Why do we dream? What science knows, and what it still doesn't

Why do we dream? Where the question stands

Modern dream science really began in 1953. Eugene Aserinsky and Nathaniel Kleitman noticed that sleepers' eyes darted under their lids at regular intervals, and that people woken during those periods often described vivid dreams. That stage became known as REM sleep (rapid eye movement).

For decades, dreaming and REM were treated as almost the same thing. We now know that is too simple. People woken from non-REM sleep also report dreams, usually shorter and less story-like. In a 2017 study, Francesca Siclari and colleagues found that activity in a region at the back of the brain predicted whether someone would report a dream, whatever the sleep stage.

Science measures when we dream and which brain areas take part. Why we dream is still a matter of theories. Here are the main ones, and where they fall short.

Are dreams the brain making sense of noise?

In 1977, psychiatrists J. Allan Hobson and Robert McCarley proposed the activation-synthesis hypothesis, one of the first theories built on brain physiology rather than psychoanalysis.

The core idea

During REM sleep, the brainstem sends bursts of activity to the rest of the brain. The cortex receives these signals and does what it always does: it tries to make a story out of them.

1. Activation: the brainstem switches the cortex on during REM sleep.
2. Synthesis: the cortex links images, sensations and memories into a scene.
3. The dream: the story you remember, with its jumps and impossible turns.

What supports it

• Brain imaging shows strong activity in emotional and visual association areas during REM sleep.
• The parts of the frontal lobe that check logic are less active, which fits the way we accept absurd scenes.

Its limits

Dreams are less random than this model suggests. Their content often follows our waking concerns, relationships and emotions, a link that researchers such as G. William Domhoff call continuity. Hobson himself later revised the model. Today, few researchers see dreams as pure noise.

Do dreams help us remember?

Here the evidence is strongest, but it concerns sleep more than dreams. Sleeping after learning helps you keep what you learned, as Robert Stickgold summarized in a 2005 review in Nature.

What the sleeping brain does with the day

During sleep, and not only during REM sleep, the brain appears to:

• Replay experiences: in rats, cells that fired while running a maze fire again in the same order during sleep.
• Link new and old memories: recent information is tied into what you already know.
• Pick out patterns: sleep seems to help extract rules from separate experiences.

Dreams could be a partial, distorted glimpse of this sorting. That would explain why fragments of the day show up mixed with older memories.

What the research shows

In a 2010 study by Erin Wamsley, Robert Stickgold and colleagues, participants learned a virtual maze, then took a nap. The few who dreamed about the maze improved much more on it afterwards than those who did not.

Striking, but based on few dreamers, and it does not prove the dream caused the gain: it may simply show the brain was working on the task. Memory benefits from sleep even when you remember no dreams.

Why so many dreams are about danger

Being chased, falling, being unable to move: unpleasant dreams are common. In 2000, Finnish researcher Antti Revonsuo offered an evolutionary explanation, the threat simulation theory.

The evolutionary argument

Dreaming would act as a rehearsal space. By simulating dangers during sleep, our ancestors could have practiced spotting threats and reacting to them, at no real risk. Individuals who rehearsed better would have survived better.

This would explain the recurring scenarios:

• Being chased or attacked
• Falling or losing control
• Conflict and social embarrassment
• Being unable to run or move, a feeling close to sleep paralysis

Threats of today

Predators are rare now, but the same scenarios fit modern worries: a speech, a job interview, a conflict, a health concern.

Dreaming that you are late for an exam or turn up naked at work may belong to this kind of rehearsal. It says that something matters to you, not that it will go wrong. The theory is still debated: it explains frightening dreams well, and pleasant or ordinary ones less well.

Do dreams soften our emotions?

Another line of research looks at emotions: a night's sleep could help us keep the memory of a painful event while taking some of the sting out of it.

The "overnight therapy" hypothesis

In 2009, Matthew Walker and Els van der Helm brought this idea together under the name "overnight therapy". Their reasoning rests on a well-established fact: during REM sleep, noradrenaline, a chemical messenger linked to stress, is almost absent from the brain.

• Emotional memories are reactivated in a calmer chemical setting.
• Their emotional charge could fade while the content stays.
• The next day, the event would feel a little less raw. That is the old advice to "sleep on it".

This is a hypothesis, supported by some experiments and not by others. It also concerns REM sleep more than the dreams we remember.

When the system seems to jam

Several conditions come with disturbed REM sleep or dreams:

• Post-traumatic stress: nightmares that replay the event are one of its typical signs.
• Depression: REM sleep often starts earlier in the night and its pattern changes.
• Anxiety: it often goes with more frequent bad dreams.

