Psychomotor
Vigilance Task

10 min · lapse ≥ 500 ms

The PVT is the standard laboratory measure of behavioural alertness. It is deliberately the simplest task in this collection: nothing to learn, nothing to remember, no strategy that helps. That is what makes it work — with aptitude and practice taken out of the picture, what moves your numbers between sessions is state, not skill. What it detects is not slowness but wake-state instability: brief attentional failures intruding into otherwise normal responding.

The task

284
An empty box sits in the middle of the screen. At an unpredictable moment a millisecond counter starts running inside it. Press SPACE the instant you see it — the counter stops and shows your reaction time for a second, then the wait begins again.

Do not press before the counter appears. Early presses are recorded as false starts and restart the wait with a fresh random interval, so pressing to a rhythm gets you nothing but a worse score. On a phone, tap anywhere instead of SPACE.

What comes out

Two numbers carry the session, and neither is your average reaction time.

LapsesResponses slower than 500 ms — the count of moments your attention simply dropped out. This is the measure the PVT was built around.
Mean 1/RTResponse speed: every reaction time flipped to its reciprocal, then averaged. It registers the slowing that precedes a lapse, which a threshold count misses, without letting a handful of two-second responses dominate the average, which a raw mean cannot avoid.

Mean RT is reported too, and pointedly not scored: an average over a distribution with a long lapse tail moves mostly with that tail, so it is noisiest exactly where the signal lives — and it cannot tell uniformly mediocre responding apart from fast responding punctuated by failures. Those are different states.

Session length10 minutes
Interval between stimuli210 s, uniform
Lapse threshold500 ms
Expected stimuli~94

These are fixed, not preferences. The parameterisation is the instrument — change the interval range or the threshold and the numbers stop being comparable with anyone else's, or with your own.

Silence your notifications. The task counts a stimulus you missed because you looked away as a lapse, because it cannot tell the difference.
The other version

Ten minutes is a lot to ask of yourself repeatedly. The 3-minute PVT-B is validated against this test and loses about 23% of its sensitivity — worth taking if the alternative is skipping the measurement.

Go to PVT-B3 min, lapse ≥ 355 ms →

They are scored as separate instruments and keep separate histories. Runs on one are not comparable with runs on the other, so pick a version you can stick with rather than alternating.

This is a research-style measurement instrument and an educational demonstration — not a medical or diagnostic device. A browser adds display and input latency that a laboratory PVT device does not, so absolute values here are not comparable with published norms; what is usable is the change in your own numbers across sessions on the same hardware. Nothing here can confirm or rule out a condition, diagnose sleep disorders, or tell you whether you are fit to drive.

Alertness · Sustained attention

Psychomotor Vigilance Task

What this test measures

The Psychomotor Vigilance Task (Dinges & Powell, 1985) is the reference measure of behavioural alertness, used in sleep research from hospital wards to the International Space Station. An empty box sits on screen; after a random wait a millisecond counter starts running, and you stop it as fast as you can — for ten minutes.

The task is trivial on purpose. With nothing to learn and no strategy, performance is nearly immune to practice, so what it measures is state: how sleep pressure and fatigue destabilise attention. Sleep loss does not slow you uniformly — it makes brief attentional failures intrude into otherwise normal responding. That is why the PVT is scored on lapses (responses of 500 ms or more) and mean response speed (1/RT), the metrics most sensitive to sleep loss, rather than average reaction time.

How to use it

Take it under consistent conditions and the trend tracks your alertness — after poor sleep, across the workday, or around caffeine. Ten minutes is long by design: sustaining vigilance is part of what is measured, and time-on-task decline is reported too. If you will not reliably sit through ten minutes, use the validated 3-minute PVT-B instead — a shorter test you actually take beats a longer one you skip.

Sapiens Bench tasks are educational instruments for self-tracking, not medical or diagnostic devices.