A night of sleep runs in cycles of about ninety minutes. Deep sleep comes first; the dream-rich REM periods come at the end of each cycle and grow longer toward morning, while the body is switched off so the dream stays inside.
😴 When you sleep, your brain stays busy. It makes little movies. Those are dreams.
🧠 Your brain looks at all the things from your day and tidies them up.
🌙 Dreams can be silly or scary. They are not real. When you wake up, they are gone!
What is a dream?
A dream is a story your brain makes while you sleep. Your eyes are shut, but your brain is busy. It mixes up things you saw, people you know, and places you have been.
When do dreams happen?
Sleep comes in rounds, like laps around a track. At the end of each round your brain gets very busy, and your eyes flick back and forth under your eyelids. That is dream sleep. Your body stays still so you do not run around in your bed.
Why do we dream?
Nobody knows for sure. Scientists think dreams help your brain sort out what you learned and how you felt. Even babies dream, more than anyone. And almost everyone forgets their dreams by breakfast!
Your Brain Never Clocks Off
Sleep is not your brain switching off. All night it runs through rounds of sleep that each last about an hour and a half. Early in each round you sink into deep sleep. At the end of each round your brain speeds up again, almost as busy as when you are awake, and your eyes dart back and forth behind your closed eyelids. Scientists call this REM sleep, for rapid eye movement, and it is when most dreams happen.
Locked Down for Safety
During REM sleep your brain sends a signal that switches off almost all your muscles. Your arms and legs go limp. That is why you can dream about running from a dinosaur and not kick your sister. Cats whose switch was damaged in an experiment stalked and pounced while fast asleep, acting out their dreams.
What Dreams Might Be For
Nobody has proven why we dream, but there are good clues. Rats that learn a maze replay the same path in their brains while they sleep, as if practicing. People who sleep after learning something remember it better than people who stay up. Dreams also seem to take the sting out of bad feelings, and many dreams rehearse trouble, like being chased or being late, which may be a safe way to practice for real life. Or a dream may just be your brain making a story out of random signals. Perhaps all of these are true.
Gone by Breakfast
You dream for about two hours every night, but you forget nearly all of it. The part of your brain that files new memories is running differently during REM sleep, so a dream usually fades within minutes of waking unless you tell someone or write it down right away.
Try This!
Keep a notebook by your bed. The moment you wake up, before you move, write down anything you remember. After a week, read it back. You will find you dream far more than you thought.
The Night Has a Shape
Sleep is not one flat state but a cycle of about 90 minutes repeated four to six times a night. Each cycle descends through light sleep into slow-wave sleep, when brain waves become large and slow and growth hormone is released, then climbs back up into REM sleep, when the brain's electrical activity looks almost like waking, the eyes flick rapidly under the lids, heart rate and breathing turn irregular, and nearly every skeletal muscle is paralysed by a signal from the brainstem. Early cycles are heavy with deep sleep; later ones are mostly REM, which is why the dreams you remember tend to be the ones from just before the alarm. Adults spend about 20 to 25 percent of the night in REM, roughly two hours. Newborns spend half their sleep there, about eight hours a day, and REM is thought to help build the brain's connections before there is much waking experience to build them from.
Two Discoveries
In 1953 Eugene Aserinsky, a graduate student in Chicago, was watching his sleeping son's eyes with an electrical recorder and noticed bursts of movement that came and went through the night. With his supervisor Nathaniel Kleitman he showed that people woken during these bursts nearly always reported a vivid dream, and that the length of a REM period matched how long the dreamer thought the dream had lasted. Dreams run in real time. A decade later Michel Jouvet in France found the switch that keeps the body still. Cats with a small lesion in the pons stood up during REM sleep, stalked invisible prey, and pounced with their eyes shut, which is as close as anyone has come to watching a dream from outside.
What the Brain Is Doing
Brain scans of sleepers explain a lot about how dreams feel. During REM the amygdala and other emotion centres are more active than when you are awake, the visual areas that build scenes are busy, and the part of the frontal lobe that checks logic and keeps you on task is quiet, which is why a dream can put your teacher on a pirate ship and you accept it without question. Meanwhile the hippocampus, which files new memories, is replaying the day: rats that learned a maze fire the same sequence of place cells in the same order while they sleep, only faster, and people who sleep after practising a task get better at it overnight while people kept awake do not.
Why, Then?
There are four main ideas, and they are not rivals. The memory idea says dreams are the side effect of the brain sorting the day, keeping what matters and discarding the rest. The feelings idea says REM sleep is overnight therapy: the brain replays upsetting things with the stress chemicals turned down, so the memory stays but the sting fades. The rehearsal idea notes that dreams are full of threats, being chased, falling, losing something, and that practising danger in a safe place would have kept our ancestors alive. The story idea, from 1977, says the brainstem fires noisy signals during REM and the cortex, which hates nonsense, weaves them into a plot. Each explains some dreams and none explains all of them. The honest answer is still being worked out, which is a fine thing to be able to say.
