Image Rehearsal Therapy for Nightmares:
The Evidence-Based Treatment Veterans Are Rarely Offered
IRT is the AASM Level A recommended treatment for PTSD-associated nightmares, what the Krakow 2001 trial, the 2019 VA RCT, and current meta-analyses show.
Key Takeaways
- IRT is the AASM Level A (highest evidence) recommended treatment for PTSD-associated nightmares, based on the 2001 Krakow JAMA RCT and subsequent meta-analyses.
- Krakow 2001 (n=168): IRT reduced nightmare frequency (d=1.24), sleep quality (d=0.67), and PTSD symptoms (d=1.53), among the largest effect sizes in any PTSD psychotherapy.
- IRT does not require re-exposure to the traumatic event. The veteran rewrites the nightmare any way they choose and rehearses the new version, no trauma processing involved.
- The 2019 VA RCT (Harb et al., n=108) found IRT+CBT-I was not superior overall but specifically benefited women veterans and lower-severity presentations.
What IRT Is and How It Works
The mechanism in plain language
IRT is built on a simple cognitive principle: nightmares are learned behaviors. Like any learned behavior, they can be modified through rehearsal. The brain that replays a trauma sequence night after night is, through repetition, strengthening the neural pathway that produces that sequence. IRT intervenes at this pathway by introducing a competing rehearsed version.
The procedure has three steps. First, the veteran selects a recurrent nightmare and writes it down in as much detail as is comfortable. Second, the veteran rewrites the nightmare any way they choose, the ending, the setting, the characters, even the entire arc. The new version does not need to be realistic, resolved, or thematically related to the original. It needs only to be different and non-threatening. Third, the veteran reads the rewritten nightmare aloud or to themselves for 10–20 minutes daily, for at least 3–4 weeks.
The mechanism is not fully understood. Leading theories include: (1) the cognitive restructuring of the nightmare as a “habit” rather than an inescapable memory, (2) the introduction of a competing neural pathway during REM sleep, and (3) the reduction of cognitive-emotional pre-sleep arousal. What the research is consistent on is the outcome: nightmare frequency, nightmare distress, and sleep quality all improve.
Why no trauma exposure is required
The contrast with exposure therapy
Standard PTSD exposure therapies (PE, EMDR) work by having the patient repeatedly confront the traumatic memory until the fear response habituates. IRT does not work through habituation, the veteran never describes or relives the original traumatic event.
The rescripting mechanism
IRT works through rescripting, creating a new narrative that the dreaming brain can access alongside the trauma narrative. The rewrite competes with the original during consolidation, not by confronting it directly but by offering an alternative.
Why it matters for engagement
Veterans who have been reluctant to engage with exposure-based PTSD treatments often find IRT more approachable precisely because the original event is not the subject of treatment. The lower engagement threshold is a clinical feature, not a compromise.
Who this applies to most
- Veterans with PTSD and nightmares ≥2 nights/week: This is the primary indication. IRT can be delivered alongside or following CBT-I for insomnia.
- Veterans who completed PTSD treatment but retain nightmares: Nightmares are one of the most treatment-resistant PTSD symptoms. IRT should be initiated when nightmares persist after PTSD remission.
- Women veterans: The 2019 VA RCT specifically found that adding IRT to CBT-I produced better outcomes for women veterans. IRT may be particularly indicated in MST-related nightmare presentations.
- Veterans who cannot engage with exposure-based therapy: IRT’s non-exposure design makes it accessible to veterans who have found Prolonged Exposure or CPT too distressing to complete.
The Evidence Base
The Krakow 2001 JAMA landmark trial
Barry Krakow and colleagues conducted the first published RCT[1] of IRT in 2001, enrolling 168 women with PTSD secondary to sexual assault. Participants were randomized to IRT (3 sessions, n=88) or waitlist control (n=80). At 3 and 6 months:
- Nightmare frequency: Cohen d=1.24 (IRT vs. control; p<0.001)
- Number of nightmares per week: d=0.85 (p<0.001)
- Sleep quality (PSQI): d=0.67 (p<0.001)
- PTSD symptoms (PSS): d=1.00 (p<0.001)
- PTSD symptoms (CAPS): d=1.53 (p<0.001)
These are among the largest effect sizes documented in any psychotherapy for PTSD-related outcomes. The JAMA publication established IRT as a credible, RCT-supported first-line intervention.
The 2019 VA veteran-specific RCT
Harb et al. conducted the most methodologically rigorous[2] IRT trial specific to military veterans: 108 US veterans of Iraq and Afghanistan with current severe PTSD and recurrent deployment-related nightmares, randomized to IRT+CBT-I (n=55) or CBT-I alone (n=53). The headline finding was complex, overall, IRT+CBT-I was not superior to CBT-I alone for the full sample (29% nightmare reduction in both groups). However, subgroup analyses found that IRT addition specifically benefited: (a) women veterans and (b) veterans with lower baseline nightmare severity.
