PTSD, Nightmares and Sleep: CBT-I, Imagery Rehearsal and What the Evidence Actually Shows

Bottom line: PTSD-related sleep problems deserve their own assessment. Insomnia, nightmares, fear of sleep, hypervigilance and sleep-disordered breathing can overlap, but they are not interchangeable. CBT-I has direct evidence for PTSD-related insomnia, and imagery rehearsal therapy has meaningful evidence for recurrent trauma-related nightmares. At the same time, improving sleep does not automatically treat every PTSD symptom, and treating PTSD does not guarantee that insomnia or nightmares disappear.[1-14]

PTSD sleep disturbance is more than “bad dreams”

Sleep problems are common in PTSD, but collapsing them into one symptom loses clinically useful information.

A person with PTSD may have difficulty falling asleep from hyperarousal, recurrent trauma-related nightmares, repeated awakenings, fear of sleep, conditioned alertness in bed, irregular timing, medication or substance effects, sleep apnea, or several of these at once.[1,3,12,14]

Those pathways do not all respond to the same intervention.

The useful question is therefore not simply “What makes someone sleepy?” It is “What is actually disrupting sleep?”

CBT-I directly treats insomnia in PTSD

A randomized controlled trial of 45 adults with PTSD and insomnia compared eight sessions of CBT-I with a monitoring-only waitlist.[1]

CBT-I was superior across sleep-diary outcomes and improved polysomnography-measured total sleep time.[1]

The remission result is especially clear:

• 41% of CBT-I participants reached full subjective sleep remission;
• 0% of waitlist participants did.[1]

The sleep gains were maintained at six-month follow-up.[1]

That is direct evidence that insomnia in PTSD can be a legitimate treatment target rather than something that must simply wait for all trauma symptoms to resolve.

Better sleep does not automatically mean a larger PTSD response

A 2025 randomized trial in 94 veterans compared CBT-I integrated with prolonged exposure therapy against sleep hygiene integrated with prolonged exposure therapy.[4]

Both groups improved substantially in PTSD symptoms, but the groups did not differ significantly in PTSD symptom reduction.[4]

The CBT-I group did show significantly larger improvements in insomnia severity, sleep efficiency, total sleep time and perceived quality of life.[4]

That is an excellent endpoint distinction:

CBT-I made the sleep treatment better without proving that it made trauma-focused psychotherapy more effective for PTSD itself.

Sleep hygiene is not equivalent to CBT-I

Sleep hygiene means advice about routines, caffeine, light, exercise or the bedroom.

CBT-I is a structured behavioral treatment that can include stimulus control, sleep restriction or compression, cognitive work around insomnia, restructuring conditioned wakefulness and relapse prevention.

For chronic insomnia, a bedtime-habits handout should not be described as “basically CBT-I.”[1,4]

PTSD treatment can improve while insomnia remains

A 2023 systematic review examined whether sleep disturbance affects outcomes from trauma-focused psychotherapy.[12]

Across 16 primary studies, sleep problems were associated with greater overall PTSD severity, but the review did not find strong evidence that insomnia necessarily prevents trauma-focused therapy from working.[12]

The notable exception was sleep-disordered breathing, which may complicate outcomes.[12]

That supports a flexible model: insomnia does not always need to be completely resolved before trauma therapy starts, and trauma-focused treatment does not guarantee that sleep will normalize.[3,4,12]

Nightmares are a separate treatment target

Trauma-related nightmares overlap with insomnia but are not identical to it.

Someone can have recurrent nightmares with otherwise adequate sleep opportunity, severe insomnia with few nightmares, or both conditions simultaneously.

That distinction changes the evidence base.

Imagery rehearsal therapy has direct nightmare evidence

Imagery rehearsal therapy (IRT) asks the patient to modify a recurrent nightmare while awake and repeatedly rehearse the revised imagery.

