why your brain won't switch off at night
You got into bed tired. Properly tired — the kind where your eyes were closing on the sofa an hour ago.
Then your head hit the pillow and something switched on.
Tomorrow’s meeting. That email you should have phrased differently. A conversation from 2019. And underneath all of it, the arithmetic: it’s 12:40, if I fall asleep now I get six hours and twenty minutes.
You are not bad at sleeping. You are stuck in a loop that has a name, a mechanism, and a treatment with a stronger evidence base than almost anything else in mental health.
The trigger is usually long gone
Chronic sleeplessness rarely stays about the thing that started it.
The clinical model most sleep researchers work from separates what started the problem from what maintains it. Something sets it off — a stressful quarter, an illness, a new baby, a house move. That thing passes. But by then you’ve adopted a set of reasonable-sounding responses: going to bed earlier to catch up, lying there longer hoping it’ll come, napping when you can, checking the clock to see how bad tonight is going to be.
Those responses are what keep it going.
The bed is the clearest example. Sleep is heavily conditioned — your brain learns what a place is for. Spend enough nights lying awake, anxious, scrolling, doing sums about tomorrow, and the bed quietly stops being a sleep cue and becomes a wakefulness cue. Some people notice they fall asleep instantly on the sofa and snap awake the moment they get into bed. That isn’t in their head in the dismissive sense. It’s classical conditioning, working precisely, against them.
Trying is the problem
Here’s the part that sounds like a riddle and isn’t.
Sleep is not a behaviour you perform. It’s a state you stop blocking. Every other route to it — wanting it, chasing it, monitoring your progress toward it — raises arousal, and arousal is the one thing that reliably prevents it.
Which means effort applied directly to sleeping makes sleeping harder. Not metaphorically. Mechanically.
This is why “just relax” fails so consistently. It converts sleep into a task with a pass mark, and hands you something you can fail at, at midnight, alone, with no way to try harder. Anyone who has lain there doing a breathing exercise and thinking this isn’t working has run the experiment.
Everything that actually works, works indirectly — on the clock, on the association with the bed, on the daytime, on your relationship with the worrying. Never on the sleeping itself.
What the evidence actually supports
The treatment is called cognitive behavioural therapy for insomnia, or CBT-I. It isn’t one technique — it’s a package, and the components carry the effect.
In the UK, NICE recommends CBT-I as first-line treatment for insomnia, before or alongside medication. Not sleep hygiene. Not a prescription. That’s the guidance.
It also works without a therapist in the room. A 2025 meta-analysis in npj Digital Medicine pooled 29 randomised controlled trials covering 9,475 participants and found fully automated digital CBT-I produced moderate-to-large reductions in insomnia severity. Worth being straight about the limits: it was still less effective than CBT-I with a real therapist supporting it, and heterogeneity across the trials was high. But the effect is real and it survives being delivered through a screen.
The gap is access. NICE’s own implementation resource notes that insomnia disorder is among the ten most common reasons people see a GP, that sleep loss costs the UK an estimated £34 billion a year, and that one specialist service’s waiting list now runs twelve to fifteen months. Most people asking their GP about sleep get sleep hygiene advice — which is the weakest ingredient in the package, handed out on its own.
Four things that carry most of the effect
None of these are relaxation techniques. That’s deliberate.
Fix your wake time, not your bedtime. Getting up at the same time every day — including weekends — is what anchors the body clock. Bedtime is downstream of it. Going to bed earlier after a bad night feels like repayment and usually backfires: more time in bed spread over the same amount of sleep produces thinner, more broken sleep and more hours awake in a place you’re trying to re-associate with sleeping.
Get out of bed. If you’ve been lying awake long enough to be annoyed about it — roughly fifteen or twenty minutes, judged by feel, not by looking at the clock — get up. Go somewhere else, keep the lights low, do something undemanding, and come back when you’re sleepy rather than when you think you should. This is stimulus control, and it’s the component that directly attacks the conditioning. It feels absurd the first few nights. It is the single highest-yield thing on this list.
Do the worrying earlier, on purpose. Set aside fifteen minutes in the early evening — not at bedtime — and write down what’s circling and the next physical action for each one. The point isn’t resolution. It’s that a worry which has been looked at directly, in daylight, with a pen, has less to say at 1am. The timing is what makes it work.
Stop measuring. See below.
Your sleep tracker may be part of the loop
This is the uncomfortable one, and it’s where most of the wellness industry goes quiet.
Researchers at the University of Bergen surveyed 1,002 adults in Norway about sleep app use, published in Frontiers in Psychology in 2026. Forty-six per cent had used one. Overall the picture was mixed — some people genuinely benefited. But people already showing insomnia symptoms were markedly more likely to report negative effects, and younger users felt the feedback more strongly in both directions.
Read that against the mechanism above and it stops being surprising. If the problem is that you’re monitoring your sleep too closely, a device that grades it every morning and shows you a chart of your failures is feeding the thing it claims to fix. Clinicians have a name for the extreme version — orthosomnia, the pursuit of perfect sleep metrics making sleep worse.
Two honest caveats. This was a self-report survey, not a trial, so it establishes an association rather than proving the app caused the harm. And it doesn’t mean trackers are useless for everyone. It means that if you are anxious about your sleep, the tracker is probably not neutral, and turning it off for a fortnight is a cheap experiment.
While we’re here: consumer sleep-stage scoring is a rough estimate, not a measurement. You are lying awake at 3am worrying about a number that was inferred from your wrist movements.
What to do tonight
Nothing that can be failed.
Get up at your normal time tomorrow regardless of how tonight goes. That’s it — that’s the whole intervention for night one, and it happens in the morning, which is the point. There’s nothing to perform at bedtime.
If you’re awake and irritated tonight, get up for a while. If you’re not, don’t.
You’ll probably have a bad night at some stage this week. Nothing has gone wrong when that happens. Waking in the night is not automatically a disorder — the panic that follows the waking usually does more damage than the waking did.
When this needs a doctor rather than an article
Please see a GP if sleeplessness has run three nights a week for three months or more, if you snore heavily or your partner has noticed you stopping breathing or gasping — that needs assessing for sleep apnoea before anything else — or if low mood, anxiety or alcohol are tangled up in it. This article is general information, not medical advice, and no app is a substitute for an assessment.
Where fullfill fits
We built Restful Nights on this literature rather than on atmosphere. Twenty-one short sessions covering stimulus control, the arousal loop, worry timing and morning light — not twenty-one variations on breathing slowly.
It teaches, then gets out of the way. No streaks. No score in the morning. Nothing to fail.
The founder wrote about the loop he couldn’t switch off — sleep was the first thing it took. This is the part of the answer that took longest to get right.
If you’d like to try it, download the app.