# Why do we dream?

> Maybe you woke with it still around you: a house that was yours and somehow wasn't, a person who had no reason to be there, a feeling you couldn't quite name. By breakfast, when you try to tell someone, it has thinned to a few flat sentences, and you're left wondering where it came from.

**In short:** Nobody knows for certain, and researchers still disagree about whether dreaming does anything itself[^1][^2]. What's clearer is what dreams are made of: fragments of recent days mixed with older ones[^3][^1], in a brain where a region for deliberate reasoning is among the least active[^4]. Sleep plainly helps memory, and a dream may simply be a trace of that work.

A dream that lingers has its own feel. While you were inside it everything made sense: the corridor that opened into your grandmother's kitchen, the friend from school who was somehow running the meeting, the hurry to find a ticket that didn't exist. Only now, sitting up in the early light, does the logic come apart, and what lasts longest is often the mood, which goes with you through the morning after the plot has gone.

The people who study dreams say plainly that whether dreaming has a purpose at all is still unresolved[^1][^2], yet the main ideas fit together better than they first seem. With the senses closed and the brain's reasoning turned down[^4], the sleeping mind pulls up the day's memories and mixes them with older ones[^1], and that mixing would go a long way toward explaining a dream's odd cast and borrowed rooms. Where the theories part is over the dream itself, and whether the experience does anything for you or simply comes along with the work.

## Why did my dream borrow from yesterday?

If you've noticed last night's dream helping itself to something from the day before, you're seeing what researchers have measured. In one study, people kept a log of their days for two weeks, wrote down their dreams as well, and then went through the two side by side, looking for anything from the day that had turned up in the night. Sixty-five percent of their dreams contained something from recent waking life, yet no more than one or two percent replayed an event as it had happened[^3]. A later diary study found the same pattern[^5], with fragments of the dreamer's own life in most dreams and a memory relived just as it was almost nowhere.

One woman in that first study had spent her day deciding which leftover muffins to take home from Starbucks, and that night she dreamt of going shopping with her father into a store filled with muffins from floor to ceiling, unable to decide which one she wanted[^1]. Nothing in the dream happened as it had in the day, yet the muffins and the indecision came straight from it, mixed in with a father and a shopping trip from somewhere else entirely.

What gets in first tends to be whatever is new and on your mind. People who spend a night in a sleep laboratory often dream of the laboratory itself: across 813 reports from REM sleep, about one dream in five plainly featured it[^1], with the experimenters and the electrodes on their heads turning up in the dream. After hours of playing Tetris, most people found Tetris images in front of them as they drifted off[^1], the blocks still dropping behind their closed eyes.

Some of the day comes back later, too. In small studies, a day's events tended to turn up in the next night or two, fade, then return about five to seven days later[^6], a delayed echo researchers call the dream-lag.

**When a day's events come back in dreams**

Counted in days after something happens: a stretch of closer matches in the next night or two, fewer matches three or four days on, then a second stretch about five to seven days on. Twenty people kept a diary for over a week, then spent two nights in a sleep laboratory, where they rated every dream against every diary day. The late echo showed up in REM dreams but not in lighter non-REM ones[^6].

- **The next night or two**

- **The dream-lag, a delayed echo**

- **The day itself**

**Dreams are full of hidden symbols waiting to be decoded**

A dream so clearly draws on your own life, and leaves such a strong feeling behind, that it seems it must be telling you something, and dream dictionaries promise to say what.

The evidence supports something more modest and still interesting. Researchers from opposite camps agree that dream content is organized and plainly shaped by waking life[^7][^1]. A dream tends to reflect what's on your mind and what matters to you, which is part of why it feels so full of meaning. Even so, when researchers have tried to predict which of the day's events will turn up in a particular dream, they have often failed[^1]. Whether particular images have fixed meanings, the way a dictionary sets them out, is a question the research simply hasn't tested.

## Why do dreams feel so real while they're happening, and so strange after?

Inside a dream you rarely notice how strange it is. Brain imaging hints at why: in people in REM sleep, the parts of the brain that handle feeling and memory are busy, and so are the seeing parts at the back[^4]. A region at the front involved in deliberate reasoning is among the least active[^4]. One explanation researchers offer is that with the front of the brain turned down, loose associations are free to run[^8], which would fit the way dreams from REM sleep tend to be the most vivid, bizarre and emotional[^8]. The pictures and the feelings are fully lit, while the part of you that would ask why your old teacher is driving the bus has stepped out of the room.

