What Your Smartwatch Can (and Can't) Actually Tell Your ADHD Brain

You wake up, and before your feet hit the floor you've checked the number. Six hours, forty-one minutes. Fourteen per cent deep sleep. A little coloured bar telling you the night was "fair." And just like that, the day has a verdict on it before it's begun.

We say this as people who have worn the watch, chased the ring score, and felt a genuinely decent morning curdle because a device decided otherwise — not as clinicians. Consumer wearables are quietly brilliant at one job and quietly terrible at two others, and if you have ADHD, knowing which is which is the difference between a useful tool and a shiny new thing to feel bad about. Here's what the science actually says your smartwatch can and can't tell you, and how to use it without handing it the keys to your morning.

Can I trust my sleep tracker?

Partly — and the honest answer is more useful than a yes or a no. Your tracker is genuinely good at knowing whether you are asleep, reasonably good at spotting broad night-to-night patterns, and genuinely bad at two things it displays most confidently: the exact minutes it gives you, and the pretty light/deep/REM breakdown. In a head-to-head test of seven consumer sleep devices against in-lab polysomnography — the gold-standard "wired-up-in-a-lab" sleep study — the devices detected sleep with high sensitivity (all at or above 0.93), meaning that when you were genuinely asleep, they usually knew it (Chinoy et al., 2021).

So the instrument in the "is this person asleep or awake right now" job is doing well. The trouble starts the moment you ask it for anything more precise than that. Trusting the shape of your sleep over weeks is reasonable. Trusting any single night's exact figures is where people — especially ADHD people, who tend toward all-or-nothing readings of a single data point — come unstuck.

What does a smartwatch actually measure well?

The presence of sleep, and coarse trends over time. Where wearables fall down hard is telling sleep from wake once you're in bed but restless. In that same seven-device comparison, the devices' specificity for detecting wake was only low-to-medium (0.18 to 0.54) — meaning when you were actually lying there awake, the watch frequently logged it as sleep (Chinoy et al., 2021). A systematic review and meta-analysis of Fitbit models against polysomnography found the same lopsided pattern: sensitivity of 0.95 to 0.96 for correctly identifying sleep, but specificity of only 0.58 to 0.69 for correctly identifying wake, with the authors concluding these devices are "convenient and economical" but "not a substitute for" a proper sleep study (Haghayegh et al., 2019).

Read practically, that lopsidedness explains a common ADHD frustration: you lay awake for what felt like an hour, and the watch insists you slept. It isn't gaslighting you. It's simply built to err toward "asleep," because still bodies look like sleeping bodies to a wrist sensor. This is also why the device is a decent trend instrument and a poor stopwatch. The direction it points — better week, worse week — carries more signal than any single figure it hands you.

Why are the absolute sleep numbers wrong?

Because they're systematically biased, not randomly noisy — and "systematic" is the important word. A 2025 meta-analysis of 24 studies and 798 participants, pooling Fitbit, Apple Watch, Garmin and other consumer wrist devices against polysomnography, found that these devices underestimate total sleep time by roughly 17 minutes on average (mean difference −16.854 minutes; 95% CI −26.332 to −7.375), and tend to overstate how long you spent awake after first falling asleep (Lee et al., 2025).

A consistent lean in one direction matters more than it sounds. It means your device isn't wrong by a random wobble that averages out — it's quietly shaving minutes off, night after night, in the same direction. So the specific number that ruins your morning is not a measurement of your night; it's a measurement plus a known bias. This holds whether the gadget lives on your wrist (Fitbit, Apple Watch, Garmin) or your finger (Oura and other rings) — same principle, same reason to read the drift over a fortnight rather than the digits from last night. If you want one habit to take from this whole piece: never let a single night's headline number pass as fact.

How accurate are the sleep stages — light, deep and REM?

This is the weakest link, and it's worth being blunt about it. The colourful hypnogram — that stacked bar promising you precisely how much "deep" and "REM" sleep you got — is the least reliable thing on the screen. In the seven-device comparison, sleep-stage assessments were described plainly as "inconsistent" (Chinoy et al., 2021). Consumer devices infer stages mostly from movement and heart-rate patterns, which are a genuinely clever proxy and still nowhere near what electrodes on your scalp can see.

So when your watch says you got 42 minutes of deep sleep, the accurate translation is: the device's best guess, from an approach known to be unreliable at exactly this task, is around 42 minutes. That's not nothing, but it is not a fact to build a mood on. If you take one number off your device with a large pinch of salt, make it the stages. The trend in your total sleep is worth watching; the nightly deep-sleep percentage is closer to a mood ring than a measurement.

What is orthosomnia, and why should ADHD brains watch for it?

Orthosomnia is the name sleep clinicians gave to an unhappy loop: becoming so fixated on achieving "perfect" tracker sleep that the pursuit itself makes your sleep worse. The term was coined in a 2017 case series describing "a growing number of patients who are seeking treatment for self-diagnosed sleep disturbances such as insufficient sleep duration and insomnia due to periods of light or restless sleep observed on their sleep tracker data" (Baron et al., 2017). The authors flagged something we find quietly alarming: tracker data can start to feel more real to people than validated methods — more real than how rested they actually feel.

