REM Sleep: What Happens and How Much You Need

Your brain becomes strikingly active, your eyes flick beneath closed lids, and most of your large muscles go quiet. REM is the strangest-looking part of a normal night—and the part most likely to be shortened by an early alarm.

Rapid eye movement sleep is the fourth sleep stage, reached after the three stages of non-REM sleep. It is strongly associated with vivid dreaming, but dreaming is only its most famous feature. REM also has a distinctive mix of waking-like brain activity, temporary muscle paralysis, irregular breathing and heart rate, and a growing presence as morning approaches.

For healthy adults, REM commonly occupies roughly 20% to 25% of total sleep. That works out to about 90 to 120 minutes in an eight-hour night, but there is no official REM quota that every person must hit. The useful target is enough total sleep for your age and circumstances, not a perfect stage percentage on a watch.

What Happens During REM Sleep?

REM is sometimes called “paradoxical sleep” because the brain and body seem to be telling opposite stories. Brain activity becomes faster and more mixed than it is during deep sleep, resembling wakefulness in several respects. Behind closed eyelids, the eyes move rapidly from side to side. At the same time, neural circuits in the brainstem strongly reduce tone in most skeletal muscles.

That muscle atonia is why an ordinary dream does not usually turn into a full-body performance. Small twitches can still occur, and the muscles that control breathing continue to work. Heart rate and breathing may become less regular than during non-REM sleep. Peever and Fuller’s review of REM biology describes this unusual state as a coordinated package rather than simply “the dreaming stage.”

Dreams can occur in every sleep stage. REM dreams are often longer, more vivid, and more emotionally charged, which helps explain the association, but REM and dreaming are not interchangeable. The separate guide to why we dream examines the main theories without treating one stage as a complete answer.

REM Builds Toward Morning

The first REM period commonly arrives about 90 minutes after sleep begins. It is usually brief. The night then cycles back through non-REM sleep, returns to REM, and repeats. Deep sleep is concentrated earlier, while REM periods generally become longer across later cycles. The exact timing varies; a night is not made from identical 90-minute blocks.

The proportions are illustrative, not a personal sleep-stage prescription.

This architecture gives an early alarm a hidden cost. Cutting the final hour from a night does not remove an average slice of every stage; it often removes a REM-rich part of the night. The broader guide to N1, N2, N3, and REM shows how the full sequence fits together, while the sleep-cycle guide explains why the familiar 90-minute rule is only an approximation.

How Much REM Sleep Do Adults Need?

There is no separate clinical recommendation saying that an adult must obtain a particular number of REM minutes. Sleep guidelines recommend total sleep because the stages organize themselves dynamically within it. The American Academy of Sleep Medicine and Sleep Research Society recommend that adults sleep seven or more hours per night on a regular basis. Individual need still varies above that minimum.

If REM makes up 20% to 25% of a night, the rough arithmetic looks like this:

Those are reference ranges, not pass marks. REM percentage changes with age and can vary from night to night. A large review found a gradual decline across adulthood rather than one sharp age cutoff. More importantly, a person can fall outside 20% to 25% on one night without anything being wrong.

What Is REM Sleep For?

The honest answer is that REM probably does several jobs, and science has not reduced them to one settled function. The sleep and memory guide follows the best-known candidate: how sleep stabilizes and reorganizes new learning, with non-REM and REM contributing in different ways. Rasch and Born’s major review of sleep and memory presents this as an interacting overnight process, not a contest in which one stage does all the useful work.

REM has also been linked with emotional-memory processing. Its brain chemistry and patterns of activity create unusual conditions for revisiting emotionally important material, but researchers still debate exactly what changes: the memory itself, its emotional charge, later reactivity, or some combination. Goldstein and Walker’s review lays out a prominent model, while later findings have kept the details open.

That nuance matters. “REM processes emotions” is a useful summary; “more REM automatically fixes mood or memory” is not. Deep sleep, lighter non-REM sleep, REM, and brief awakenings form one architecture. The comparison of REM sleep versus deep sleep separates their strongest associations without pretending either stage works alone.

For anyone learning a skill or using cognitive training exercises, the practical lesson is simple: protect the night that follows the practice. Training supplies the experience; sleep helps the brain decide what survives it.

Why Would a Tracker Show Low REM?

The most ordinary explanation is a short night, especially one cut off at the morning end. Fragmented sleep can also interrupt normal cycling. Age, alcohol, some medications, illness, sleep disorders, and plain night-to-night variation can change the pattern as well.

Then there is the measurement itself. A clinical sleep study identifies stages using brain waves, eye movements, muscle activity, and other signals. Most consumer devices infer them mainly from movement and cardiovascular patterns. In a laboratory comparison of seven devices, sleep-versus-wake detection was generally stronger than stage classification, while sleep-stage estimates varied across devices.

So a watch reporting 14% REM once is not the same as a sleep study showing a persistent abnormal pattern. Treat wearable REM as a trend estimate, not a nightly verdict. If the number stays unusual alongside persistent daytime sleepiness, unrefreshing sleep, frequent awakenings, breathing concerns, or physically acting out dreams, the symptoms matter more than the app score.

Can You Increase REM Sleep?

You cannot order the brain to enter REM for another 20 minutes. The most defensible strategy is less glamorous: allow enough time for a complete night and stop repeatedly trimming the REM-rich morning portion.

Start with your actual schedule. For seven nights, record when you got into bed, roughly when you fell asleep, and when you woke. If the window regularly leaves less sleep than you intend, fix the window before trying to optimize a stage percentage. Our calculator builds in sleep-onset time and expected awakenings rather than pretending that eight hours in bed always means eight hours asleep.

Regular timing can make that sleep opportunity easier to use. Reducing avoidable disruption helps the stages unfold without guaranteeing any exact percentage. The goal is not to “hack REM”; it is to stop crowding it out.

The Better Question to Ask

Instead of asking, “Did I hit 25% REM last night?” ask, “Did I leave enough uninterrupted time for the later cycles to happen?” That question turns an uncertain wearable number into something you can act on.

Use the total night as your baseline, protect the morning end for a week, and then judge the pattern by how consistently you slept and functioned—not by whether every bar in an app turned the right color. The Sleep section continues from stage architecture into duration, schedules, sleep debt, dreams, and daytime performance. For a waking baseline, try the Short Term Memory Test or browse the Memory & Recall section; the broader brain tests collection adds speed, attention, and reasoning. None measures REM itself.