What is CBT for insomnia, and what would it ask of me?

It may be 3am as you read this, after another night of watching the clock, or the morning after, with the day already looking long. Being told yet again to try CBT‑I can sound like being told it's all in your head, or like being asked to sleep even less when you're running on so little. It does ask for less time in bed at first, and there's a good reason for that. If you're awake now, one small thing is enough for tonight.

If your nights are short because the hours simply aren't there, because of a baby, someone you care for, night shifts or long days, this piece is about something else, and its advice would only take more from you. The insomniainsomnia Trouble falling asleep, staying asleep or waking too early, that keeps happening and affects how you feel or get on during the day. CBT‑I treats means trouble sleeping at least three nights a week for three months or more, despite enough time and the right conditions to sleep⁵5 Edinger et al., 2021b. And if your bad nights began only a few weeks ago, after something hard, you're probably not there yet. By the usual account, a short spell like that needn't become a long one unless certain coping habits keep it going⁶6 Patel et al., 2018. Those habits are what the rest of this is about, so you can read on knowing what to watch for.

CBT‑I is a short, practical course in which most of what it asks of you happens at the edges of the night: when you go to bed, when you get up and what you do with the time you'd otherwise spend lying awake. It brings together a little learning about how sleep is regulated, a set window in bed, getting up when sleep won't come and some work on the thoughts that keep you awake, with a sleep diary kept all the way through³3 Edinger et al., 2021. The window is called sleep restrictionsleep restriction A part of CBT for insomnia that limits time in bed to about the amount of sleep you're getting at the moment, then widens it as sleep becomes more solid. It's usually guided by a clinician or a tested program. and getting up belongs to stimulus controlstimulus control A part of insomnia treatment that helps the bed feel linked with sleep again: going to bed when sleepy, and getting up for a while when sleep won’t come., and each has more to it than its name suggests.

Of all the ways to treat insomnia that has gone on for months, this is the one with the strongest backing. The guideline American sleep doctors work from gives it its only strong recommendation for chronic insomnia in adults³, and that's true if you live with other health conditions too, and whether it's given in person or digitally¹²1 Qaseem et al., 20162 Riemann et al., 2023.

If you've already cut caffeine, darkened the room and kept the phone away from the bed, you've been following sleep hygienesleep hygiene Everyday habits often suggested for better sleep, such as keeping caffeine to the first part of the day, sticking to regular hours and sleeping in a dark room away from noise., which is sensible advice and a different thing. When researchers took CBT‑I apart to see which pieces made the difference, the hygiene advice wasn't one of them: it added nothing clear, while changing unhelpful thoughts, newer acceptance-based parts, the window in bed and getting up mattered most⁷7 Furukawa et al., 2024. Offered on its own, it did less well than active treatments, and the same guideline suggests not relying on it for chronic insomnia³3 Edinger et al., 2021. So if the hygiene advice hasn't been enough, that's what the trials would lead anyone to expect, and it says nothing about how hard you've tried.

Perhaps it began with something you can name, a loss or a hard winter at work, and that has eased while the nights stayed bad. Chronic insomnia often lasts years, fewer than half of people recover without treatment, and it tends to become a problem in its own right, whatever set it off⁵5 Edinger et al., 2021b.

What keeps insomnia going is mostly what anyone would do. After a few rough nights you go to bed a little earlier to make up for them, lie in when the morning allows, nap if you can and stay in bed trying, since getting up feels like giving in. In the model most clinicians teach from, extra time in bed, napping and a growing dread of bedtime are the habits that turn a short spell of insomnia into a long one, and they're what the behavioral parts of CBT‑I set out to undo⁶6 Patel et al., 2018. It's a working model more than a tested finding.

Perhaps you doze off on the sofa in the evening and then lie wide awake the moment you get into bed, sleepy everywhere except the one place you're meant to sleep. Night after night of lying awake there, the bed comes to mean wakefulness, and stimulus control is designed to break that link³3 Edinger et al., 2021. Meanwhile the hours in bed stretch while the sleep inside them stays much the same, so it's spread thin: in a small study in Glasgow, people with long-standing insomnia were spending about eight hours in bed and sleeping for about sixty-eight percent of it⁸8 Kyle et al., 2014.

