Dream Research
Everything worth knowing about dreams, in one page
After seventy years of laboratory research the confident claims are fewer than the popular literature suggests. Here is what has actually held up, and what has not.
This site has covered a great deal of ground. It is worth setting out, in one place, what the evidence actually supports — and what it does not, including several things that are repeated constantly.
What is well established
Everyone dreams. People who say they do not are people who do not recall. Woken from REM in a laboratory, they report dreams at close to normal rates.
Dreaming is not confined to REM. Mentation occurs in non-REM sleep too, and lesion studies show REM and dreaming are doubly dissociable — each can occur without the other.
Dream content is continuous with waking life. The best-supported general principle in the field. Dreams reflect preoccupations, relationships and concerns. They are mostly mundane, mostly social, and mostly about people you know.
Dream content skews negative. Aggression outnumbers friendliness, misfortune outnumbers good fortune, negative emotion outnumbers positive roughly two to one.
The prefrontal cortex is deactivated during REM, which explains the uncritical acceptance of absurdity, the poor working memory, and the failure of reality checks performed inside dreams.
Dream imagery is built from stored perceptual material. Congenitally blind people do not dream visually; those blinded later do. Damage to visual association cortex removes visual dreaming.
Lucid dreaming is real and verifiable. Demonstrated by pre-agreed eye signals recorded during unambiguous REM, and since used for two-way communication with sleeping participants.
Sleep consolidates memory. Replay of waking neural sequences occurs during sleep and predicts subsequent performance. Targeted memory reactivation — cueing during slow-wave sleep — measurably biases what gets consolidated.
Nightmares are treatable. Imagery rehearsal therapy has good evidence and takes a few sessions.
That dreams predict the future. That symbols have fixed universal meanings. That you can learn new information from audio played during sleep. That everyone dreams in black and white, or that they used to. That dreams are disguised wish fulfilments. That you die if you die in a dream. That sleepwalkers must not be woken. That the Senoi had an elaborate dream-control culture. That dreaming has a demonstrated function.
What is contested
Why we dream. Threat simulation, emotion regulation, memory consolidation and by-product accounts all have supporters and none has won. It is entirely possible that dreaming has no function and is a by-product of processes that do.
Whether the glymphatic clearance finding holds. The original result — that clearance increases dramatically during sleep — has been directly contradicted by more recent work, and the mechanism is disputed.
Whether animals dream. Their brains do what dreaming brains do. Whether that is accompanied by experience is not answerable with current methods.
Whether segmented sleep was the natural human pattern. The historical evidence is real; the anthropological evidence from societies without electric light does not support it.
How much dream reports can be trusted at all. The dream colour literature demonstrated that a basic feature of reported experience was shaped by cultural assumption. The same concern applies to everything else that gets coded.
What is worth doing
If you want to remember your dreams: keep a journal, and do not move on waking. Those two things account for most of the achievable improvement.
If you want to have lucid dreams: journal first, then wake back to bed, then MILD. Devices add little that the practices around them do not.
If you have nightmares: ask for imagery rehearsal therapy by name.
If you have insomnia: ask for CBT-I, not a sedative.
If you are exhausted despite adequate hours, snore, or have been told you stop breathing: get assessed for sleep apnoea. It is common, under-diagnosed, and treatable.
If you act out your dreams and you are over fifty: see a neurologist. This one matters more than anything else on the list.
How to think about your own dreams
The defensible position sits between two errors.
The first is treating dreams as messages requiring decoding, which has no evidential basis and a long history of confident nonsense.
The second is dismissing them as meaningless noise, which does not fit the continuity evidence — dream content demonstrably tracks what is going on in a life.
The middle position: dreams are an indicator. They show what has been occupying you, weighted roughly by how much. The emotion is more informative than the plot. Recurring themes mark unresolved situations. And none of it tells you what to do, because the dreaming brain has no judgement, no social filter, and no access to information you do not already have.
Which is less exciting than a message from beyond, and considerably more useful than a dictionary.
