1. How Rapid Eye Movements Became Linked to Dreams
In 1953, two researchers at the University of Chicago, Eugene Aserinsky and Nathaniel Kleitman, published a short paper about periods during sleep when the eyes move in bursts[1]. Then they did something that sounds rude but was clever. They woke volunteers up at different moments and asked what had been going through their minds.
When the volunteers were woken during eye movement, 20 out of 27 times they described a detailed dream, usually full of pictures. The other 7 times they either could not remember anything, or they felt they had been dreaming but could not recall any details[1]. That is about three out of four.
When the eyes were still, the answers were very different. Out of 23 wake-ups, 19 brought no memory of a dream at all. The remaining 4 were split evenly between "felt like dreaming, no details" and a detailed dream[1].
This was not "everyone dreams whenever their eyes move." It was a pattern, counted one wake-up at a time, and it was strong enough to get many scientists interested. The stage of sleep with bursts of eye movement came to be called REM sleep, for "rapid eye movement."
2. REM and Non-REM Sleep Repeat Through the Night
REM sleep is not one long block. The U.S. National Heart, Lung, and Blood Institute explains that sleep alternates between REM and non-REM phases, and that the cycle starts over every 80 to 100 minutes, with usually four to six cycles a night[2].
During REM, the eyes twitch and the brain is active, about as active as when you are awake. The muscles, meanwhile, normally go limp, which keeps you from acting out your dreams[2]. The same page adds that you usually get more REM sleep later in the night[2].
So REM sleep, and the dreams that go with it, are spread across the whole night, with more REM toward the morning. That is why the timing of a wake-up matters when scientists ask people about their dreams.
3. People Dream During Non-REM Sleep Too
After 1953, it was tempting to conclude that dreams happen in REM sleep and nowhere else. More wake-up studies said otherwise. A 2000 review by the sleep researcher Tore Nielsen concluded that mental activity in non-REM sleep is real. Averaged across earlier studies, people reported something in 81.9 percent of REM wake-ups and 43.0 percent of non-REM wake-ups, with a lot of spread from study to study. What gets counted is "mental activity," a wider category than "a dream"[3]. A later paper puts the REM figure for at least one dream at about 80 percent[4]. REM gets more dream reports, but non-REM sleep is far from empty.
The details are worth a look. In one study, the neuroscientist Francesca Siclari and colleagues woke sleepers 735 times during the deeper non-REM stages (N2 and N3) and asked what they had experienced. About one third of the time (33.5 percent), people described a dream. Another 38.6 percent said they had been dreaming but could not remember what it was about. Only 27.8 percent said they had experienced nothing[5].
Put the last two numbers together and you get a surprising idea: "I didn't dream last night" and "I don't remember my dream" can look exactly the same in the morning. A blank does not prove there was no dream.
The same team also looked at where in the brain this was happening. Before dream reports, they found less slow-wave activity, a slow, rolling kind of brain activity, in a region at the back of the brain they called a "hot zone." Their summary is that in both REM and non-REM sleep, dreaming goes with local drops in slow-wave activity in the back of the brain[5]. In other words, the clue may not be the kind of sleep you are in. It may be which part of your brain is switched on.
4. What Are Dreams For? Three Ideas, No Winner Yet
Here is the honest part. Scientists have ideas about why we dream, but no idea has won. Here are three that are taken seriously enough to be published in major journals.
Idea 1: Dreams Rehearse Dealing with Danger
In 2000, the Finnish researcher Antti Revonsuo argued that dreams are not random noise. He pointed out that dream content is organized and selective: some kinds of things are over-represented compared with waking life, and others are under-represented. He proposed that the biological function of dreaming is to simulate threatening events and rehearse noticing and avoiding them[6]. This is a hypothesis, not a settled fact, and it was published alongside a long set of comments from other researchers.
Idea 2: Dreams Build a Model of the World
In 2009, the psychiatrist J. Allan Hobson suggested that REM sleep might be a "protoconscious" state, which means a kind of early version of waking consciousness. On this view, the brain builds a virtual-reality model of the world that helps the brain develop and keep up its waking awareness[7]. This, too, is a theory.
Idea 3: Dreams Help the Brain Sort What You Learned
A third line of research asks whether dreams are linked to memory. In a 2010 study, Erin Wamsley and colleagues trained volunteers on a virtual navigation task, then let them nap. People who had dream images about the task during the nap tended to do better when retested than people who did not. Thinking about the task while awake did not predict better scores[8]. Be careful with this one. It shows that the two things go together. It does not prove that dreaming caused the improvement.
Three ideas, three different answers, and none of them knocks out the others. That is not a failure of science. It is what a real open question looks like.
How to Look into Dreams Yourself
- Keep a dream notebook for a week. Put a pencil and paper beside your bed. As soon as you wake up, before you get up, write one line: a picture, a sound, or a feeling, whichever you can catch. A single word such as "flying" or "scared" is fine. At the end of the week, count the mornings when you remembered nothing. The surprising part of this article predicts those blank mornings are not proof you did not dream.
- Read the original paper. The 1953 Aserinsky and Kleitman paper is only a few pages long, and a scan of it is free online from the University of California, Los Angeles. See how two researchers described "27" and "20" in their own words.