Dreams come outside REM as well: people woken from non-REM sleep in sleep laboratories report a dream up to seventy percent of the time[^9]. Even so, the dreams gather toward morning, when REM sleep fills more of the night[^2][^6]. If you're wondering what these stages are and how they take turns through the night, that's [What happens while we sleep?](https://nitez.com/stages-of-sleep), and REM has its own question in [What is REM sleep?](https://nitez.com/rem-sleep)

**How can researchers tell when someone is dreaming?**

Mostly they wake people and ask them.

A team led by Francesca Siclari went further, recording sleepers in a laboratory under caps of 256 electrodes and waking them again and again through the night[^9], each time asking whether they had been experiencing anything just before. They woke thirty-two newcomers 233 times in all, and seven trained sleepers 815 times[^9].

Whether someone had been dreaming, in REM or non-REM sleep, tracked activity in a 'hot zone' toward the back of the brain[^9]. In a final test the researchers watched that zone as people slept and, from it alone, predicted whether a person woken from non-REM sleep would report a dream. The same signal appeared when people knew they had dreamt but couldn't remember what about. This is one group's work so far. It shows where in the brain dreaming seems to happen, and it leaves open the question of what dreaming is for.

## Why do most dreams slip away so fast?

You may have felt a dream going as you reached for it, the details thinning while you were still lying there. One leading account says a dream is remembered only if we wake briefly while it's happening or soon after[^10], and people who remember a lot of dreams do report waking more in the night. In a laboratory in Lyon, eighteen people who often recalled their dreams and eighteen who rarely did heard first names played among tones while they slept, and the frequent recallers' brains reacted more strongly[^10], as if their sleep left the door to the world a little further open.

Much of what we dream seems to go unrecorded. People woken in the laboratory sometimes know they were dreaming yet can't say what about[^9], and people who keep a dream journal say they remember more as the habit grows[^2], both of which suggest we dream a good deal more than we recall. The dream you woke with this morning was a piece of a longer night, the piece you happened to wake into.

## Are dreams the brain sorting its memories?

This is the idea with the most behind it, and its story begins with rats. As a rat runs a maze, particular cells in its brain fire in order along the route, and during sleep those same cells fire again in sequence, as if retracing the way, faster than in life and rarely exactly the same[^1]. Is the rat dreaming of the maze? Nobody knows whether it experiences anything at all. What the recordings do show is a sleeping brain going back over its day, and in people, study after study has found that sleep helps memory[^1].

The harder question is whether dreams take part. At Harvard, ninety-nine students learned their way through a virtual maze, then either napped for ninety minutes or stayed awake, and were tested again five hours later. Four of the nappers mentioned the maze when asked about their dreams, and they improved about ten times as much as the other nappers[^11]. None of their dreams replayed running the maze, though: in two of the dream reports the maze's music was playing[^11], and some of the rest were set in other maze-like places. Thinking about the maze while awake didn't predict improvement in the same way.

That result came from four people, and it looked weaker when the same team repeated the study overnight: the maze dreamers again improved more, though the effect was about half the size it had been the first time[^12]. The researchers were candid about the rest. They noted that other labs have tried and failed to find the effect[^12]. They also pointed out that the maze dreamers were the people who had struggled most with the maze at first, and once they allowed for that, in an analysis of seventeen people, the dream effect was no longer statistically significant[^12].

Another kind of evidence comes from people whose memory system is damaged. Researchers woke four men with damage to both hippocampi, the brain's hub for remembering events, through the night at home, along with ten healthy men of similar ages. They woke the patients just as often, yet the patients reported fewer dreams with thinner content, and one reported none at all[^13]. The group is tiny, and earlier studies of damage like this were mixed. What it suggests is that dreaming draws on the memory system, which isn't the same as dreaming serving it.

The memory researchers' own reading is a careful one: a dream may be the visible trace of useful memory work without the dream itself doing the work[^11][^1], and the way it mixes fragments into new and sometimes bizarre combinations may reflect how sleep links new memories to old ones. That would fit the student who heard the maze's music and the woman with her store full of muffins, and perhaps the aunt, years gone, who walks into a dream as if she had only stepped out for a minute and seems strange only once you're awake.

## Do dreams help us work through our feelings?

Some mornings a worry you took to bed has lost its edge, or a hurt doesn't sting the way it did last night, as if something worked on it while you slept.