This is the part that sounds hardest when you're already worn out. You and your therapist, or the program if you're working through one, start from your diary: roughly how long are you sleeping now? You set your time in bed close to that³3 Edinger et al., 2021, with the same getting-up time every morning. Each week you look at the diary together and decide whether to let the window out step by step as sleep comes to fill it, or draw it in a little when it doesn't³. It doesn't go below a floor, such as five hours, or four and a half in some programs⁸⁹8 Kyle et al., 20149 Ellis et al., 2015.

The thinking is that staying up until the window opens lets sleep pressuresleep pressure The body’s need for sleep, which builds the longer we stay awake and eases while we sleep. build. It may also steady the body clock and turn down the alertness that keeps you awake⁸¹⁰10 Maurer et al., 2018. How it does all that is still an open question, because the evidence for the mechanism was thin and mostly from poor-quality studies¹⁰. What most studies saw is clearer. A review of fifteen of them found that in most, time in bed shrank and sleep came more quickly once people lay down¹⁰, which is the change you'd be hoping for.

So what are those first weeks like? That same Glasgow study is one of the few that set out to measure them. Its sixteen volunteers agreed to four weeks of sleep restriction, and in the first week they went from about eight hours in bed to a little under six⁸8 Kyle et al., 2014. The first night was the roughest: they slept about an hour and a half less than usual⁸. For the next three weeks they were sleepier than usual during the day⁸.

The guideline describes the same early stretch for CBT‑I as a whole. Tiredness, irritability and trouble concentrating are its main downsides early on, and they usually ease by the end of treatment³3 Edinger et al., 2021. Until they do, the shorter nights and the sleepiness can make it hard to keep going³.

Please don't drive or use heavy machinery on days when you feel very sleepy during these weeks. In an earlier study by the same team, more than a third of people keeping audio diaries said their driving was affected, and one said, “I felt, really, I was a danger on the road.”⁸

By three months the people in Glasgow were back to their usual daytime alertness⁸. They were asleep for about ninety-one percent of their time in bed, and sleeping about fifty-six minutes longer each night than before they began⁸, in a window that had widened to make room for it.

Is that worth a few hard weeks? The guideline's authors set the early tiredness against what follows and judged that the benefits strongly outweigh it³. On this evidence that looks right. The weeks are still hard, and it's better to go in knowing they will be.

For a few people the window is something to do only with a clinician beside them. Guidelines say sleep restriction may not suit people prone to mania or hypomania, such as those with bipolar disorder; people whose seizures aren't well controlled; people already very sleepy in the day; or people in jobs such as driving or operating heavy machinery³3 Edinger et al., 2021. The caution about mania has trial evidence behind it: in a large trial of UK university students, a digital sleep course brought a small, lasting rise in manic symptoms¹¹11 Freeman et al., 2017 alongside its benefits. The same guideline expects that most people, once the reasons are explained, would choose sleep restriction with a clinician keeping an eye on how it goes³.

The other half of the behavioral work asks you to go to bed only when sleepy, get up when you can't sleep, keep the bed for sleep and sex, rise at the same time each morning and let daytime naps go³3 Edinger et al., 2021. In one large trial in Quebec, people went to another room whenever they hadn't fallen asleep, or back to sleep, within about twenty minutes, and came back to bed only once sleepy again⁴4 Morin et al., 2009. The twenty minutes is a rule of thumb from that trial's protocol, so a rough sense of time is plenty and the clock can stay turned away. If you're at risk of falls, find moving about difficult or take sleeping pills³, getting up is worth adapting with whoever is guiding you.

After months of this, sleep can start to feel like an exam you sit every night, and the cognitive part of CBT‑I works on the thinking that grows up around it. It looks at beliefs such as needing a full eight hours every night, or putting every hard day down to the night before, which feed the habits and the pressure to perform at bedtime³⁴3 Edinger et al., 20214 Morin et al., 2009.

