The phrase "mental health" appears on the marketing pages of almost every consumer wearable now. Apple has a mental wellbeing feature in the Watch. WHOOP quantifies stress. Garmin's Body Battery is, implicitly, a stress and recovery score. Oura reports on your capacity to handle the day. The category has clearly decided that the mind is its next growth territory.
We think this is mostly good, and we think it needs to be treated with more care than it has been. Wearable sensors can observe things that correlate with mental health. They cannot observe mental health itself. That difference is the whole story, and the more honest we are about it, the better wearables will end up being for the people who use them.
What sensors actually measure
Start with what is physically true. A wearable on your body can read your heart rate, your heart rate variability, your respiratory rate, your skin temperature, your skin conductance in some devices, your movement and posture, and your sleep architecture inferred from these signals. That is most of the sensing layer. It is genuinely impressive, and it is also a bounded set.
These signals correlate, in population averages, with states we colloquially call stress, recovery, and emotional load. When you are under sustained acute stress, your HRV tends to drop. When you are physically ill or emotionally exhausted, your resting heart rate tends to rise. When you are anxious, your respiratory rate tends to tick up. These correlations are real and medically meaningful.
But a correlation in a population is not a diagnosis in an individual. Your HRV being low today is compatible with grief, flu, sleep deprivation, a hard workout yesterday, a hangover, a stressful meeting, or a perfectly ordinary day that happens to sit in the low end of your normal range. The sensor can flag the signal. It cannot tell you what the signal means.
What the current products do well
The best mental health features in the current generation of wearables are the ones that stay on the safer side of this line. Apple's mindfulness prompts, for example, are not diagnostic: they are behavioural nudges, cued by context, and they do not claim to know how you feel. That is honest and useful.
Garmin's Body Battery is similar. It combines heart rate, HRV, activity, and sleep into a single readiness-ish score. It never tells you that you are depressed or anxious; it tells you that your body has more or less fuel in the tank today. That is a reasonable framing.
WHOOP's stress tracking is more ambitious and more prone to overreach. The company is careful in its clinical communication, but the product still presents continuous "stress" readings in a way that can read as mental-state detection, which it is not.
What the category is getting wrong
The main failure mode is confusing passive sensor data with psychological state. A readiness score is useful information about your body. It is not information about how you feel. When users start reading their readiness score before they read themselves, the wearable has started to replace self-awareness rather than enhance it.
There is a well-documented phenomenon in sleep tracking called orthosomnia, where people who focus heavily on sleep scores report worse subjective sleep even when their objective sleep is unchanged. The same pattern is possible for stress. The person who checks their HRV every hour and panics when it drops is, ironically, in a worse mental state because of the tracker than they would be without it.
The honest framing is this: wearables are good at giving you data that, for most people, adds up to a clearer picture of their body over time. They are not therapists, they are not diagnosticians, and they should not be sold as either.
Where wearables can really help
This is not a pessimistic piece. There are specific mental health use cases where wearables are genuinely valuable, and the industry should lean into them.
First, burnout pattern detection. Sustained drops in HRV combined with sustained rises in resting heart rate and disrupted sleep architecture are a real early warning sign of burnout, and most people do not notice until they are too deep in. A wearable that flags the pattern before the user feels it is doing useful work.
Second, illness onset. Many mental health dips are downstream of physical illness that has not yet become subjective. Skin temperature rising, respiratory rate ticking up, sleep quality falling, all a day or two before you realise you have a virus. Catching this early is materially useful.
Third, behavioural nudges. Mindfulness prompts, breathing exercises, movement reminders. These are not diagnostic. They are ambient interventions, and they can shift behaviour in ways that cumulatively matter.
Fourth, context for therapy. For users already in therapy or medical care, objective data about sleep, activity, and physical stress is useful ground truth. Clinicians increasingly appreciate this, and the better consumer platforms are starting to make it easy to share.
How to use a wearable for mental health without being used by it
Our honest guidance is this. Look at the mental health features on a weekly rather than daily cadence. Trust trend lines, not single readings. Do not diagnose yourself from a readiness score. If a tracker is making you more anxious rather than more self-aware, take it off for a week and see what changes. Use the behavioural features; be sceptical of the label features.
Most of all, remember that the sensor is watching your body, not your mind. The two are linked. They are not the same thing. A wearable that remembers this is a useful companion. A wearable that forgets it becomes part of the problem it was supposed to help with.