Field note
Your watch draws clean blocks of light, deep and REM sleep. Those blocks are an estimate. They are not a lab sleep study, and treating them like one is the most common way to overread a consumer wearable.
Lab sleep staging is polysomnography. It scores stages from brain waves, eye movements and muscle tone, usually overnight in a clinic. A wrist device mostly sees motion and heart-rate patterns through the skin. That is a different instrument answering a related question, not the same measurement with a prettier chart.
If you only remember one sentence from this note, make it this: the stages are an estimate; the duration is not in the same category of uncertainty. Everything below exists to make that sentence usable on a Tuesday morning when the chart looks dramatic and you feel fine, or quiet and you feel wrecked.
When researchers want to know how close a consumer tracker gets to lab staging, they put people in polysomnography and wear the device at the same time. Three papers in that line sit behind Aera's own Field Guide material on this topic.
Chinoy and colleagues compared seven consumer sleep-tracking devices with polysomnography. Lee and colleagues ran a prospective multicenter validation across eleven wearable, nearable and airable consumer sleep trackers. Schyvens and colleagues validated six commercial wrist-worn devices specifically for sleep stage scoring against polysomnography.
I am citing those studies as they are named in Aera's source list. I have not read the full text of each paper for this note, so I am not attaching a specific agreement statistic, sensitivity figure, brand ranking or "X percent accurate" claim to any of them. If a later note needs a number from those papers, it will quote that number only after the text is opened, or it will not quote a number at all.
What those studies are doing, as a class, is measuring disagreement between instruments. Stage labels from the wrist and stage labels from the lab often fail to line up night by night. That disagreement is the honest limit, not a footnote. Duration and timing against your own recent nights are usually the sturdier signals from the same devices.
A consumer stage chart is a model. It maps wrist signals onto labels that look like lab categories. The chart can still be useful as a soft hint: some nights look more fragmented, some look more continuous. It becomes harmful when you chase "more deep sleep" as if the minutes were a clinical score you can optimise like a race time.
Deep sleep and REM matter in sleep science. That does not mean your watch has measured them the way a sleep lab would. Carskadon and Dement's overview of normal human sleep is the textbook framing for what lab stages are supposed to mean. A wrist chart borrows those labels. Borrowing a label is not the same as repeating the measurement.
If an app never says the stage labels are noisy, you are looking at confidence theatre. A precise-looking ribbon is easy to design. An honest ribbon is harder to sell.
Prefer total sleep time and consistency against your own baseline. Ask whether last night was shorter or longer than your usual, and whether that pattern has held for a week. That is a question the sensor is better built to answer than "how many minutes of deep sleep did I bank." That duration question is the one taken up in How much sleep do I actually need?
Use stages as a conversation with how you felt. If you feel wrecked and the chart looks pristine, believe your body first. If you feel fine and the chart looks broken, do not invent a crisis from a colour block. The point of a personal baseline is exactly that: your usual, not a stranger's stage recipe.
When you compare apps, remember they may disagree with each other as well as with the lab. Different models, different update windows, different ways of filling gaps when the watch was off the wrist. Disagreement between two watches is not proof that one of them secretly knows the truth.
Also remember what the watch cannot see. It does not score your night from brain waves. Alcohol, illness, a late meal, a bad mattress and a stressful week can all change how you feel without producing a stage chart that looks like a thriller. The absence of a dramatic chart is not proof that nothing happened.
The stages are an estimate. The duration is not in the same category of uncertainty. That pair of claims is the whole note. Everything else is context so you do not throw the watch away, and so you do not treat its stage ribbon like a prescription.
The useful move, for anyone with a watch, is to stop asking the stage chart for a resolution it does not have, and to start asking the questions it can answer well. When a morning grade folds noisy stage minutes into one number, read What does a sleep score mean? When short nights start stacking across the week, What is sleep debt? is the paired note.
I write these because I build Aera, an iPhone app that reads Apple Health against your own baseline and puts the research and the error bars underneath every number it shows you, including this one. You do not need it to use anything above.