Learn › CBT-I

Sleep restriction therapy: why spending less time in bed can mean more sleep

Sleep restriction therapy is the deliberate, temporary narrowing of the hours you allow yourself to be in bed, so that more of the time you spend there is actually spent asleep. It works by concentrating sleep into a shorter window, which builds sleep pressure — for many people that means falling asleep faster and staying asleep longer — and by stopping your bed from being a place where you lie awake. The name is misleading: the goal is not less sleep, it is less wakefulness in bed. The first week or two is genuinely hard, and for many people it starts to ease after that. It is also a technique with real contraindications, so before you change anything about your schedule, read the safety section below.

What it actually is (and what it isn't)

Sleep restriction therapy (SRT) was developed by Arthur Spielman and colleagues and first published in 1987. It is one component of CBT-I — cognitive behavioural therapy for insomnia — alongside stimulus control, cognitive work on sleep beliefs, and relaxation methods.

It is not sleep deprivation for its own sake, and it is not a sleep hygiene tip. If you lie in bed for eight and a half hours and sleep five and a half of them, SRT does not take sleep away from you. It takes away the three hours of lying there awake, then gives them back gradually as your sleep proves it can fill them.

In the American Academy of Sleep Medicine's 2021 guideline, multicomponent CBT-I carries the only strong recommendation among the behavioural treatments; sleep restriction on its own gets a conditional one. The technique tends to do better inside a full programme than as a standalone trick.

Why shrinking the window works

Three things happen at once, and they reinforce each other.

• Sleep pressure builds. The longer you have been continuously awake, the stronger your biological drive to sleep, so sleep tends to come faster and deeper.
• Sleep consolidates. Broken, fragmented sleep tends to knit back into one solid block when there is less room for it to spread out.
• The bed stops meaning "awake". Months of lying in bed frustrated teach your brain that bed is a place for alertness and worry. Cutting the awake-in-bed time weakens that association — which is also what stimulus control is for.
• A fixed wake time anchors your body clock, because the window is set from a consistent morning wake-up seven days a week.
• It breaks the compensation loop. Going to bed at nine after a bad night, or lying in until eleven at the weekend, dilutes sleep pressure and tends to make the next night worse.

How your sleep window gets set

Before you set a window at all, check whether this is safe for you. Sleep restriction is not suitable for everyone without medical advice — if you have a seizure disorder, bipolar disorder, a parasomnia such as sleepwalking or night terrors, possible untreated sleep apnoea, narcolepsy or unexplained heavy daytime sleepiness, a safety-critical job, a raised risk of falls, or you are pregnant or managing another significant medical or psychiatric condition, talk to a clinician before you change your sleep schedule. The full list is in the safety section below.

Assuming none of that applies, there is no shortcut past the diary — the prescription comes from your own data, not a general rule. You log one to two weeks: roughly what time you got into bed, roughly how long you were awake, what time you got up. Fill it in the next morning from rough estimates rather than during the night; clock-watching at 3am works against you, and precision matters less here than consistency.

Then the window is built from that average. In Spielman's protocol, your prescribed time in bed starts at your average total sleep time. A gentler and also evidence-based variant (Edinger and Carney) sets it at your average sleep time plus 30 minutes. The wake time is fixed first — no later than your usual workday wake-up — and the bedtime is counted backwards from there.

There is always a floor, and it is the most safety-relevant number in the whole protocol. Spielman's modified protocol does not go below 5 hours in bed; his original 1987 study used a 4-hour-30-minute lower limit. Many clinical protocols use 5 hours 30 minutes, and some consumer guidance sets 6. When nobody is reviewing you week to week, the more conservative floor is the safer choice: if the arithmetic says less, use the floor instead. Quitude never prescribes less than 5 hours 30 minutes in bed — that floor was raised from 5 hours precisely because Quitude is unsupervised self-help with no clinician watching your week.

A worked example: you average 5 hours 20 minutes of sleep across 8 hours in bed. That is a sleep efficiency of about 67%. With a 5-hour-30-minute floor and a 6:30am wake time, your window is 1:00am to 6:30am — for now.

Which convention Quitude uses

Since those two conventions point in different directions, here is exactly what the app does. Quitude uses Spielman's convention, reviewed once a week on 7 days of diary data: 90% or above widens your window by 15 minutes, 85% to just under 90% holds it, and below 85% tightens it by 15 minutes — never past the 5-hour-30-minute floor. It shows you the efficiency number behind each change.

One limitation before you rely on it: Quitude applies those adult thresholds to everyone and does not adjust them by age. If you are over about 65, the source protocol would widen your window at 85% rather than 90%, so the app will hold you at a tighter window than the age-appropriate convention calls for. Worth raising with a clinician, particularly if you are also managing balance or fall risk.

What the first hard weeks actually feel like

The first week or two is usually the hardest stretch. Expect daytime sleepiness, a foggy or slow feeling, lower patience, sometimes headaches or a flat mood. That is the predictable cost of the method rather than a sign it is failing — but it is also a genuine safety issue while it lasts, which is why the rules below are not optional.

Do not drive or operate machinery when you feel sleepy. That applies to every reader, not only people who drive for a living: in Kyle and colleagues' study of sleep restriction, more than one in three participants reported that their driving ability had been compromised during treatment. If you are too sleepy to be safe, that outranks the protocol every time.

