Why do we dream?

A strange night

Imagine this. You are late for a train, except the station is your old school. Someone you have not seen for fifteen years is there, but neither of you finds that strange. You open a door and suddenly you are driving. The road tilts upwards like a wall. You are worried about a suitcase you do not remember packing.

Then you wake up.

For a few seconds the whole thing feels important. Ten minutes later, most of it has gone.

Dreaming is one of the oddest ordinary things humans do. We spend part of almost every night having experiences that can contain places, voices, fear, sex, arguments, dead relatives, impossible architecture and people who somehow change identity halfway through a conversation. While it is happening, much of it feels real enough.

We have been trying to explain this for a very long time. Dreams have been treated as messages from gods, omens, buried wishes, symbolic codes and windows into the unconscious. Modern sleep science has given us much better tools for observing the sleeping brain. It has not given us one agreed answer to the obvious question.

Why do we dream?

What sleep research sees

The first useful correction is that dreaming is not simply another word for REM sleep.

Rapid eye movement sleep is strongly associated with vivid, elaborate dreaming, and REM dreams tend to be longer, more story-like and more bizarre. But researchers can also collect dream reports after waking people from non-REM sleep. Dreaming appears across sleep states, although its character can change.

That matters because scientists cannot put a dream on a screen and inspect it directly. They can measure the sleeper instead.

In a sleep laboratory, researchers can record electrical activity from the brain, eye movements, muscle tone, heart rate and other physiological signals. They can wake somebody at a chosen point in the night and ask what, if anything, was happening in their mind immediately before they woke. They can compare those reports with the sleep stage and patterns of brain activity that came before them.

This has produced a much richer picture of sleep than the old idea that the brain simply switches off for eight hours. Sleep moves through different states. Neural activity changes. Memories are processed. Emotional systems behave differently. Parts of the brain associated with visual imagery and emotion can be active while some regions involved in executive control behave differently from waking life.

It also creates a basic scientific problem. The dream itself is private. Researchers usually know about it because the dreamer reports it afterwards, and dream memory is notoriously fragile.

You have probably experienced the problem personally. A dream can seem enormous when you first open your eyes, then collapse while you are trying to explain it. Details disappear. Sequences become uncertain. What remains is already a reconstruction.

So sleep science has become increasingly good at studying the conditions in which dreams happen without acquiring a simple meter labelled PURPOSE OF DREAMING.

Several possible answers

One family of ideas connects dreaming with memory.

Sleep is involved in memory processing and consolidation, and researchers have long asked whether the conscious experience of dreaming is part of that work. Bits of waking life certainly appear in dreams, although rarely as neat recordings. A place from yesterday might be combined with somebody from childhood and a problem that does not exist. The sleeping brain seems perfectly happy to remix its source material.

That makes an appealing story: perhaps dreams are what it feels like while memories are being sorted, strengthened or integrated with older knowledge. There is evidence linking sleep, memory and dream content, but that does not prove the dream narrative itself is necessary for memory consolidation. The machinery of sleep could be doing useful work while the strange story is simply what consciousness experiences alongside it.

Another set of ideas focuses on emotion.

Dreams are often emotionally loaded. Fear, anxiety and social situations turn up with impressive enthusiasm. REM sleep is associated with the processing of emotional memories, and researchers have proposed that dreaming may contribute to emotional regulation, rehearsal or the reworking of difficult experiences.

Again, the interesting word is may.

Reviews of the field still describe the functional role of dreaming as unresolved. Emotional regulation is a plausible hypothesis, not a licence to look at a nightmare and announce that your brain has diagnosed the problem.

There are other possibilities. Threat-simulation theories suggest dreams may provide a kind of rehearsal space for danger. Other accounts emphasise creativity and loose association, arguing that the unusual state of the sleeping brain allows ideas and memories to combine in ways that waking thought would reject. Some researchers take a less romantic position: dreaming may not have a separate adaptive function at all. It could be a by-product of the brain activity required for other processes during sleep.

That last possibility is worth sitting with because humans are very good at assuming that a vivid experience must contain a message.

A dream about your teeth falling out does not arrive with an official translation. There is no scientifically established universal dictionary in which teeth equal anxiety, water equals emotion and being chased means one specific unresolved issue. Personal associations can obviously matter to the person having the dream, but that is different from a code that applies to everybody.

Nor are dreams reliable predictions. The fact that a dream later resembles something that happens is much less mysterious when you remember how many dreams are forgotten, how flexible dream memories can be and how good human brains are at noticing a striking match after the event.

The uncertainty is not a failure of the science. It is the interesting part.

We know that the sleeping brain is active. We know dreaming can occur in REM and non-REM sleep. We know waking experiences, memories and emotions can find their way into dreams. We know sleep itself performs important biological and cognitive work.

What we do not yet know is whether dreaming has one function, several functions, different functions in different sleep states, or no independent function at all.

There may not even be a single answer hiding at the bottom of the problem. Asking what dreams are for could eventually turn out to be like asking what thoughts are for. Some solve problems. Some rehearse conversations. Some remember. Some wander. Some are noise.

And some leave you standing in a school that is also a railway station, desperately looking for a suitcase that never existed.

Sleep science can observe the sleeping brain in extraordinary detail. The dream itself is still reported from the inside.