One landmark randomized trial included 168 women with chronic nightmares, most of whom had moderate-to-severe PTSD after sexual assault.[5]

Compared with waitlist, IRT produced large improvements in nights per week with nightmares (d = 1.24), number of nightmares per week (d = 0.85) and sleep quality (d = 0.67), along with PTSD symptom improvement.[5]

The gains were sustained from three to six months without additional intervention.[5]

Meta-analytic evidence supports imagery-based nightmare treatment

A 2012 meta-analysis of 13 studies found large pre-to-post effects of imagery rehearsal on nightmare frequency, sleep quality and post-traumatic stress symptoms.[6]

Adding CBT-I components to imagery rehearsal produced greater improvement in sleep quality than imagery rehearsal alone, but combined treatment did not clearly add benefit for nightmares or PTSD outcomes.[6]

A 2024 review focused on military veterans also found supportive imagery-rescripting signals while emphasizing variability and limitations in the veteran literature.[9]

A 2025 review concluded that IRT and prazosin remain among the most supported nightmare-focused treatments, while results still vary across studies and newer CBT-based nightmare approaches continue to develop.[11]

Adding imagery rehearsal to CBT-I does not always improve outcomes

A randomized trial of 108 U.S. veterans with severe PTSD and recurrent deployment-related nightmares compared CBT-I plus imagery rehearsal with CBT-I alone.[2]

Both groups improved, but imagery rehearsal did not significantly outperform CBT-I alone on the primary nightmare-frequency or nightmare-distress outcomes.[2]

Across treatment, 29% achieved a meaningful reduction in nightmare frequency and 22% reached nightmare remission.[2]

The lesson is not that IRT “does not work.” It is that a treatment can beat waitlist or usual care and still fail to add benefit when layered onto another active treatment.

Insomnia treatment itself can reduce some trauma-related nightmares

A randomized study of 91 veterans found brief behavioral treatment for insomnia reduced trauma-related nightmare frequency relative to progressive muscle relaxation immediately after treatment.[13]

However, the nightmare reduction was not maintained at six months, and non-trauma-related bad dreams did not improve.[13]

That suggests some nightmare burden can shift when sleep becomes more stable without proving that insomnia treatment replaces nightmare-focused therapy.

Combined insomnia + nightmare treatment can add value

A 2023 pilot randomized trial in 93 active-duty military personnel compared cognitive behavioral treatment for insomnia and nightmares plus cognitive processing therapy with cognitive processing therapy alone.[3]

The trial ended early and was underpowered, so its findings should be treated as promising rather than definitive.

Compared with CPT alone, adding sleep-focused treatment produced larger improvements in PTSD symptoms (d = -0.36), insomnia (d = -0.77), sleep efficiency (d = 0.62) and nightmares (d = -0.53).[3]

Prazosin has evidence—but not a universal verdict

A 2022 network meta-analysis included 29 randomized trials involving 2,214 trauma survivors and identified both prazosin and imagery rehearsal therapy as interventions with evidence for trauma-related nightmares.[7]

A larger 2024 pharmacotherapy network meta-analysis included 99 RCTs with 10,481 participants.[8]

Prazosin ranked favorably for insomnia, nightmares and poor sleep quality, but the authors emphasized underpowered comparisons and confidence ranging from very low to moderate.[8]

So the calibrated conclusion is:

• prazosin has meaningful trial and meta-analytic support;
• results vary across studies and populations;
• medication selection requires individualized clinical review;
• this evidence does not justify a universal site-wide dose or self-treatment recipe.

Sedation is not the same as nightmare treatment

A drug can make someone sleepy without treating trauma-related nightmares, conditioned hyperarousal or PTSD.

The 2024 pharmacotherapy network meta-analysis found weak or uncertain sleep-specific evidence for several commonly used sedating or psychiatric medications.[8]

The sleep cluster therefore keeps four endpoints separate:

feeling sedated ≠ resolving insomnia ≠ reducing nightmares ≠ treating PTSD.

Trauma-focused therapy, CBT-I and IRT answer different questions

Trauma-focused therapies such as prolonged exposure or cognitive processing therapy target the PTSD syndrome.