Researchers have asked whether REM sleep is what does that, and so far the studies point both ways. At Berkeley, young adults looked through a set of emotional pictures and rated them. Twelve hours later they came back and saw the same pictures again, some after a night's sleep and some after a day awake. For those who had slept, a part of the brain that handles emotion reacted less to the same pictures and the most intense ratings fell, while over a day awake its response rose[^14]. A team at the University of Massachusetts found the reverse. Over a day awake, reactions to negative pictures faded, but over a night's sleep they were largely kept, and more REM sleep went with more of the feeling staying[^15]. And what were the dreams doing all this time? Both teams measured sleep and neither collected dreams, so neither can say.

One study did look at the dreams themselves. In Geneva, people who reported more fear in their dreams showed less reactive brain responses to frightening images when awake[^16], and the researchers read this as dreams possibly rehearsing feelings rather than only replaying them. It's a correlation, so it can't show the dreams caused the difference, and there's also the finding that most dreams contain negative feelings, failures or misfortunes, and rarely solve problems in any obvious way[^1].

## Are dreams practice for danger?

Perhaps you know the dream where something is coming after you and your legs won't run, or the one where you sit down to an exam you haven't studied for, years after you last took one. In 2000 the Finnish researcher Antti Revonsuo proposed a reason for dreams like these: that dreaming evolved to rehearse spotting and escaping danger[^7], so a night spent being chased would be a kind of practice for the day. His case was built on what dreams contain. Dream content is organized rather than random, he argued, and he pointed to the dreams of children, trauma survivors and hunter-gatherers. Most of the support since has come from his own group, through studies of what people's dreams contain[^17]. That's where we'd hesitate. However many chases turn up, counting what dreams contain can't show whether rehearsing in them helps anyone once they're awake.

A newer version from the same group says dreams rehearse our relationships. To test it, fifteen students wrote down their dreams and their waking days at home, and their dreams did include more social events than their accounts of waking life, though other predictions failed[^18], which leaves the idea with partial support.

## Or are dreams a side effect with no purpose?

You may have heard the confident answer that dreams are noise and nothing more, the brain's static with nothing behind it. The researchers who have reviewed the field still count the by-product view among the live options[^1][^2]. Its best-known form came in 1977, when the Harvard psychiatrists Allan Hobson and Robert McCarley proposed that dreams begin as bursts of activity from the brainstem during REM sleep, which the forebrain stitches into a story from stored memories[^19]. They suggested too that shifts in the brain's chemicals during sleep could explain why dreams are forgotten[^19].

Any account of dreaming, this one included, has to make room for two later observations. Sleepers woken from non-REM sleep, outside REM altogether, often report dreams too[^9]. And in early case studies, which a review has gathered, patients with damage to a region deep in the front of the brain stopped dreaming[^8]. Hobson and McCarley had already given the forebrain the job of building the story, so their theory has room for the second. Our own reading of those early case studies is simply that dreaming needs more of the brain than the brainstem alone.

In 2021 Erik Hoel offered a way to give the strangeness itself a job. Engineers add noise when training artificial neural networks so they don't learn their examples too narrowly, and he suggests the oddness of dreams might likewise help the brain generalize from the day's learning[^2], an idea whose predictions are still to be tested.

Most studies of dreams are small, and the ones on emotion measured sleep rather than the dreams themselves.

- **Dreams draw on recent life but rarely replay it** good-evidence

- **Dreams happen in non-REM sleep as well as REM** good-evidence

- **Dreams from REM sleep tend to be the most vivid, bizarre and emotional** good-evidence

- **Whether dreaming about a task goes with learning it better: findings differ** still-debated

- **Whether REM sleep weakens the emotional charge of memories: studies disagree** still-debated

- **Whether dreams evolved to rehearse threats: not yet clear** still-debated

- **Whether dreams are a by-product with no purpose of their own: not yet clear** still-debated

- **Whether the strangeness of dreams helps the brain generalize: an untested idea** not-well-studied

## So how do the theories fit together?

Two things are well supported, and they make a fair place to begin: sleep helps the brain keep what it has learned[^1], and dreams are built from memory, recent days mixed with older ones[^3][^1]. One way to read them together, and it is our reading rather than a settled finding, is that a dream is what that night work on memory feels like from the inside. What's still open is whether the dream you remember does something itself or is the glow from a machine working in the dark. The memory researchers favor the glow, a trace of memory work the dream doesn't itself do[^11], and the older by-product idea of brainstem bursts stitched into a story[^19] gives the dream no job of its own either. The threat and emotion theories, and Hoel's, give the dream itself a job, and the nearest thing to evidence for them is the Geneva finding that more fear in dreams went with less reactive waking responses[^16].

