Why Humans Sleep—and What Scientists Are Still Discovering

Sleep is strange when you look at it plainly. For hours, an animal that needs to find food, avoid danger and care for its young becomes nearly unresponsive. Evolution tends to trim costly habits that do not pay off, yet sleep has persisted. That suggests it does something important. What, exactly, is a question researchers have not settled.

This article separates three things: what we can describe with confidence, what is well supported but not fully explained, and what is still being argued about. It is general education, not medical advice. If you have persistent sleep problems, a clinician is the right person to ask.

What a night of sleep looks like

Sleep is not a single state. The U.S. National Institute of Neurological Disorders and Stroke (NINDS) describes two broad types: non-REM sleep, which has three stages, and REM (rapid eye movement) sleep. Stage 1 is the light drift away from wakefulness. Stage 2 is still light, but heart rate and breathing slow and body temperature drops. Stage 3 is deep sleep, the stage from which a person is hardest to wake.

REM sleep typically first appears about 90 minutes after you fall asleep. Your eyes move, breathing becomes faster and less regular, and your limbs are temporarily unable to move. Most dreaming is reported from REM, which is why it is tempting to say that dreams happen only then. That is a simplification: the NINDS summary says most dreaming occurs in REM, not all of it.

The clock that tells you when

Why do you get drowsy in the evening even if you have done nothing tiring? Part of the answer is an internal timekeeper. The body runs a roughly 24-hour cycle, called the circadian rhythm, that influences wakefulness, body temperature, metabolism and hormone release. A small region of the hypothalamus, the suprachiasmatic nucleus, receives light information from the eyes and helps keep that clock aligned with the outside day. The pineal gland releases melatonin as darkness arrives, which helps promote sleepiness.

Light is the strongest influence on this cycle, according to NINDS. That is one reason jet lag and night-shift work are so disruptive: the clock and the schedule point in different directions. It is also why the common advice about bright mornings and dimmer evenings has a reasonable basis, even though individual people vary a good deal.

How much sleep do people need?

There is no single number for everyone. The U.S. Centers for Disease Control and Prevention (CDC) lists recommended ranges that shift with age: for example, 14 to 17 hours for newborns, 9 to 12 hours for children aged 6 to 12, 8 to 10 hours for teenagers, and 7 or more hours for adults aged 18 to 60. For people 65 and older, the CDC range is 7 to 8 hours. NINDS gives a slightly different wording for adults, 7 to 9 hours.

These are population guidelines, not personal prescriptions. Some people feel well at the low end of a range and others at the high end, and needs can change with illness, pregnancy, age and life circumstances.

So what is sleep for?

Here the story gets more honest and more interesting. NINDS states plainly that the biological purpose of sleep is still a mystery, and that the purpose of dreaming is not known. Several ideas compete, and they are not mutually exclusive.

• Memory and learning. Sleep seems to help the brain stabilize and organize what it has taken in. This is supported by many studies, though the precise mechanisms are debated.
• Housekeeping. Some researchers propose that sleep helps clear waste from the brain.
• Energy and body regulation. Sleep touches nearly every body system, from metabolism to immune function.

The housekeeping idea became famous after a 2013 mouse study from the lab of Maiken Nedergaard, which proposed a fluid-flow system now called the glymphatic system. It has been widely discussed, including in popular articles that describe sleep as the brain taking out the trash. But it is a live dispute, not settled fact.

A case study in an unsettled question

In May 2024, a team led by Nick Franks at Imperial College London published a study in Nature Neuroscience that injected a fluorescent dye directly into mouse brain tissue and tracked how fast it cleared. They reported that clearance was about 30 percent lower during sleep and about 50 percent lower under anesthesia than during wakefulness, the opposite of what a simple version of the glymphatic idea predicts. They also found that dye movement looked driven by diffusion rather than by flow.

Critics responded quickly. Nedergaard argued that the experiment lacked key controls, that the injection itself could cause problems, and that the sleep model was not natural sleep. Other researchers not involved in the work said some technical worries were valid while finding the modeling persuasive. The dispute was still unresolved when reported, and a formal critique followed in 2025. Both camps agree that better tools are needed.

Sorting a few common claims

Why the open questions are worth sitting with

It can feel unsatisfying that a universal human activity lacks a settled explanation. But this is how science tends to look near its frontier: robust descriptions, partial mechanisms and a few good arguments. The 2024 disagreement is a useful example, because both sides are doing what researchers should do, which is to test an attractive idea and argue about the method.

For everyday life, you do not need to wait for a final theory. The reasonably solid parts, such as regular timing, a dark enough room, and treating persistent trouble as something to discuss with a professional, are available now. If sleep and attention interest you, you might also enjoy why your memory is a reconstruction, since the two topics overlap in research on learning.