Meta-analytic evidence
The Casement and Swanson (2012)[3] meta-analysis of imagery rehearsal for post-trauma nightmares found small to moderate effects on nightmare frequency and sleep quality across 13 studies, with consistent effects despite heterogeneous samples. The 2024 Hicks et al. systematic review[4] specifically focused on military veterans confirmed significant reductions across nightmare, sleep, and PTSD outcomes, with the caveat that dropout rates are higher in veteran samples than civilians.
What the critics say
The 2019 Harb et al. finding, that IRT added to CBT-I was not superior to CBT-I alone in the full sample, has led some clinicians to question whether IRT provides independent benefit beyond CBT-I. The critique deserves careful framing. First, CBT-I alone produced significant nightmare improvement, suggesting that addressing conditioned arousal and sleep consolidation may be sufficient for many veterans. Second, the IRT benefit in women and lower-severity cases is real and should not be discarded. Third, a single VA RCT should not override consistent meta-analytic evidence from broader populations. IRT remains the AASM Level A recommendation.
What the Evidence Doesn’t Say
Optimal IRT protocol for veterans. The number of sessions, the degree of trauma exposure in rescripting, and format (individual vs. group) have not been definitively optimized for veteran populations. Standard protocols use 3–6 sessions; some research suggests single-session formats may be comparably effective.
Whether nightmares predict suicide risk, and whether IRT addresses that pathway. Nightmares are independently associated with elevated suicidal ideation in veterans. Whether IRT specifically reduces suicide risk through nightmare reduction is not yet established.
Long-term durability beyond 6 months. Most IRT trials follow participants to 3–6 months. Whether gains persist at 12–24 months in veteran populations is not well characterized.
Men vs. women differences in mechanism. The Harb et al. subgroup finding that women benefit more from IRT addition to CBT-I than men has not been explained mechanistically.
Clinical Implications
| Application | Evidence | Strength | Notes |
|---|---|---|---|
| Ask specifically about nightmare frequency at every VA mental health intake | Only 11–38% of veterans with clinically significant nightmares discuss them with providers, systematic inquiry is required | Strong (epidemiological data) | Add “How many nights per week do you have nightmares?” to standard intake |
| Initiate IRT when nightmares persist after PTSD treatment | Nightmares require separate treatment, they do not reliably remit after PE, CPT, or EMDR alone | Strong (AASM Level A) | IRT can be delivered during or after PTSD treatment; does not require PTSD remission first |
| Add IRT to CBT-I for women veterans and lower-severity nightmares | Subgroup benefit documented in the largest VA veteran RCT | Moderate (Harb et al. 2019 subgroup) | For high-severity nightmares with replicative content, consider IRT regardless of sex |
| Offer IRT as an alternative when veterans cannot engage with exposure therapy | IRT’s non-exposure design is the key differentiator for veterans who have found PE or CPT too distressing | Moderate (clinical practice; engagement literature) | Frame IRT as nightmare-specific, not trauma therapy |
| Provide telehealth IRT to overcome access barriers | Telehealth delivery has comparable outcomes to in-person; significant access barriers in rural veteran populations | Moderate (telehealth literature) | VA Video Connect is the delivery platform; request specifically |
What Can You Do?