If nightmares come back several times a week, wreck your sleep or leave you anxious during the day, talk to a doctor or a sleep specialist. Effective treatments exist.

What happens in the dreaming brain

The sleeping brain does not simply switch off. In 1996, Pierre Maquet's team in Liège mapped its activity during REM sleep with PET scans.

Areas that are busier

• The amygdala: involved in emotions, especially fear. This fits the intensity of many dreams.
• The anterior cingulate cortex: linked to attention and emotion.
• Visual association areas: they help build images, which fits the very visual nature of dreams.
• The brainstem and thalamus: they drive and relay the REM state.

Areas that are quieter

• The dorsolateral prefrontal cortex: involved in reasoning and self-monitoring. Its lower activity may explain why we accept that a room changes shape.

Meanwhile, the body is held still by a muscle paralysis specific to REM sleep. That is why you can run in a dream without moving in bed.

A different chemical setting

• Acetylcholine is high, close to waking levels, and keeps the cortex active.
• Noradrenaline and serotonin are very low, which may play a part in the strange logic of dreams and in how hard they are to remember.

The result is a state where emotion and imagery take the lead while critical judgment steps back. These are correlations, not a full explanation of the dream experience.

Do animals dream too?

We cannot ask a cat what it dreamed. We can, however, observe sleep states that look like ours.

Who has REM-like sleep?

• Most mammals studied show phases resembling REM sleep, with notable exceptions and variations.
• Birds also have them.
• Some reptiles might as well: in 2016, a study described two alternating sleep states in the bearded dragon, a lizard.

Replays seen during sleep

• Rats: in a 2001 study, Kenway Louie and Matthew Wilson found that maze-running patterns in the hippocampus came back during REM sleep.
• Songbirds: in young zebra finches, neurons involved in singing fire during sleep in patterns that echo the song (Dave and Margoliash, 2000).
• Dogs: their twitching paws are familiar, but we can only guess what they "see".

A replay is not proof of a dream. But a state this widespread, kept despite its costs (an asleep animal is not feeding and is more exposed), probably serves useful functions. Which ones is still debated.

What dreams teach us about consciousness

Dreams raise a question that goes beyond sleep: how can we live a whole world when our senses are almost cut off from the outside?

Experience without input

• In a dream, you see, hear and feel, with almost no input from your senses.
• You are conscious, but cut off from your surroundings.
• The brain builds a convincing world from its own resources.

The brain, in other words, can produce a rich experience without the outside world feeding it.

What lucid dreams have shown

In a lucid dream, you know you are dreaming. Researchers use this to study dreams from the inside:

• Since the early 1980s, lucid dreamers have signalled to experimenters with agreed eye movements, recorded during REM sleep.
• In 2021, a study led by Karen Konkoly reported two-way exchanges: dreamers answered simple questions, such as maths problems, while still asleep.

These experiments show that dream awareness is not all-or-nothing: you can be asleep and partly aware of your own state.

Waking life as a guided simulation?

Philosopher Thomas Metzinger defends an unsettling idea: waking consciousness would also be a model built by the brain, constantly corrected by the senses. In a dream, that model runs without the senses holding it in check.

Old questions

"How do I know I am not dreaming right now?" The question goes back at least to Descartes, and philosophy still has no settled answer.

Several functions at once?

These theories are not mutually exclusive. Many researchers think that dreaming, or the sleep it goes with, plays several roles at the same time:

• sorting and linking memories;
• softening the emotional charge of some experiences;
• rehearsing threatening situations;
• bringing together unusual associations, sometimes useful for creativity.

Others, like Domhoff, see dreams above all as a kind of thinking under sleep conditions, without a function of its own. Either way, content keeps reflecting what matters to us when we are awake, which is why it is worth noting.

What we still don't know

• Why some dreams are so bizarre and others almost realistic.
• Why we forget most of our dreams so quickly.
• What decides which memories and worries end up in a dream.
• Whether dreams themselves have a function, or are a by-product of sleep.

What you can do with this

Science will not tell you what a particular dream "means". It does suggest a practical method: write down the scene, the emotion and what was going on in your life at the time, then compare over several weeks.

Fictional example: this entry shows how to separate the dream from a possible link with your day.

• Dream: "I was redoing my driving test in an empty car park. The examiner kept changing faces."
• Emotion: "Tense, then strangely calm once the car started."
• Recent context: "First week in a new job, lots to learn."
• Question to keep: "Do 'test' dreams come back when I am learning something new?"

One entry proves nothing. Patterns only show up across many nights, including quiet ones.