Can writing down nightmares actually help?

I've had the same nightmare maybe a dozen times over the past year. The details shift but the shape is always the same: I'm somewhere familiar, something goes wrong, and I can't fix it. I wake up with my heart going and that specific kind of dread that takes a few minutes to shake off.

I always figured nightmares were just something you dealt with. Write it off, move on, hope it doesn't come back tonight. But when I started looking into it, I found there's a clinical technique that actually works, it's surprisingly simple, and writing down the nightmare is the first step.

Most people don't think nightmares are treatable

This is maybe the most frustrating thing I found. Only about a third of adults with clinically significant nightmares believe the condition is treatable. Among veterans, who experience trauma-related nightmares at much higher rates, only 11-38% even bring it up with a healthcare provider.

Meanwhile, about 4-5% of adults meet the criteria for nightmare disorder, which the DSM-5 defines as recurrent disturbing dreams that cause real distress or mess with your ability to function. That's a lot of people quietly suffering through something that has effective treatment.

The technique that actually works

It's called Image Rehearsal Therapy, or IRT. A psychiatrist named Barry Krakow developed it in the mid-1990s, and the American Academy of Sleep Medicine considers it the only "recommended" treatment for nightmare disorder. Everything else gets a "may be used" classification.

Three steps:

1. Write down a recurring nightmare in detail while you're awake
2. Change the storyline. Rewrite it with a different outcome. It doesn't have to be the opposite of the nightmare, just different.
3. Rehearse the new version in your mind for 10-20 minutes a day

That's it. You can learn it in a single session and practice it on your own.

The numbers are hard to argue with

I found a 2001 study by Krakow in JAMA that looked at 168 women who were sexual assault survivors, 95% with moderate-to-severe PTSD. Their nightmare frequency dropped from 3.88 nights per week to 1.33. The effect size was large (Cohen's d of 1.24 for nightmare nights).

A 2012 meta-analysis pooling 13 studies and 511 participants found nightmare frequency dropped with an effect size of 0.69, sleep quality improved at 0.68, and PTSD symptoms dropped at 0.72. Those effects held up at 6-12 month follow-ups. In an 18-month follow-up, 68% of subjects had dropped below the threshold for nightmare disorder.

People started seeing partial results after 2-3 weeks. That's fast for something that might have been haunting you for years.

About 30% of patients don't respond to IRT, which is worth knowing. It's not a magic fix for everyone. But for the majority, it works and it lasts.

Why it works is interesting

This is the part that got me. A 2019 study from the University of Geneva found that people who experienced moderate fear in their dreams actually showed reduced emotional arousal when facing threats while awake. Their amygdala (the brain's fear center) was less reactive, and their medial prefrontal cortex (which inhibits fear) was more active.

Bad dreams are basically overnight training for emotional regulation. Your brain rehearses scary situations in a safe environment so you handle them better when you're awake.

But there's a threshold. The researcher Lampros Perogamvros put it this way: "if a certain threshold of fear is exceeded in a dream, it loses its beneficial role as an emotional regulator." That's the line between a bad dream and a nightmare. Bad dreams help you. Nightmares overwhelm the system.

So IRT is doing by hand what healthy dreaming does on its own. You take the nightmare that overwhelmed the system, give it a different ending, and rehearse it until your brain picks up the new version.

The piano chord study

OK so this is where it gets weird. The same Geneva team published a study in 2022 combining IRT with something called targeted memory reactivation.

They had 36 nightmare disorder patients do IRT, but half the group also heard a specific piano chord (C69) played while they rehearsed the positive version of the dream. Then for two weeks, a headband device played that same chord during REM sleep.

The IRT-only group went from 3 nightmares per week to 1. The group with the sound cue went from 3 to 0.2. They also reported having happier dreams overall. The effects held at three months.

A piano chord. Played during sleep. I realize this sounds like something from a sci-fi movie, but the paper is real and the results were significant.

Journaling alone isn't enough

Writing down nightmares is step one of IRT, and it's important. The journal gives you the raw material to work with. It helps you see patterns and creates enough distance to engage with the content instead of just being overwhelmed by it.

But journaling without the rescripting step has a real downside. You're re-exposing yourself to the traumatic content without the therapeutic resolution. James Pennebaker's expressive writing research showed that writing about traumatic experiences can help, but only when there's emotional engagement and processing, not just recounting facts.

If you're going to write down a nightmare, follow it with the rewrite. Always. Don't just record it and close the notebook.

Things that make nightmares worse

A few things I didn't expect. Beta-blockers cause nightmares in roughly a third of users, which is a lot considering how commonly they're prescribed. SSRIs can intensify dream content too.

Alcohol is a sneaky one. It suppresses REM sleep early in the night, but when it wears off around 3 or 4 a.m., you get a REM rebound that produces more vivid, often disturbing dreams. So the nightcap might actually be making things worse.

And then there's melatonin, which is genuinely confusing. It can cause vivid dreams and nightmares in some people, while also helping counter nightmares caused by beta-blockers. I don't fully understand that one yet.

Where I am with this

I tried the IRT technique on my recurring nightmare. Wrote it down, changed the ending (the thing that breaks, I fix it, then I leave), and spent a few minutes each night before bed running through the new version.

It's been about three weeks. The nightmare came back once, but the new ending sort of bled into it. I woke up less rattled. I don't know if that counts as the technique working or just knowing what the nightmare is takes some of its power away. Maybe both.

The thing that stuck with me most from the research is that nightmares aren't just unpleasant. They're a sign that something in your brain's emotional processing is stuck. And unlike a lot of things that go wrong with sleep, this one has a fix that doesn't require medication, a sleep lab, or years of therapy. Just a pen, a rewrite, and some rehearsal time before bed.