| How to Implement | Expected Benefit (and Why) | Evidence Strength | Context Notes |
|---|---|---|---|
| Report nightmare frequency explicitly at your VA appointment | |||
| Tell your provider: “I have nightmares [X] nights per week. I want to discuss treatment options specifically for nightmares.” | Opens access to IRT referral, because nightmare disorder is systematically underreported and often attributed to general PTSD without specific treatment initiation | Strong | If your provider doesn’t mention IRT, ask by name: “Is Imagery Rehearsal Therapy available here?” |
| Request an IRT referral, not just CBT-I | |||
| Ask: “I would like a referral specifically for Imagery Rehearsal Therapy for nightmare disorder.” | Ensures nightmare-specific treatment rather than general insomnia treatment, because CBT-I addresses conditioned arousal and sleep avoidance but not trauma nightmare content | Strong (AASM Level A) | If IRT is unavailable at your facility, ask about telehealth delivery to another VA site |
| Complete IRT with a trained provider, do not attempt solo | |||
| Find a VA mental health or sleep medicine provider trained in IRT; ask during initial contact whether they have IRT training | Supervised IRT produces the largest effect sizes; self-directed versions are less studied in veteran populations | Moderate–strong | IRT is brief (3–6 sessions); ask about the provider’s specific experience with veteran nightmare presentations |
| Practice the rescripted dream daily for at least 3 weeks | |||
| After your first IRT session, commit to 10–20 minutes of daily rehearsal of the new dream version | Builds the competing neural pathway through repetition, because the brain strengthens rehearsed sequences through the same consolidation processes that maintain the trauma replay | Strong (IRT protocol requirement) | Miss a day and resume, consistent practice over weeks, not perfect daily compliance |
| Add CBT-I if sleep initiation and maintenance are also problems | |||
| Ask your IRT provider about combining IRT with CBT-I if you have both nightmares and difficulty falling or staying asleep | Combined CBT-I + IRT produces better sleep quality and PTSD symptom outcomes than either alone, because CBT-I addresses conditioned arousal while IRT addresses nightmare content, two separate mechanisms | Moderate (meta-analytic support) | Women veterans: the combination shows particularly strong evidence |
How to Use AI With This Information
When to Work With a Professional
IRT requires a trained provider. Self-directed attempts may produce limited benefit and, in rare cases, may increase nightmare distress if the rewriting process inadvertently intensifies trauma engagement. Seek evaluation from a VA mental health or sleep medicine provider if:
- You have nightmares 2 or more nights per week that cause significant distress or avoidance of sleep
- Your nightmares have persisted after completing PTSD treatment
- You experience behavioral sleep enactment (yelling, thrashing, leaving the bed), this requires polysomnographic evaluation before nightmare therapy
- You have been offered CBT-I but your nightmares remain the primary sleep complaint
FAQ’s
Do I have to describe the traumatic event in detail during IRT?
No. IRT asks you to describe the nightmare content, not the original traumatic event. Many IRT protocols do not require any description of the trauma itself. You describe the nightmare as it recurs and then create a new version, the original event is not the subject of treatment.
How is IRT different from the nightmare work done in Prolonged Exposure or CPT?
PE and CPT address the traumatic memory through controlled exposure and cognitive restructuring respectively. IRT does not address the trauma memory directly, it targets the nightmare as a learned behavioral pattern, introducing a competing rehearsed narrative. The two approaches can be complementary and are often combined.
What if my nightmare rewrite feels absurd or doesn’t make sense?
That is expected and acceptable. IRT does not require the new version to be realistic, resolved, or satisfying in a narrative sense. Many veterans find that a clearly fantastical or even humorous rewrite works as well as a realistic one. The cognitive principle is that any different, non-threatening sequence creates a competing pathway.
How quickly does IRT work?
Most clinical protocols expect measurable improvement within 3–6 weeks of consistent daily rehearsal. Some veterans notice changes sooner; some require longer. The research shows significant effects at 3 months maintained at 6 months.
REFERENCES
- Krakow B et al. (2001). Imagery rehearsal therapy for chronic nightmares in sexual assault survivors with PTSD: a randomized controlled trial. JAMA, 286(5), 537–545. doi:10.1001/jama.286.5.537
- Harb GC et al. (2019). Randomized controlled trial of imagery rehearsal for posttraumatic nightmares in combat veterans. J Clin Sleep Med, 15(5), 757–767. doi:10.5664/jcsm.7704
- Casement MD & Swanson LM. (2012). A meta-analysis of imagery rehearsal for post-trauma nightmares. Clin Psychol Rev, 32(6), 566–574. doi:10.1016/j.cpr.2012.06.002
- Hicks AJ et al. (2024). Imagery rescripting interventions for military veterans with nightmares: Systematic review and meta-analysis. Clin Psychol Psychother, 31(4), e3025. doi:10.1002/cpp.3025
- Albanese M et al. (2022). Nightmare rescripting: using imagery techniques to treat sleep disturbances in PTSD. Front Psychiatry, 13, 866144. doi:10.3389/fpsyt.2022.866144
- El-Solh AA. (2018). Management of nightmares in patients with PTSD. Nature and Science of Sleep, 10, 409–420. doi:10.2147/nss.s166089
- Yucel DE et al. (2020)[7]. Comparative efficacy of imagery rehearsal therapy and prazosin for trauma-related nightmares. Sleep Medicine, 71, 1–9. doi:10.1016/j.sleep.2019.12.008
- Morgenthaler TI et al. (2018)[8]. Position paper for treatment of nightmare disorder in adults: AASM position paper. J Clin Sleep Med, 14(6), 1041–1055. doi:10.5664/jcsm.7178
- Margolies SO et al. (2019). RCT of imagery rehearsal for recurrent nightmares in 108 veterans. J Clin Sleep Med, 15(5). doi:10.5664/jcsm.7704
- Seda G et al. (2015). Comparative meta-analysis of prazosin and imagery rehearsal for nightmare frequency and sleep quality. J Clin Sleep Med, 11(1). doi:10.5664/jcsm.4354