Sleep restriction therapy is a core part of CBT-I, the first-line treatment for insomnia. It treats broken sleep by limiting how long you spend in bed each night. That sounds backwards — you fix your sleep by spending less time in bed. But it works: if you spend nine hours in bed and only sleep six of them, your time in bed shrinks to about six hours. That shorter window makes you sleepier at night, your sleep stops breaking apart, and the window widens again as your nights improve. The American Academy of Sleep Medicine suggests it can be used on its own.
For the first few weeks it deliberately makes you sleepier — that extra sleepiness is the engine of the treatment.
Sleep pressure builds the whole time you are awake and gets spent when you sleep. Chronic insomnia usually comes with a set of sensible-looking responses that quietly bleed that pressure away: an early bedtime after a bad night, a long lie-in on Saturday, a nap when the afternoon collapses. Each one reduces the drive you need at bedtime. Compressing the window does the opposite — it concentrates the pressure into fewer hours, so you fall asleep faster and stay asleep longer inside a shorter window. Then the window grows.
The other half is that a compressed window kills the dead time. Eight hours in bed for six hours of sleep means two hours of lying there, and those two hours are what teaches your brain that bed is a place where you are awake. That is the same problem stimulus control attacks from a different angle, which is why the two are usually delivered together inside CBT-I.
Your clinician sets it from the numbers in your own sleep diary, not from how tired you feel. A 2025 paper in the Journal of Behavioral and Cognitive Therapy set out standard definitions for how this is actually implemented, and the sequence below is theirs. Read it to understand what your clinician is doing — not to set your own window, because the whole point is that the numbers are yours and the judgement calls are theirs.
Two things stand out in that sequence. The floor is real — five hours, and the original protocol never went below 4.5 — so "sleep restriction" does not mean sleeping as little as possible. And the expansion criterion is sleep efficiency rather than how you feel, which is why the schedule can widen on a week when you are still tired.
Yes. Used by itself, it helps people fall asleep faster, wake less during the night, and spend more of their time in bed actually asleep — a review of the standalone evidence found moderate-to-large weighted effect sizes for sleep onset latency, wake after sleep onset, and sleep efficiency after therapy, from the four trials with adequate methodological strength. What it does not reliably do is add hours: total sleep time barely moved, which is true across the CBT-I research — the treatment makes your sleep more solid, not longer.
The AASM suggests clinicians use sleep restriction therapy as a single-component therapy for chronic insomnia in adults, a conditional recommendation. Multicomponent CBT-I is the only strong one.
And there is a practical argument for the standalone version. In a randomized trial of 150 postmenopausal women, sleep restriction was delivered in two sessions with interim phone contact against six sessions for full CBT-I, which makes it a viable route for someone who cannot complete a full course. The same review notes that variability in how clinicians implement it makes strong conclusions harder, and that measures of daytime functioning are thin — more research needs to be done in this area.
Because for a couple of weeks, it is worse, and the evidence on how much is genuinely mixed.
The study that measured it most closely put 16 patients with well-defined insomnia through four weeks of single-component sleep restriction with overnight lab recordings. Objective total sleep time dropped by an average of 91 minutes on the first night of treatment, 78 minutes by night 8, and 69 minutes by night 22, relative to baseline. Reaction-time lapses increased and reaction times slowed at most assessment points during the acute phase, and Epworth sleepiness scores were up in weeks 1, 2, and 3. All of it returned to baseline by three months, and the subjective sleep outcomes improved throughout — but that is real, measured daytime impairment while the schedule is compressed.
Two later studies did not find it. A driving-simulator study reported no significant change in daytime sleepiness, reaction time, or simulated driving performance during the acute phase, with the caveat that its participants held their total sleep time roughly steady and there was no control group. And the randomized trial in postmenopausal women found the odds of becoming excessively sleepy after sleep restriction were no different from the odds of not becoming sleepy, and no worse than after full CBT-I.
The risk looks transient where it shows up at all, and it is not settled. The 2025 best-practice paper takes the conservative line: patients should be forewarned about drowsiness and safety risks that can arise while implementing sleep restriction. If you drive for work, operate machinery, or your job depends on sustained vigilance, that is a conversation to have with your clinician before the window is set, not after.
Compressing time in bed assumes you can leave the bed, and for some people that assumption does not hold. The best-practice definitions name four situations where the standard instruction needs changing: limited mobility, a restricted environment such as a dorm, barracks, or prison, chronic pain that requires extended time in bed, and chronic fatigue or a comorbid disease. For those, the recommended substitute is counter control — staying in bed but out of a sleeping posture, sitting up without the pillow and blanket, so the bed still stops being a place you lie awake in.
That list is about circumstances rather than diagnoses, and it is not the whole picture. Whether a psychiatric or neurological condition changes the calculus is an assessment question, which is the argument for having the schedule set by someone who does the assessment. A clinician can also spot the reason a schedule stalls that no amount of compression fixes: untreated sleep apnea.
There is a middle path between setting your own window and booking weekly appointments. Prescription digital CBT-I applies the same titration rules — the diary, the floor, the 85% threshold — without you having to hold the arithmetic, and a clinician orders it after reviewing your intake. The safest version is still one where a clinician sets the starting window and adjusts it. At the RISE Sleep Clinic that is the same intake either way: you can work with a sleep psychologist, start with the digital program, or move between them. What you pay depends on your plan.
Sleep restriction works on sleep pressure, which is the same quantity sleep debt measures — the difference is direction. Sleep debt tracking exists to help you keep pressure low; sleep restriction deliberately drives it up for a few weeks so that your sleep consolidates, then lets it settle. Both rest on the same fact: sleep pressure is real, it accumulates, and what you do in the daytime spends it.
Timing is the piece a compressed window cannot fix on its own. A schedule that sits against your circadian rhythm asks you to sleep while your body clock is running its wake signal, and no amount of pressure makes that comfortable.
RISE makes it easy to improve your sleep and daily energy to reach your potential
Yes, for what it targets. The standalone evidence shows moderate-to-large effects on how fast you fall asleep, how long you're awake in the night, and sleep efficiency. Total sleep time improves only slightly. The AASM suggests it can be used as a single-component therapy for chronic insomnia, a conditional recommendation.
Your clinician calculates it from your sleep diary: the standard rule sets your window at the average total sleep time from your own sleep diary, with a floor of five hours — the original protocol never prescribed less than 4.5. Guessing at it is how people end up compressed too far.
Weeks. In the lab study that tracked it night by night, subjective sleep and insomnia severity improved across a four-week protocol while objective sleep time was still down. The window starts widening as soon as your sleep efficiency reaches 85% over about five days, which for many people is sooner than they expect.
Yes, for most people — the one real cost is feeling sleepier during the day for the first few weeks. One lab study measured that cost: more daytime sleepiness and slower reaction times early on, gone by three months. Two later studies looked for the same problem and did not find it. Clinicians are still told to warn patients about drowsiness, so if you drive for work or operate machinery, say so before your window is set.
Partly — and there's a safer version of "on your own" than setting your own window. This is the component where getting the numbers wrong costs most: the five-hour floor, the 85% efficiency threshold, and the decision to widen or narrow are judgement calls made against your own diary. Prescription digital CBT-I applies those rules for you, ordered by a clinician who has reviewed your intake.
Usually because sleep pressure has run low by the small hours while your sleep is at its lightest and cortisol is climbing. That's exactly the pattern a compressed window targets. If it's happening three or more nights a week for three months, it meets the threshold for sleep-maintenance insomnia.
