What is Imagery Rehearsal Therapy?

Imagery Rehearsal Therapy is a brief cognitive-behavioral protocol developed by Barry Krakow for recurring nightmares. You write out a recurring nightmare, rewrite it however you choose with one rule that the new version does not reproduce the traumatic threat, then rehearse the rewritten version in waking imagination for about 10 to 20 minutes a day. Nightmare frequency usually declines over two to four weeks.

How many sessions does IRT take?

IRT is short. Most published protocols run four to six sessions, and some deliver the core instruction in as few as one to three. The daily home practice between sessions does most of the work, so the total clinician time is modest compared with longer trauma therapies.

Does IRT work for non-trauma nightmares?

It helps both, and the American Academy of Sleep Medicine recommends it for idiopathic nightmares and PTSD-associated nightmares alike. The largest effects show up with trauma-replay nightmares, where the frightening content is fairly literal. When the nightmare is driven by a developmental or relational pattern rather than a discrete threat, rewriting the surface story reaches less of what is generating it.

Is IRT recommended by clinical guidelines?

Yes. The 2018 American Academy of Sleep Medicine position paper lists IRT as the psychotherapy recommended for nightmare disorder and PTSD-associated nightmares, its top designation, ahead of treatments the paper says only may be used. That endorsement rests on multiple randomized trials.

Can I do IRT on my own?

The rewrite-and-rehearse structure is simple enough to try alone, and self-guided versions have some support. Clinician guidance tends to produce stronger and more reliable results, especially with trauma-related nightmares, because dysregulation can spike during the rewrite and a therapist helps keep the new version from accidentally re-staging the threat.

Imagery Rehearsal Therapy: The Protocol That Actually Treats Recurring Nightmares

TL;DR: Imagery Rehearsal Therapy is the psychotherapy the American Academy of Sleep Medicine recommends for recurring nightmares, including nightmares in PTSD. You rewrite a recurring nightmare so it no longer reproduces the threat, then rehearse the new version in waking imagination about 10 to 20 minutes a day. Frequency usually drops over two to four weeks.

A patient once told me that cognitive behavioral therapy for insomnia had fixed everything except the reason she came in. She fell asleep faster. She stopped clock-watching at 3 a.m. And every few nights the same nightmare still arrived on schedule, the same locked room, the same footsteps, and she woke up with her heart going and the rest of the night ruined. The sleep-onset problem had a treatment. The dream did not, or so she thought.

The dream did have a treatment. It has had one since the 1990s, and most people who live with recurring nightmares have never heard its name.

What Imagery Rehearsal Therapy is

Imagery Rehearsal Therapy, or IRT, is a brief cognitive-behavioral protocol for recurring nightmares. The clinical psychologist Barry Krakow and colleagues developed and tested it, much of the early work with survivors of sexual assault whose nightmares had persisted for years. The core idea is direct. A recurring nightmare behaves like an overlearned script. IRT teaches the mind a different script by having you rewrite the nightmare while awake and practice the new version until it competes with the old one.

You do not interpret the dream. You do not decode its symbols. You change it and rehearse the change.

The four-step protocol

The protocol is simple to describe and takes practice to do well.

First, you select one recurring nightmare to work on. Early in treatment, clinicians often steer you toward a nightmare of moderate intensity rather than the most severe one, so the method has room to work before you apply it to the worst material.

Second, you write the nightmare out. Putting the sequence on paper turns a flood of images into something with edges you can see and, later, alter.

Third, you rewrite it. You change the dream in any way you want. You can change the setting, the other figures, the turning point, or the ending. The one firm rule sits here, and I will return to it.

Fourth, you rehearse the rewritten version in waking imagination for roughly 10 to 20 minutes a day. This is the active ingredient. You run the new dream like a scene, in detail, on purpose, while awake. Over two to four weeks of daily rehearsal, the original nightmare typically drops in frequency, in intensity, or both.

Most courses of IRT run about four to six sessions, with the daily home practice carrying the load between visits.

The one rule about the rewrite

The rewrite has a constraint that matters more than any other instruction: the new version must not reproduce the traumatic threat. You are not rehearsing a braver way to survive the attack. You are not scripting a fight where you win. Re-staging the danger, even a version you now escape, keeps rehearsing the threat itself, which is the thing the nervous system is stuck on.

So the new dream moves somewhere else entirely. The locked room becomes a room with a door that was always open. The footsteps belong to someone bringing coffee. The point is to give the imagination a different track to run on, not a rematch. Getting this right is the main reason clinician guidance helps, because under stress people tend to write rescue scenes that quietly reinstall the fear.

What the evidence shows

IRT has the strongest guideline support of any nightmare psychotherapy. The 2018 American Academy of Sleep Medicine position paper on nightmare disorder in adults lists IRT as the treatment that is “recommended,” its top designation, for both nightmare disorder and PTSD-associated nightmares. Every other listed option, including several medications, received the weaker language that it “may be used.” That recommendation rests on multiple randomized controlled trials and on meta-analyses showing reductions in nightmare frequency and in overall sleep disturbance.

Two honest caveats. The evidence is strongest for reducing nightmare frequency and distress, and somewhat less consistent on downstream measures like broad PTSD symptom scores. And trial quality varies, which is why guideline bodies describe the support as good rather than settled. IRT works for a lot of people. It is not a guarantee for everyone.

Where IRT fits, and where it does not reach

The effect is largest where the nightmare replays a discrete trauma more or less literally. There the content is close to the surface, and changing it reaches most of what is driving the dream. You can read more about why certain nightmares lock into place in why the same nightmare keeps coming back, and about the trauma-specific pattern in dreams after trauma.

The reach is shorter when the nightmare is driven by a developmental or relational structure rather than a single threat. If the dream is generated by an early attachment wound, a pattern of betrayal, or a way of relating to yourself that formed long before any nameable event, rewriting the surface story treats the symptom and leaves the source. The nightmare may fade and then return in a new costume, because the thing producing it was never the plot. For a fuller version of this problem, see the same nightmare every night.

This is not a knock on IRT. It is a map of what a content-level intervention can and cannot do.

Combining IRT with deeper work

In practice, the sequence often runs like this. IRT first, because it is brief and it can give a person real relief from the nightmare within a few weeks, which frees up the capacity to do harder work. Then, if a developmental or relational driver is present, depth-oriented psychotherapy or EMDR to reach the structure underneath. The two are not rivals. IRT quiets the dream; the deeper work addresses what the dream was expressing.

My patient with the locked room did IRT. Within about three weeks the dream stopped arriving on schedule. What surfaced next, once the nights were quiet enough to notice it, was a much older question about safety that the nightmare had been shouting over. That is usually the shape of it. The protocol clears the noise. Then you can hear what the noise was covering.