Ask ten people how vividly they remember last night’s dream and most of them will pause, look slightly embarrassed, and tell you they don’t really dream anymore.
They do. Everybody does. You spend something close to two hours a night in REM sleep, and the dreaming happens whether or not any of it comes back with you.
I find this a strangely moving fact. There is a whole second life running every night, and for most adults it is almost entirely unwitnessed.
The usual explanation you’ll read is that your alarm is stealing your dreams. That’s the version I believed for years, and there is something real in it. But when I went looking for the actual research, the picture turned out to be more interesting than the popular story, and in one important way it contradicts it.
The Part That Is True: Your Dreams Are at the Back of the Night
Sleep is not one flat state. It moves in cycles of roughly 90 to 120 minutes, and the shape of those cycles changes as the night goes on.
Deep slow-wave sleep is front-loaded. It’s concentrated in the first third of the night, which is why the first few hours feel so heavy and why they’re the hardest to give up.
REM sleep, where most vivid dreaming happens, runs the other way. The first REM period of the night is short, often under ten minutes. The last one can stretch to thirty, forty, sometimes sixty. Your longest, densest, most cinematic dreaming is waiting for you in the final stretch of a full night.
So if you sleep six hours instead of eight, you are not trimming a uniform slice off the whole night. You’re cutting the end, and the end is where the dreams live.
There’s a second layer to this. Under sleep restriction, the body does not sacrifice everything equally. In a 2025 study of recurring short sleep, deep slow-wave sleep held its ground even when total sleep time dropped from about seven and a half hours to under six. REM shrank along with the night. The body defends deep sleep first, which means the proportion of your night given to REM falls as the night gets shorter.
I want to be careful here, because this is exactly the point where wellness writing usually overreaches. No study has taken a group of people, chopped ninety minutes off the end of their night, and measured precisely how much REM disappeared. What we have is the architecture, and the architecture points one way. Read it as a strong inference, not a proven number.
The Part That Isn’t True: Sleeping Longer Doesn’t Make You Remember
Here is where it gets genuinely surprising.
If short sleep were the main reason people don’t recall dreams, then longer sleepers should remember more. They don’t.
In a 2025 survey of 1,151 preteens published in the Journal of Sleep Research, sleep duration showed no relationship at all with dream recall frequency. The researchers ran a Bayesian analysis that actively supported the null result, which is a stronger statement than simply failing to find an effect. What did correlate with every single dream measure was something else entirely: nocturnal awakenings.
A 2025 adult study makes the same point from a different direction. Researchers in Italy followed 204 people aged 18 to 70 for fifteen days at home, collecting about 2,900 morning voice reports with wrist actigraphy throughout and portable EEG on a subgroup. On average people recalled a dream with actual content about four mornings a week.
The things that predicted recall were not what you’d guess. Age wasn’t among the predictors of whether you remembered anything at all, though it did predict whether a recalled dream had actual content or was just a vague sense of having dreamed. Sex wasn’t a predictor. Neither verbal nor visual memory was.
What did predict it: how much lighter sleep a person got, their proneness to mind-wandering during the day, and, remarkably, their attitude toward dreaming. People who found dreams interesting remembered more of them.
Notice that the amount of REM was not the winning variable. Lighter sleep was.
So What Actually Erases a Dream
The best explanation researchers have is called the arousal-retrieval model, and it is almost insultingly simple.
A dream is not stored while you are asleep. To make it into long-term memory, there has to be a moment of waking awareness while the dream is still there. No moment, no memory. It happened, and then it was gone.
This explains a finding that otherwise makes no sense. When you wake people up in a lab directly out of REM, they report a dream about 82% of the time. Out of non-REM sleep, roughly half that, though the non-REM figures vary enormously between studies depending on how loosely “a dream” is defined. Either way, the stage you surface from matters enormously.
And it explains the awkward corollary. In one small study of 36 people, those who reported remembering dreams often had no more REM awakenings than those who rarely did. What they had was more awakenings overall, and specifically more — and longer — surfacings out of lighter stage-two sleep. Their sleep was a little more porous. In people with insomnia, elevated dream recall disappeared entirely once researchers accounted for how often they woke in the night.
Which leads somewhere uncomfortable and worth saying plainly: remembering a lot of dreams is not by itself a sign of good sleep. Where it comes packaged with waking up exhausted, it is often a sign of interrupted sleep. That deserves its own article, and it gets one next, because in Traditional Chinese Medicine dream-disturbed sleep is treated as a symptom rather than a trophy.
