Why you forget your dreams (and what actually helps)
Why you forget your dreams (and what actually helps)
I went looking for tips on remembering dreams and found out the brain is actively working against you.

Jacob Lowe · Founder of Sandman
Published on Mar 18th · 5 min read
I used to wake up from a dream and think, if I just lie here for a second, I'll remember it. Then I'd check my phone. Gone. Every time. I figured I just had a bad memory for dreams.
Turns out that's not really what's happening. I went down a rabbit hole on dream recall and the science is wilder than I expected. Your brain isn't passively forgetting your dreams. It's shredding them on purpose.
Your brain has a delete button for dreams
In 2019, researchers at SRI International, Nagoya University, and Hokkaido University published a study in Science that changed how I think about this. There are neurons in your hypothalamus called MCH neurons that fire during REM sleep and actively suppress your hippocampus from forming memories.
About 53% of these MCH cells fired specifically during REM sleep. When the researchers turned them on artificially, memory got worse. When they turned them off, memory improved. Your brain has a dedicated system for making sure you don't remember what you just dreamed.
On top of that, norepinephrine, the neurotransmitter your brain needs to encode memories, drops to almost zero during REM sleep. So the chemical you need for memory is gone, and there's a separate system actively blocking it. Dreams aren't slipping away. They're being wiped.
That "90% in 10 minutes" thing? Nobody knows where it came from
You've probably seen this stat everywhere: you forget 50% of a dream within 5 minutes, 90% within 10. I assumed it came from some classic sleep study. It doesn't. I couldn't find a single controlled experiment behind those specific numbers. It gets attributed to Ebbinghaus sometimes, but his forgetting curve research used nonsense syllables, not dreams. The general idea, that dreams fade fast, is true. But those exact percentages are basically folk wisdom that got repeated until everyone assumed someone proved it.
People who remember dreams aren't better at remembering
Research by Perrine Ruby at INSERM in Lyon looked at this. Her team ran PET scans on 41 people, 21 who remembered dreams often (about 5 per week) and 20 who rarely did (about 2 per month). The high recallers showed more activity in the temporo-parietal junction, a brain region involved in noticing things.
But here's the thing: it wasn't that they were better at holding onto dreams during sleep. They woke up more. Even just briefly, just enough for the brain to grab a fragment and write it down before falling back asleep. Your sleeping brain literally cannot form new memories. It needs those micro-awakenings to encode anything. Dream recall isn't a memory skill. It's a waking pattern.
What actually works
So given all that, what can you do?
The single biggest thing is to stay completely still when you wake up. Don't move, don't open your eyes. Just mentally scan for any dream fragments. Movement and sensory input start overwriting those traces immediately. Checking your phone is the worst thing you can do, not because of the blue light, but because the cognitive engagement floods your working memory with new stuff.
Before bed, tell yourself you're going to remember your dreams. Sounds like nothing, right? But a 2025 study by Giulio Bernardi tracking 217 adults found that attitude toward dreams was one of the three strongest predictors of recall. People who cared about their dreams remembered more of them. Simple as that.
Sleep longer if you can. REM periods get longer as the night goes on. Your first one might be 10 minutes. By hour seven or eight, they can stretch to 45-60 minutes. Cut sleep short and you lose the best dreaming windows.
There's also a technique called wake-back-to-bed (WBTB) where you wake up after 6 hours, stay up for about an hour, then go back to sleep. A sleep lab study by Erlacher and Stumbrys found a 95% dream recall rate using this approach. That number surprised me.
One more: vitamin B6. A placebo-controlled study from the University of Adelaide with 100 participants found that 240mg before bed for five nights improved recall. It didn't make dreams more vivid. It specifically helped people remember them. The theory is B6 increases cortical arousal during REM by helping convert tryptophan into serotonin.
Things that quietly kill your dream recall
Alcohol suppresses REM sleep, especially in higher doses. Less REM, fewer dreams to remember.
Cannabis does the same thing. Regular users often report dreaming less or not at all. One study found THC reduced REM to about 17.7% of total sleep and delayed REM onset to nearly two hours. But here's the interesting part: when people quit, there's a REM rebound effect where dreams come flooding back, vivid and intense. Your brain apparently has a backlog.
Harsh alarms are worse than gentle ones. Getting jolted awake spikes cortisol in a way that interferes with episodic memory. A gradual wake, like a light based alarm, gives you more of that still, scanning for fragments window.
And irregular sleep schedules mess with your sleep architecture. Your body can't settle into consistent REM cycles if bedtime moves around every night.
The weird stuff
A few things I didn't expect to find.
Sleep position apparently matters. A 2012 study of 670 adults found that people who sleep face down have the most vivid dreams. Calvin Kai-Ching Yu at Hong Kong Shue Yan University thinks pressure on the body and restricted breathing get woven into dream content. I sleep on my side, so I can't confirm this, but I kind of want to try it.
Dream recall also drops in winter. The Bernardi study found this seasonal pattern that hadn't been documented before. People remember fewer dreams in winter compared to spring and autumn. No clear explanation yet.
And this one messed with my assumptions: some research suggests that differences in dream recall come from more awakenings during NREM stage 2 sleep, not REM. The popular understanding that dreams equal REM might be too simple.
Where I landed
I've been trying the stay-still-and-scan thing for a couple weeks now. It works. Not every morning, but enough that I've gone from remembering maybe one dream a week to three or four. The biggest change was putting my phone across the room so I can't reach for it on autopilot.
Knowing that my brain is actively working against me actually made it easier to be patient with the process. It's not that I have bad dream memory. It's that nobody has good dream memory — we all have a system designed to erase this stuff. The people who remember are just the ones who've figured out how to grab fragments before they disappear.
If you're starting from zero, just try lying still tomorrow morning. Eyes closed. See what surfaces. You might be surprised.
About the Author

Jacob Lowe
Founder of Sandman
Jacob is a web developer with over a decade of experience in the field. His passion for coding and open-source technologies drives his desire to create and innovate. He believes that through technology, we have the power to increase access to new experiences and make a positive impact in the world. At the heart of his work lies a love for nature and the beauty of the natural world. He finds solace in the stillness of nature and the abstractions of code.
Remember your dreams
A dream journal that helps you track, understand, and learn from your dreams.
Remember your dreams