So what do we make of it? The studies tying the dream itself to anything useful are few and small, and none yet shows that a dream caused a benefit. For now we'd side with the memory researchers and the glow. What could change our minds is Hoel's idea, because it gives the strangeness itself a use[^2]. If its predictions turn out right, the dream would have a job of its own after all.

A dream still means something in a plainer sense: it's made from what's on your mind. If your dreams are frightening more than strange, or keep coming back, [Why do I keep having nightmares?](https://nitez.com/nightmares) is about that.

**If you’d like to try something**

Next time a dream stays with you, you could jot down a few words about it and then look back over the last day or two, and again about a week before, for the pieces it borrowed: a place, a worry, a person, a phrase[^3][^6].

The dream that followed you to breakfast, the one that sounded so flat when you tried to tell it, was put together in the dark from your own days and older memories, as your brain went over its night's work. Reason was turned low the whole time. Whether the dream did anything for you is still an open question. Where it came from is often there to be found, in a muffin, a melody or a worry from the day before.

## Related questions

- [What happens while we sleep?](https://nitez.com/stages-of-sleep)
- [What is REM sleep?](https://nitez.com/rem-sleep)
- [Why do I keep having nightmares?](https://nitez.com/nightmares)
- [Why am I still tired after a full night’s sleep?](https://nitez.com/tired-after-sleeping)