The strongest urge will be to go to bed early "just tonight". That is the sleep pressure you are trying to accumulate. A few things that make these weeks more manageable:

• Keep the wake time fixed, weekends included. It is the anchor for everything else.
• Plan the late evening in advance — a lit room, something mildly engaging, out of the bedroom and off the sofa. Boredom at 11pm is what sends people to bed early.
• No naps is the standard rule during the restriction phase, because a nap discharges the sleep pressure you are building. Safety is the exception, and it wins: if you are too sleepy to be safe — especially before driving — take a short nap early in the day rather than pushing through, then widen your window and get advice. Never trade alertness behind the wheel for protocol purity.
• Get bright light and some movement in the first hour after waking.
• Keep caffeine steady and early rather than escalating it through the day — late caffeine will directly undercut the sleep you are trying to consolidate.
• When it works, the pay-off usually arrives as consolidation first: fewer awakenings, faster sleep onset, sleep that feels more like one block. The window widening comes after that, 15 minutes at a time. Some people take longer than others, and some do not respond.

When to stop and get help

Two rules apply to every reader, whatever your health. First: do not drive or operate machinery when you feel sleepy — everyone on the programme, not only people who drive for a living. Second: add time back to your sleep window and speak to a clinician, rather than pushing through, if any of the following happen.

Escalate if you fall asleep unintentionally during the day, if you nod off at the wheel or at work, if your mood becomes elevated, agitated or unusually energised, or if your mood drops significantly. None of those is the treatment working, and none is something to grit your teeth through. Widening the window is always available to you, and so is stopping.

If your mood drops badly, or you feel hopeless or have thoughts of harming yourself, please do not wait for a routine appointment. In the US you can call or text 988 for the Suicide and Crisis Lifeline; elsewhere, contact your local crisis line or emergency services. This page is general information, not medical advice, and an app cannot examine you. If your sleep problems persist, or you are unsure whether any of the above applies to you, see a clinician.

Why the hard weeks are worth considering

The case for putting yourself through this is durability. In a systematic review and network meta-analysis by Furukawa and colleagues, at a median follow-up of 24 weeks, long-term remission was 41% (95% CI 31–53) for CBT-I versus 28% for sleeping medication — an odds ratio of 1.82 (1.15–2.87), rated high certainty. Dropout was lower for CBT-I too: 21% versus 39%.

Those are population figures, not a forecast for you. Remission is also a strict all-or-nothing threshold: plenty of people who never quite reach it still sleep meaningfully better than when they started, and some people do not respond at all. Anyone telling you a behavioural programme is a guaranteed fix, or a cure, is overselling it.

Combining CBT-I with medication showed no clear advantage over CBT-I alone in that analysis (odds ratio 1.07, 0.63–1.80) — a conclusion the AASM's 2026 combination-treatment guideline reached independently, conditionally suggesting against combining the two rather than doing CBT-I alone. That guideline also conditionally favours combination over medication alone, and notes that someone whose priority is more total sleep early in treatment might reasonably choose to combine them. That is a conversation for a clinician, not an app.

From the medication side, the AASM's 2017 pharmacologic guideline made 14 recommendations and graded every one of them weak, with evidence quality mostly low or very low. It states that chronic hypnotic use should be reserved for people for whom CBT is inaccessible or ineffective. The plain framing: pills work while you take them and generally act faster in the first week or two, whereas CBT-I teaches skills that, for many people, keep working after the programme ends.

If the full cut is too much: gentler variants

Cutting straight to your average sleep time is the fastest version, not the only one. Two established, evidence-based alternatives trade speed for tolerability, and either is reasonable if you are worried about the sleepiness or your days have no slack in them.

Sleep compression (Riedel and Lichstein) narrows the window gradually rather than all at once: at the first step you cut the gap between your baseline time in bed and your baseline sleep time by half, then at each of the next two steps you cut it by a further quarter of that same original gap — closing the gap over about three sessions. The plus-30-minutes variant simply starts the window a little wider than your average sleep time. Both take longer to consolidate sleep, and both beat abandoning the method in week one.

Doing this without a clinician

The technique is not intellectually complicated; the bookkeeping is where people fall off. A week or two of diary, a rolling average, an efficiency percentage, a weekly rule applied consistently, and a floor you never cross — miss any of those and the method quietly stops being the method.

That is the part software is genuinely good at. Quitude's 7-week CBT-I programme does the arithmetic for you: it builds your window from your own diary, reviews it weekly on 7 days of data, holds a 5-hour-30-minute floor, and shows you the numbers behind each change. It is the free part of the app, not a paid tier, because a programme that only works if you finish it should not have a paywall in the middle of week two.

One thing to be clear about, because it matters more than any feature list: the trial results above are for CBT-I as delivered in research settings — mostly by clinicians, or through digital programmes such as Sleepio and Somryst that were tested in their own trials. Quitude follows the same published protocol, but Quitude itself has not been tested in a clinical trial. Treat it as a structured way to run the method, not as a studied treatment in its own right, and not as evidence about what will happen for you.

It is also not clinician supervision. Before you start, go back through the safety list above. If any of those conditions apply, if your sleep does not improve, or if something simply feels wrong, an app is not the place to sort that out — a doctor is.

Frequently asked questions

Keep reading

• Stimulus control for insomnia: rebuilding the link between your bed and sleep

  Stimulus control retrains the bed to mean sleep, not lying awake: the rules, the 20-minute get-up rule, who should avoid it, and the mistakes that undo it.

• Why do I wake up at 3am and can't get back to sleep?

  Waking at 3am is mostly normal sleep architecture — the trouble is what happens next. Why it happens, and the CBT-I approach that tends to help.

• CBT-I vs sleeping pills: what the evidence actually says

  CBT-I vs sleeping pills: how they compare short and long term, what sleep medicine guidelines actually say, and who each option suits.