Think About It
People who were born blind dream in sounds, touch, and smell, while people who lost their sight later keep seeing in their dreams. What does that tell you about where dream pictures come from?
A State With Its Own Physiology
REM sleep is defined by three signs recorded together: a low-voltage, mixed-frequency EEG that resembles waking, rapid conjugate eye movements, and atonia of the skeletal muscles, with heart rate, breathing and body temperature regulation all becoming irregular. It occupies 20 to 25 percent of adult sleep, about 90 to 120 minutes, arranged in four to six episodes that lengthen from a few minutes in the first cycle to half an hour or more before waking. Aserinsky and Kleitman described the state in 1953; Dement and Kleitman showed in 1957 that awakenings from it yield dream reports about 80 percent of the time and that dream length tracks REM length. Jouvet's lesion studies in cats located the atonia switch in the pons, and REM sleep behaviour disorder, in which that switch fails in people and they act out their dreams, is now recognised as an early sign of Parkinson's disease, sometimes appearing a decade before the tremor. Non-REM dreaming is real but different: David Foulkes found in 1962 that around half of non-REM awakenings produce a report, typically brief, plausible and thought-like.
The Dreaming Brain on a Scanner
Positron emission tomography during REM, first reported by Maquet and colleagues in 1996, shows the pattern that explains dream phenomenology: the amygdala, anterior cingulate and other limbic structures are more active than in waking, along with the extrastriate visual cortex that constructs scenes, while the dorsolateral prefrontal cortex, the seat of working memory, planning and reality testing, is markedly deactivated. Emotion is amplified, imagery is vivid, and the machinery that would say "this cannot be happening" is offline. Hobson and McCarley's activation-synthesis hypothesis of 1977 put the source of the imagery in the brainstem: cholinergic bursts from the pons activate the cortex more or less at random and the cortex synthesises a narrative from the input. Later versions of the model concede that the content is not random at all, since dreams draw on recent memories and long-standing concerns, and the debate has shifted to what the replay is for.
Three Functional Hypotheses
Beyond consolidation, three ideas have real evidence behind them. Revonsuo's threat simulation theory (2000) starts from the content: dreams over-represent threats and negative emotion relative to waking life, and children and trauma survivors dream of threats most of all, which would fit an evolved rehearsal system. Walker and van der Helm's "overnight therapy" account (2009) starts from neurochemistry: REM is the only time of day when noradrenaline is essentially absent from the brain, so emotional memories can be replayed and reconsolidated without the stress signal that accompanied them, which fits the observation that sleep dampens the emotional charge of upsetting images while preserving their content, and that in post-traumatic stress disorder this decoupling seems to fail. Roffwarg's ontogenetic hypothesis (1966) explains the newborn's eight hours of REM as endogenous stimulation that wires the developing brain before the eyes have much to look at. These are complementary; a state as metabolically expensive as REM probably earns its keep in more than one way.
Lucidity, Blindness, and Other Animals
Lucid dreaming, knowing that one is dreaming while it continues, was demonstrated objectively in 1981 when LaBerge's subjects signalled from inside REM sleep with a prearranged sequence of eye movements that showed up on the recording; a 2016 meta-analysis of fifty years of surveys found that about 55 percent of people have had at least one lucid dream. People blind from birth report dreams built from sound, touch and smell, while those who lost sight after about age seven retain visual dreams, so dream imagery is assembled from stored perception rather than generated fresh. Nearly all mammals show REM, and recent recordings have extended the list: zebra finches replay song, octopuses cycle between a quiet state and an active one in which their skin flashes through colour patterns, and rats, as above, rehearse their mazes.
What Is Still Unknown
Whether the subjective dream does any work, or is a by-product of processes that would run identically without it, is the central open question, and hard to test because the dream is available only by report; the relationship between REM dreaming and non-REM dreaming, with their different brain states and different content, is unresolved; and no one has yet shown that any particular dream content predicts any particular waking benefit, which is the experiment every theory needs.
Why Your Kid Asked
Because a dream is the first thing that happens to a child that no one else can see, and because the frightening ones arrive around age three to six, exactly when the question does.
The Short Version
Sleep cycles every 90 minutes or so through deep slow-wave sleep and then REM sleep, when the brain is nearly as active as when awake, the eyes flick, and the muscles are paralysed by a brainstem switch so the dream cannot be acted out. Most vivid dreams happen in REM, about two hours a night, mostly in the longer REM periods toward morning; shorter, thought-like dreams occur in non-REM sleep too. The brain in REM has its emotion centres turned up and its reality-checking frontal cortex turned down, which is why dreams are vivid, emotional, and absurd without seeming so. Why we dream is not settled. The evidence supports overlapping roles: sorting and strengthening the day's memories (the replay seen in sleeping rats and in people who learn better after sleep), taking the emotional charge off upsetting experiences, rehearsing threats, and, in babies, stimulating a brain that is still wiring itself. You forget dreams because REM is a poor state for laying down new memories, not because they were unimportant. That part is normal.