One leading theory traces how that worry works on the night. Worrying about sleep and about the day after raises your alertness and sets you watching for signs that sleep isn't coming¹²12 Harvey, 2002. Watched that closely, a night tends to seem shorter than it was, and the growing anxiety can then cost real sleep¹². What happens if you give someone a clock to watch? In small experiments, researchers tried it, and people asked to watch a clock while they tried to fall asleep worried more and took longer to drop off than those who didn't¹³13 Tang et al., 2007. It happened to good sleepers as well as to people with insomnia¹³, and that's the striking part: nobody needed insomnia for the watching to cost them sleep. If you've spent nights checking the time, it may sound familiar.

Effort seems to work the same way. Sleep can't be put fully under willpower, so trying hard may itself help keep insomnia going¹⁴14 Broomfield & Espie, 2005. In a laboratory study, good sleepers asked to fall asleep as fast as they could while their minds were kept busy took longer than those left to drop off whenever they liked¹⁵15 Ansfield et al., 1996. The theory and these experiments are all early and small, but they point the same way: the harder a night is watched and worked at, the longer sleep may take.

So part of what CBT‑I offers is leave to stop working so hard at sleep. The same analysis that found little in the hygiene advice put the newer acceptance- and mindfulness-based parts among those that helped most⁷7 Furukawa et al., 2024.

Between sessions, a course asks less of you than it may sound. Much of it is a few lines in your diary each morning: when you went to bed, about how long you slept and when you got up. At each session the diary guides the window for the week ahead³3 Edinger et al., 2021. Most courses run four to eight visits, and the improvement comes gradually, which can be discouraging³, though what you gain lasts after the course ends³.

Often the nights get better from the inside before they get any longer. In trials of face-to-face CBT‑I for insomnia that came without other health conditions, people were falling asleep about nineteen minutes sooner by the end of treatment and spent less of the night awake, while their total sleep had barely changed¹⁶16 Trauer et al., 2015.

While the window is short there are evening hours to spend somewhere other than bed, and some people, older adults or those living with pain or depression among them, find that extra time hard to fill³3 Edinger et al., 2021. It may help to plan those hours in daylight, a lamp by a chair in another room or a corner away from the bed, a book you can easily put down, music turned low, so there's nothing left to decide once you're tired.

For most people it does, though not for everyone, and this is where the evidence is strongest. In Quebec, 160 adults who had lived with insomnia for an average of sixteen years met in groups for six weekly sessions of an hour and a half⁴4 Morin et al., 2009. By the end, about four in ten were free of insomnia, and more again were clearly better⁴.

The benefit fades a little with time, and a year on it's smaller but still clinically meaningful¹⁷17 van der Zweerde et al., 2019. That staying power is what sets it apart from sleeping pills. The two seem to work about as well in the short term, while only CBT‑I has shown lasting effects after treatment stops¹⁸¹⁹18 Perlis et al., 202219 Mitchell et al., 2012. Even so, only a handful of trials have compared them directly, none of them with the newest sleep medicines¹⁹.

It comes in more forms than most people expect, and in the trials the forms with a therapist somewhere in them all worked well, whether one-to-one in person, in a group, by video or phone, or as a self-help book with some therapist contact²⁵25 Simon et al., 2023. Online programs helped somewhat less, and smartphone versions didn't show a clear effect²⁵. One-to-one CBT‑I with a trained provider is the best studied, though such providers can be hard to find and can cost more out of pocket than pills³3 Edinger et al., 2021. A doctor can talk you through in-person, group and online options to fit what's near you and what you can afford³.

If an online course is what's within reach, it can still be a real choice, with gains that were still there a year later, though face-to-face treatment did somewhat better²⁶26 Soh et al., 2020.

If face-to-face help isn't near you or isn't affordable, or the waiting list is long, a self-help book gives you a way to begin while you wait. People working through self-help CBT‑I sleep better, drop out no more often than with a therapist and do better still with brief phone support²⁷27 Ho et al., 2015. A doctor may also be able to offer a brief version over one to four sessions, built around the sleep window and getting up³3 Edinger et al., 2021, when a full course isn't practical.

Wherever you find it, the beginning is the same modest thing: a few lines written over breakfast about the night before. From those lines you and your therapist or program set your first window³3 Edinger et al., 2021. In the studies, the nights that followed tended to get less broken before they got longer⁸¹⁶8 Kyle et al., 201416 Trauer et al., 2015.