CBT-I targets insomnia.

IRT targets recurrent nightmares.

Those interventions can overlap in downstream benefits, but they should not be treated as interchangeable.[2-4,12]

A person may need trauma-focused therapy, insomnia treatment, nightmare-focused treatment, sleep-apnea evaluation, medication review, or some combination of these.

Sleep apnea deserves special attention in PTSD

Sleep-disordered breathing can coexist with PTSD and insomnia.

Airway obstruction can fragment sleep, insomnia can coexist with OSA, and sedation does not correct airway collapse.[12]

The 2023 psychotherapy-outcome review found sleep-disordered breathing was the sleep problem most consistently associated with poorer trauma-focused treatment outcomes.[12]

If snoring, witnessed apneas, gasping or marked daytime sleepiness are present, objective sleep-apnea evaluation may be more relevant than another calming supplement.

See Sleep Apnea vs Insomnia and Snoring vs Sleep Apnea.

Fear of sleep can maintain insomnia

PTSD-related insomnia is not always explained by nightmares alone.

Some people develop fear of losing vigilance, fear of another nightmare, or conditioned fear around the bed or nighttime.

In a secondary analysis of the PTSD CBT-I trial, greater fear of sleep was associated with more severe PTSD symptoms, more frequent nightmares and stronger hypervigilance, and fear of sleep decreased with CBT-I.[14]

That is a plausible mechanism by which insomnia can remain even when the original danger is long past.

What about supplements?

There is no strong basis for claiming that a generic calming supplement specifically treats PTSD-related nightmares.

A supplement trial in generalized stress or ordinary insomnia should not be silently inherited into PTSD.

That means magnesium evidence is not PTSD-nightmare evidence, melatonin-timing evidence is not IRT evidence, herbal sedation is not trauma-focused treatment, and a general anxiety study does not establish efficacy for PTSD-related insomnia.

Practical decision map

Mostly insomnia

Think about CBT-I, adequate sleep opportunity, conditioned wakefulness, schedule regularity, substance and medication effects, and coexisting sleep apnea.

Recurrent trauma-related nightmares

Think about nightmare-focused assessment, imagery rehearsal or related rescripting approaches, whether insomnia treatment is also needed, and clinician-guided medication options when appropriate.

PTSD symptoms remain severe but sleep has improved

That does not mean the sleep treatment failed. Sleep and PTSD are overlapping but distinct treatment targets.

PTSD improved but insomnia or nightmares remain

That does not mean trauma treatment failed either. Residual sleep disorders may still deserve their own treatment.

What the evidence does not prove

Current research does not justify saying that:

• every person with PTSD needs IRT;
• every nightmare is caused by PTSD;
• CBT-I alone treats the full PTSD syndrome;
• adding IRT to CBT-I is always superior;
• prazosin works for everyone;
• a sedating medication is automatically treating nightmares;
• trauma-focused psychotherapy always resolves insomnia;
• treating insomnia must happen before trauma-focused treatment can begin;
• supplements validated in ordinary insomnia are proven for PTSD-related sleep disturbance.

Related sleep research

• Anxiety and Sleep
• Depression and Sleep
• CBT-I vs Sleep Supplements
• Sleep Apnea vs Insomnia
• Snoring vs Sleep Apnea
• Insomnia vs Sleep Deprivation
• Subjective vs Objective Sleep

Verdict

PTSD-related insomnia and nightmares are treatable sleep problems, but they are not one problem.

CBT-I has direct randomized evidence for insomnia in PTSD. Imagery rehearsal has meaningful evidence for recurrent trauma-related nightmares. Prazosin has supportive but variable pharmacologic evidence. Trauma-focused treatment remains central for PTSD itself.

The most accurate model is therefore a parallel, diagnosis-specific approach in which insomnia, nightmares, PTSD symptoms and sleep-disordered breathing are assessed according to the evidence for each target.