To be precise, because precision is the whole point of this article: no one has measured how common orthosomnia is. It's described qualitatively — "a growing number" — with no prevalence figure attached, and we won't invent one. What we can say is why it's worth naming for an ADHD audience specifically. ADHD brains are prone to hyperfixation and to all-or-nothing thinking, and a device that issues a nightly grade is a near-perfect machine for feeding both. When last night's "62" becomes tomorrow's anxiety, the tool has stopped serving you. The score was never the goal. Feeling functional was.

So how do I use my wearable without spiralling?

Treat it as a weather forecast, not a report card — and change what you look at. Practically, that means three shifts. First, read the trend line, never the nightly headline: is this a better week or a worse one? The device is honest at that resolution and biased at the single-night one. Second, let your body's report override the gadget's. If you feel rested and the watch says "poor," the watch is the thing that's wrong; the roughly 17-minute underestimate and the shaky staging are documented, and your felt experience isn't (Lee et al., 2025). Third, cover the score if it hooks you. If the first number you see rewrites your morning, that's a design problem, and the fix can be as simple as not looking until the evening.

This is also, frankly, a design stance we've taken a hard line on. A lot of sleep tech is engineered to maximise the very fixation Baron and colleagues warned about — nightly scores, streaks to protect, a red badge when you "fail." We think that's a dark pattern wearing a wellness costume, and for an ADHD user it's actively harmful. Momentum shouldn't be manufactured by making you anxious about a number that a peer-reviewed meta-analysis says is biased in the first place.

Can my wearable read my energy or my mood?

Not reliably yet — and anyone claiming otherwise is ahead of the evidence, including us if we ever did. Inferring an inner state (energy, mood, affect) from signals like voice or movement is a genuinely promising research area and not a validated clinical one. A PRISMA systematic review of 127 studies on using speech to assess psychiatric state concluded that the technology "could aid mental health assessments, but there are many obstacles to overcome, especially the need for comprehensive transdiagnostic and longitudinal studies" (Low et al., 2020).

We mention this because Ecstasis does experiment with inferring your energy from how you speak, and we'd rather tell you plainly where that sits: it's an investigational signal we offer as a prompt for your own reflection, not a verdict about your brain. The same discipline that says "don't trust the deep-sleep bar" says "don't trust an app that claims to know your energy." Both are estimates. Both should defer to how you actually feel. An honest tool tells you when it's guessing.

How Ecstasis helps

Ecstasis is built on exactly the posture this article argues for: trends over verdicts, your felt experience over the device's confidence. There are no sleep scores, no streaks, and no red "you failed" badges — deliberately, because the evidence on orthosomnia and the reality of ADHD hyperfixation both point the same way. Where it uses wearable or voice data, it shows you drift and pattern, labelled as estimates, and it's transparent about what it's inferring and what it's merely guessing.

The iOS app is in public TestFlight beta right now — you can try it today. If you'd rather wait for the wider release, join the waitlist at ecstasis.app and we'll let you know when it's ready. The whole idea is a quieter relationship with your own data: enough signal to act on, none of the theatre that turns a tool into a stick to beat yourself with.

This is education, not medical advice

Everything above is general information to help you read your own devices more sceptically — not a diagnosis, and not a treatment plan. If you're genuinely worried about your sleep, persistent daytime tiredness, or your ability to function, that deserves a proper assessment rather than a consumer gadget's best guess. Talk to your GP or prescriber. A wearable can help you arrive at that appointment with useful patterns to describe; it can't replace the clinician who interprets them.

References

• Baron KG, Abbott S, Jao N, Manalo N, Mullen R. Orthosomnia: Are Some Patients Taking the Quantified Self Too Far? Journal of Clinical Sleep Medicine. 2017;13(2):351–354. DOI: 10.5664/jcsm.6472.
• Chinoy ED, Cuellar JA, Huwa KE, Jameson JT, Watson CH, Bessman SC, Hirsch DA, Cooper AD, Drummond SPA, Markwald RR. Performance of seven consumer sleep-tracking devices compared with polysomnography. Sleep. 2021;44(5):zsaa291. DOI: 10.1093/sleep/zsaa291.
• Haghayegh S, Khoshnevis S, Smolensky MH, Diller KR, Castriotta RJ. Accuracy of Wristband Fitbit Models in Assessing Sleep: Systematic Review and Meta-Analysis. Journal of Medical Internet Research. 2019;21(11). PMID: 31778122.
• Lee YJ, Lee JY, Cho JH, Kang YJ, Choi JH. Performance of consumer wrist-worn sleep tracking devices compared to polysomnography: a meta-analysis. Journal of Clinical Sleep Medicine. 2025;21(3):573–582. DOI: 10.5664/jcsm.11460.
• Low DM, Bentley KH, Ghosh SS. Automated assessment of psychiatric disorders using speech: A systematic review. Laryngoscope Investigative Otolaryngology. 2020;5(1):96–116. DOI: 10.1002/lio2.354.