The Numbers
A sleep cycle runs about 90 minutes and repeats four to six times a night, with REM taking about 20 to 25 percent of adult sleep, roughly two hours, while newborns spend about eight hours a day in it, half their sleep. Dream reports on about 80 percent of REM awakenings and 20 to 50 percent of non-REM awakenings. About 55 percent of people have had a lucid dream. Six or so years of a lifetime are spent dreaming. Nightmares are most common between about three and six; night terrors, which are different, peak in the same years.
Two Common Misconceptions Worth Correcting
First, "dreams have hidden meanings that can be decoded": Freud's 1899 book made the idea famous, and dreams do draw on what preoccupies us, but a century of research has found no dictionary of symbols, and the useful question for a child is not "what does it mean" but "what were you feeling in it". Second, "some people do not dream": everyone with REM sleep dreams; people who say they never do simply wake from non-REM sleep or forget within minutes, and can be shown their own dreams in a sleep lab.
When to Pay Attention
Know the difference between a nightmare and a night terror. A nightmare is a REM dream, late in the night; the child wakes fully, is frightened, remembers it, and wants comfort, which is what to give. A night terror comes from deep non-REM sleep in the first hours: the child sits up screaming with open, unseeing eyes, cannot be woken or comforted, settles in a few minutes, and remembers nothing in the morning. Night terrors are alarming to watch and harmless; keep the child safe, do not try to wake them, and mention it to a doctor only if they happen most nights or the child is sleep-deprived. Nightmares that recur weekly for months, or that follow a frightening event, are worth raising with a doctor, as is any child who acts out dreams physically, since that is not normal at any age.
Something to Do Together
Start a dream diary on the breakfast table: one line each, first thing, before the day crowds it out, with a rule that nothing in it is silly. After two weeks read it back together and look for the everyday material, the classroom, the dog, the argument, showing up in disguise. Then watch the cat or the dog asleep and wait for the twitching paws and the flickering eyes; explain that the same switch that holds them still is holding your child still at night, and that the twitches are what leaks through.
Sources
- Aserinsky, E. and Kleitman, N. "Regularly occurring periods of eye motility, and concomitant phenomena, during sleep." Science 118 (1953): 273–274.
- Dement, W. and Kleitman, N. "The relation of eye movements during sleep to dream activity: an objective method for the study of dreaming." Journal of Experimental Psychology 53 (1957): 339–346.
- Foulkes, D. "Dream reports from different stages of sleep." Journal of Abnormal and Social Psychology 65 (1962): 14–25.
- Jouvet, M. "Recherches sur les structures nerveuses et les mécanismes responsables des différentes phases du sommeil physiologique." Archives Italiennes de Biologie 100 (1962): 125–206.
- Roffwarg, H.P., Muzio, J.N. and Dement, W.C. "Ontogenetic development of the human sleep-dream cycle." Science 152 (1966): 604–619.
- Hobson, J.A. and McCarley, R.W. "The brain as a dream state generator: an activation-synthesis hypothesis of the dream process." American Journal of Psychiatry 134 (1977): 1335–1348.
- LaBerge, S., Nagel, L.E., Dement, W.C. and Zarcone, V.P. "Lucid dreaming verified by volitional communication during REM sleep." Perceptual and Motor Skills 52 (1981): 727–732.
- Wilson, M.A. and McNaughton, B.L. "Reactivation of hippocampal ensemble memories during sleep." Science 265 (1994): 676–679.
- Maquet, P. et al. "Functional neuroanatomy of human rapid-eye-movement sleep and dreaming." Nature 383 (1996): 163–166.
- Revonsuo, A. "The reinterpretation of dreams: an evolutionary hypothesis of the function and meaning of dreaming." Behavioral and Brain Sciences 23 (2000): 877–901.
- Walker, M.P. and van der Helm, E. "Overnight therapy? The role of sleep in emotional brain processing." Psychological Bulletin 135 (2009): 731–748.
- Saunders, D.T., Roe, C.A., Smith, G. and Clegg, H. "Lucid dreaming incidence: a quality effects meta-analysis of 50 years of research." Consciousness and Cognition 43 (2016): 197–215.
- Medeiros, S.L.S. et al. "Cyclic alternation of quiet and active sleep states in the octopus." iScience 24 (2021): 102223.
- Foulkes, D. Children's Dreaming and the Development of Consciousness. Harvard University Press, 1999.