Skin temperature vs core temperature: what a wrist reads

Your watch shows a wrist temperature each morning, and the number invites an obvious reading: this is how warm my body was last night. It is a reasonable assumption. It is also the wrong one. The sensor held against your skin measures a different quantity than a thermometer under your tongue would — and on the nights you most want explained, the two can move in opposite directions.

Two numbers, two different questions

Core temperature is the one your body defends. It drifts through a modest daily rhythm, lowest in the small hours, and the whole apparatus of sweating and shivering exists to keep it inside that narrow band. Ask what your core temperature is, and you are asking about a regulated set point.

Skin temperature is not defended. It is the temperature of the surface where heat leaves, so it moves with whatever affects that exchange: how open your blood vessels are, how warm the room is, what you are sleeping under, whether your watch arm spent the night above or below the covers. It can swing by several degrees in minutes while core barely moves.

So the gap between the two numbers is not measurement error to be corrected. The gap is the physiology — it is a readout of how much heat your body is currently dumping.

Your wrist can warm while your core is cooling

This is where the intuitive reading breaks down completely.

Falling asleep requires losing heat, and the body loses it through the hands and feet: the vessels there open, the extremities warm, and core temperature drops. In a 1999 study in Nature, researchers found that the distal-to-proximal gradient — how much warmer the hands and feet were than the trunk — was the best physiological predictor of how quickly sleep arrived. Better than sleepiness ratings. Better than melatonin.

Read that again with a watch on. Warmer skin at the periphery is not the signature of a body overheating. It is often the signature of a body successfully cooling itself down into sleep. The wrist sits partway between the extremities and the trunk, so it tells a quieter version of the same story — but the direction still holds, and it is the opposite of the one the number seems to suggest.

The event you want to see is smaller than the instrument

The other half of the assumption is that a hot flash must be a temperature spike, and a temperature sensor should therefore catch it.

Research from the late 1990s described the mechanism: in symptomatic women, the thermoneutral zone — the range of core temperature between the sweating threshold above and the shivering threshold below — is dramatically narrowed, so very small changes are enough to trigger heat-loss responses. That is well established, and it makes flashes sound like a core temperature story.

Then a 2019 analysis pooled 108 episodes objectively recorded in the laboratory in 72 symptomatic postmenopausal women across five prior studies, and measured the trigger directly. In the 20 minutes before a spontaneous episode, core temperature rose by 0.03 ± 0.12°C — and only 51% were preceded by any rise at all. Under mild heating, 76% were. The authors concluded that these episodes are unlikely to be triggered by a rise in core temperature.

Two things follow, and neither is in the usual explainer. First, even a perfect ingested core thermometer would fail to register roughly half of them as temperature events. Second, three hundredths of a degree is orders of magnitude below anything a wrist sensor summarizing an entire night can resolve. What a watch has a better chance of registering is downstream — the flush of vasodilation, the sweat, the awakening that follows — which is why an overnight temperature line and a night sweat are related but genuinely different records, and why distinguishing the two kinds of nighttime episode relies on timing and duration rather than on degrees.

Why the watch reports a change instead of a temperature

Apple’s design quietly concedes all of this. The watch samples every five seconds through the night, needs roughly five nights of sleep tracking before it will show you anything, and then reports a deviation from your baseline rather than a temperature. The documentation states plainly that the feature is not a medical device and is not intended for diagnosis.

That is the honest engineering answer to a hard problem. There is no useful population-normal wrist temperature to publish, because so much of the number belongs to your room, your bedding, your sleeping position and your own vasomotion. This is the site’s recurring point in its most literal form: the only fair comparison is you against your own recent nights, because a between-person comparison here is mostly a comparison of bedrooms.

It is also why the cycle research is built the way it is. Studies of wrist skin temperature against basal body temperature, and of overnight finger temperature in ring wearers, look for a multi-night shift within one person’s own record. That approach survives everything above. An absolute reading does not — and when cycles turn irregular, the shift itself often stops appearing.

How Perigee reads a wrist temperature night

We never label the overnight reading as your body temperature, because it is not one. We hold it against your own recent nights, and we show it beside what else moved — when sleep broke, what your heart rate did around those breaks — because a surface temperature is most informative in company.

Everything an Apple Watch records overnight is an estimate, not a laboratory measurement, and the wrist temperature signal page sets out what the sensor can and cannot support. That framing is the entire premise of what we built, and the evidence behind it is deliberately kept where you can check it.

One small thing to try this week

For seven nights, write down the three things the sensor cannot separate from your body: roughly how warm the room was, what you slept under, and whether your watch arm was inside or outside the covers.

Then read the week with those notes beside the readings. Most of the variation will explain itself, which is useful in itself — it tells you how much of that line was ever about you. The nights worth carrying to a clinician are the ones that still stand out after the room, the duvet and the arm are accounted for.