睡眠階段解析：深層睡眠 vs. REM 睡眠

Most healthy adults spend roughly 13 to 23 percent of total sleep time in deep sleep, which works out to about 1 to 1.5 hours across a 7 to 9 hour night. Deep sleep percentage tends to be highest in children and teenagers, gradually declines through adulthood, and drops further after age 60, so a healthy older adult getting 10 percent deep sleep is not necessarily a problem. What matters more than hitting an exact number is whether your own deep sleep is trending down over several weeks, since that is a more reliable signal of accumulating sleep debt or poor sleep quality than any single night.

Deep sleep, also called slow-wave sleep, is the stage where the body does most of its physical repair: growth hormone release peaks, muscle tissue rebuilds, and the immune system is reinforced, and it is hardest to wake someone from this stage. REM sleep is when the brain is nearly as active as when awake, dreaming is most vivid, and the brain consolidates memories and processes emotional information; heart rate and breathing become irregular and most skeletal muscles are temporarily paralyzed. Deep sleep concentrates in the first half of the night while REM sleep concentrates in the second half, so cutting a night short from either end costs you disproportionately more of one stage.

A single night of low deep sleep is usually explained by something identifiable: alcohol close to bedtime, a late or heavy meal, a shortened total sleep window, elevated stress or cortisol, illness, heat in the bedroom, or unusually high training load the day before, all of which suppress slow-wave sleep. Wearable stage estimates also carry night-to-night measurement noise, so a single low reading is not necessarily meaningful on its own. If deep sleep stays low across multiple consecutive nights rather than bouncing back, that pattern is worth paying attention to rather than one off night.

Neither stage matters more; they serve different jobs, so a REM-heavy night with very little deep sleep and a deep-sleep-heavy night with very little REM are both incomplete in different ways. Deep sleep drives physical recovery and immune function, while REM sleep drives memory consolidation, learning, and emotional regulation, and healthy sleep architecture needs adequate time in both across a full night. Someone training hard and feeling physically run down should look first at deep sleep trends, while someone struggling with focus, mood, or memory should look first at REM trends.

Consumer wearables like Oura, Apple Watch, Fitbit, and Whoop estimate sleep stages from heart rate variability, movement, and sometimes skin temperature, and they agree with clinical polysomnography (EEG-based lab sleep studies) on total sleep time and overall sleep versus wake reasonably well, but they are noticeably less precise at distinguishing light sleep from deep sleep on a night-by-night basis. That precision gap is why stage percentages are more useful as a multi-week trend than as an exact nightly measurement, and why a 5 to 10 percent swing between two similar nights usually reflects device estimation noise rather than a real change in sleep quality.