But it also means the alarm story needs rewriting. The alarm isn’t only shortening the dream-rich hours. It’s changing the quality of the transition out of them. You go from a long REM period to fully vertical in about four seconds, and then straight into a screen. There is no soft edge, no drifting, no minute of lying there with the images still hanging around. The dream needed that minute.
The Doorway Is About Sixty Seconds Wide
Sleep researchers call the fog after waking sleep inertia. Performance usually climbs back toward baseline within fifteen to thirty minutes, though full recovery can take longer, and prior sleep debt makes it worse.
Most of us treat those first minutes as dead time to be powered through. For dreams, they are the entire window. What’s in your mind on waking is sitting in short-term memory and will be overwritten by the first genuinely engaging thing you look at. Your phone is engineered to be that thing.
This is also, I think, why the “attitude toward dreaming” finding is less mystical than it sounds. If you believe dreams matter, you pause for a moment on waking and go looking. That pause is the mechanism.
What the Hun Was Doing All Night
Traditional Chinese Medicine has its own account of this, and I want to hold it a little loosely, because it is a philosophical and observational system rather than a physiological one. But it maps onto the experience in a way I’ve never quite gotten over.
In the classical texts, the Shen (神), the conscious spirit, is housed by the Heart, and the Hun (魂), sometimes translated as the ethereal soul, is housed by the Liver. The Ling Shu puts it as 肝藏血，血舍魂 — the Liver stores the Blood, and the Blood houses the Hun.
The image the tradition uses is one of daily movement. The nineteenth-century physician Tang Zonghai wrote that the Hun resides in the eyes during the day and returns to the Liver the moment we close them. At night it comes home; when the Liver Blood is abundant, it rests there quietly, and sleep is deep and undisturbed. When Liver Blood is deficient, the Hun has nowhere to settle and drifts, and the night fills with restless, tiring dreams.
In the organ clock, 1am to 3am belongs to the Liver. The hours the tradition marked as the Hun’s homecoming are the same hours in which modern polysomnography shows those long REM periods gathering.
I’m not claiming the classical physicians were describing REM sleep. They had no way to see it. But they were watching the same nights we are, and they concluded that what happens in the last stretch of darkness is where the restoration actually occurs. So does the sleep lab.
There’s one more piece of the image worth keeping. The Hun returns to the eyes at waking. In the tradition, that transition is meant to be gradual, and an abrupt one is understood to leave the whole system rattled. Anyone who has been jolted out of deep sleep by a phone alarm knows exactly what that feels like without needing a theory for it.
How to Get Your Dreams Back
Nothing here requires believing anything in particular. It’s mostly about protecting the end of the night and softening the exit.
- Give yourself the last cycle. This is the one non-negotiable. Going to bed thirty to forty-five minutes earlier restores a stretch of the night that is disproportionately REM. Nothing else on this list works without it.
- Wake gently if you possibly can. A sunrise-simulating lamp, a gradual alarm, or a wake window rather than a fixed jolt. If your schedule allows a natural waking even one or two days a week, that’s when dreams tend to come back first.
- Stay still for sixty seconds. Don’t move, don’t open your eyes fully, don’t reach for anything. Let whatever is there surface. This single habit changes more than anything else on this list.
- Write down one fragment. Not the whole dream. One image, one word, one feeling. Notebook, not phone, because the phone will hand you eleven other things.
- Get curious on purpose. The research says attitude predicts recall, which means interest is a skill rather than a personality trait. People who start paying attention generally start remembering within a week or two.
- Protect the whole night, not just the hours. A glass of wine in the evening suppresses REM early and fragments the second half. If you’re trying to get your dreams back, that’s the first variable to test.
- Watch the pattern, not one night. If you use a wearable, look at how your longest REM stretches line up with your wake time over a couple of weeks. Most people find their alarm sitting right on top of them.
Where Karviva Fits In
The honest answer is that no drink will hand you your dreams back. What supports this is the shape of your evening and the length of your night, and I’d rather say that clearly than sell you something else.
What a good evening ritual does is make the wind-down repeatable, and repeatability is what actually moves bedtime earlier. That’s the thinking behind the Karviva Wellness Reset: a whole-plant blend with 5 to 8 grams of naturally occurring sugar per serving and no added sugar, with herbs like schisandra, which classical materia medica place in the Heart and Kidney and use in the formulas for restless sleep, and goji berry, a Liver-and-Kidney herb long associated with the restorative, yin side of the cycle. These are foods and beverages, not sleep aids, and I won’t pretend otherwise. But as the thing you reach for at nine instead of a second glass of wine, it supports the hours where the dreaming actually happens.
Your dreams didn’t go anywhere. They’re still running, every night, for about two hours. You just have to arrive at the end of the night, and stay still long enough to notice.