## Sources

[^1]: Wamsley EJ. Dreaming and offline memory consolidation. Current neurology and neuroscience reports. 2014;14(3):433. [PubMed](https://pubmed.ncbi.nlm.nih.gov/24477388/) · [PMC](https://pmc.ncbi.nlm.nih.gov/articles/PMC4704085/) · [DOI](https://doi.org/10.1007/s11910-013-0433-5)
[^2]: Hoel E. The overfitted brain: Dreams evolved to assist generalization. Patterns (New York, N.Y.). 2021;2(5):100244. [PubMed](https://pubmed.ncbi.nlm.nih.gov/34036289/) · [PMC](https://pmc.ncbi.nlm.nih.gov/articles/PMC8134940/) · [DOI](https://doi.org/10.1016/j.patter.2021.100244)
[^3]: Fosse MJ, Fosse R, Hobson JA, Stickgold RJ. Dreaming and episodic memory: a functional dissociation? Journal of cognitive neuroscience. 2003;15(1):1–9. [PubMed](https://pubmed.ncbi.nlm.nih.gov/12590838/) · [DOI](https://doi.org/10.1162/089892903321107774)
[^4]: Maquet P. Functional neuroimaging of normal human sleep by positron emission tomography. Journal of sleep research. 2000;9(3):207–31. [PubMed](https://pubmed.ncbi.nlm.nih.gov/11012860/) · [DOI](https://doi.org/10.1046/j.1365-2869.2000.00214.x)
[^5]: Malinowski JE, Horton CL. Memory sources of dreams: the incorporation of autobiographical rather than episodic experiences. Journal of sleep research. 2014;23(4):441–7. [PubMed](https://pubmed.ncbi.nlm.nih.gov/24635722/) · [DOI](https://doi.org/10.1111/jsr.12134)
[^6]: Blagrove M, Fouquet NC, Henley-Einion JA, Pace-Schott EF, Davies AC, Neuschaffer JL, et al. Assessing the dream-lag effect for REM and NREM stage 2 dreams. PloS one. 2011;6(10):e26708. [PubMed](https://pubmed.ncbi.nlm.nih.gov/22046336/) · [PMC](https://pmc.ncbi.nlm.nih.gov/articles/PMC3202556/) · [DOI](https://doi.org/10.1371/journal.pone.0026708)
[^7]: Revonsuo A. The reinterpretation of dreams: an evolutionary hypothesis of the function of dreaming. The Behavioral and brain sciences. 2000;23(6):877–901; discussion 904–1121. [PubMed](https://pubmed.ncbi.nlm.nih.gov/11515147/) · [DOI](https://doi.org/10.1017/s0140525x00004015)
[^8]: Scarpelli S, Bartolacci C, D'Atri A, Gorgoni M, De Gennaro L. The Functional Role of Dreaming in Emotional Processes. Frontiers in psychology. 2019;10:459. [PubMed](https://pubmed.ncbi.nlm.nih.gov/30930809/) · [PMC](https://pmc.ncbi.nlm.nih.gov/articles/PMC6428732/) · [DOI](https://doi.org/10.3389/fpsyg.2019.00459)
[^9]: Siclari F, Baird B, Perogamvros L, Bernardi G, LaRocque JJ, Riedner B, et al. The neural correlates of dreaming. Nature neuroscience. 2017;20(6):872–878. [PubMed](https://pubmed.ncbi.nlm.nih.gov/28394322/) · [PMC](https://pmc.ncbi.nlm.nih.gov/articles/PMC5462120/) · [DOI](https://doi.org/10.1038/nn.4545)
[^10]: Eichenlaub JB, Bertrand O, Morlet D, Ruby P. Brain reactivity differentiates subjects with high and low dream recall frequencies during both sleep and wakefulness. Cerebral cortex (New York, N.Y. : 1991). 2014;24(5):1206–15. [PubMed](https://pubmed.ncbi.nlm.nih.gov/23283685/) · [DOI](https://doi.org/10.1093/cercor/bhs388)
[^11]: Wamsley EJ, Tucker M, Payne JD, Benavides JA, Stickgold R. Dreaming of a learning task is associated with enhanced sleep-dependent memory consolidation. Current biology : CB. 2010;20(9):850–5. [PubMed](https://pubmed.ncbi.nlm.nih.gov/20417102/) · [PMC](https://pmc.ncbi.nlm.nih.gov/articles/PMC2869395/) · [DOI](https://doi.org/10.1016/j.cub.2010.03.027)
[^12]: Wamsley EJ, Stickgold R. Dreaming of a learning task is associated with enhanced memory consolidation: Replication in an overnight sleep study. Journal of sleep research. 2019;28(1):e12749. [PubMed](https://pubmed.ncbi.nlm.nih.gov/30091247/) · [PMC](https://pmc.ncbi.nlm.nih.gov/articles/PMC6338510/) · [DOI](https://doi.org/10.1111/jsr.12749)
[^13]: Spanò G, Pizzamiglio G, McCormick C, Clark IA, De Felice S, Miller TD, et al. Dreaming with hippocampal damage. eLife. 2020;9:e56211. [PubMed](https://pubmed.ncbi.nlm.nih.gov/32508305/) · [PMC](https://pmc.ncbi.nlm.nih.gov/articles/PMC7279885/) · [DOI](https://doi.org/10.7554/eLife.56211)
[^14]: van der Helm E, Yao J, Dutt S, Rao V, Saletin JM, Walker MP. REM sleep depotentiates amygdala activity to previous emotional experiences. Current biology : CB. 2011;21(23):2029–32. [PubMed](https://pubmed.ncbi.nlm.nih.gov/22119526/) · [PMC](https://pmc.ncbi.nlm.nih.gov/articles/PMC3237718/) · [DOI](https://doi.org/10.1016/j.cub.2011.10.052)
[^15]: Baran B, Pace-Schott EF, Ericson C, Spencer RM. Processing of emotional reactivity and emotional memory over sleep. The Journal of neuroscience : the official journal of the Society for Neuroscience. 2012;32(3):1035–42. [PubMed](https://pubmed.ncbi.nlm.nih.gov/22262901/) · [PMC](https://pmc.ncbi.nlm.nih.gov/articles/PMC3548452/) · [DOI](https://doi.org/10.1523/JNEUROSCI.2532-11.2012)
[^16]: Sterpenich V, Perogamvros L, Tononi G, Schwartz S. Fear in dreams and in wakefulness: Evidence for day/night affective homeostasis. Human brain mapping. 2020;41(3):840–850. [PubMed](https://pubmed.ncbi.nlm.nih.gov/31663236/) · [PMC](https://pmc.ncbi.nlm.nih.gov/articles/PMC7267911/) · [DOI](https://doi.org/10.1002/hbm.24843)
[^17]: Valli K, Revonsuo A. The threat simulation theory in light of recent empirical evidence: a review. The American journal of psychology. 2009;122(1):17–38. [PubMed](https://pubmed.ncbi.nlm.nih.gov/19353929/)
[^18]: Tuominen J, Stenberg T, Revonsuo A, Valli K. Social contents in dreams: An empirical test of the Social Simulation Theory. Consciousness and cognition. 2019;69:133–145. [PubMed](https://pubmed.ncbi.nlm.nih.gov/30769273/) · [DOI](https://doi.org/10.1016/j.concog.2019.01.017)
[^19]: Hobson JA, McCarley RW. The brain as a dream state generator: an activation-synthesis hypothesis of the dream process. The American journal of psychiatry. 1977;134(12):1335–48. [PubMed](https://pubmed.ncbi.nlm.nih.gov/21570/) · [DOI](https://doi.org/10.1176/ajp.134.12.1335)

---
https://nitez.com/why-we-dream · How sleep works · updated 2026-10-03