: The development of the pre-sleep arousal scale. Behaviour Research and Therapy, 23(3), 263–271.
• Harvey, A. G. (2002). A cognitive model of insomnia. Behaviour Research and Therapy, 40(8), 869–893.
• Beaudoin, L. P., Digdon, N., O’Neill, K., & Racour, G. (2016). Serial Diverse Imagining Task: A new remedy for bedtime complaints of worrying and other sleep-disruptive mental activity.
• Laborde, S., et al. (2019). Influence of a 30-day slow-paced breathing intervention compared to social media use on subjective sleep quality and cardiac vagal activity. Journal of Clinical Medicine, 8(2), 193.
• Palagini, L., et al. (2021). Investigating racing thoughts in insomnia: A neglected piece of the mood-sleep puzzle? Comprehensive Psychiatry, 107, 152232.
• Marques, D. R., et al. (2023). Functional connectivity of the default mode network predicts subsequent polysomnographically measured sleep in people with symptoms of insomnia. Sleep.
• Jo, H., et al. (2024). Neural effects of one’s own voice on self-talk for emotion regulation. Brain Sciences, 14(7), 637. DOI: 10.3390/brainsci14070637
• Jansson-Fröjmark, M., et al. (2024). Stimulus control for insomnia: A systematic review and meta-analysis. Journal of Sleep Research. DOI: 10.1111/jsr.14002
• van der Zweerde, T., et al. (2024). Clinical effectiveness of internet-based cognitive behavioral therapy for insomnia in routine secondary care: results of a randomized controlled trial. Frontiers in Psychiatry, 15, 1301489.