Your Reaction Time at 7 AM Is Worse Than You Think

It is 7:04 AM. You are three minutes into the day. You tilt the kettle, and the water goes past the rim of the mug and onto the back of your hand. You did see it happening. You watched it happen for what felt like a long time, and your arm did not get the message.

Twenty minutes later you would not have spilled it. Nothing about your character changed in those twenty minutes. Something about your speed did, and unlike most things people say about mornings, that something has a unit: milliseconds.

The test that turns grogginess into a number

Sleep researchers stopped arguing about how tired people feel a long time ago. They measure instead, and the instrument they reach for first is deliberately, almost insultingly simple.

The psychomotor vigilance task was described by David Dinges and John Powell in 1985 as a portable simple visual reaction time task for use during sustained operations (Dinges and Powell, Behavior Research Methods, Instruments, and Computers, 1985). You watch a blank box. At random intervals, usually somewhere between two and ten seconds apart, a counter starts running inside it. You press a button. The counter stops and shows you your time in milliseconds. Then you wait for the next one. For ten minutes.

There is no skill in it and no strategy to discover. That is the entire point. The task was built to survive repeated administration: it shows minimal practice effects and is relatively free of aptitude effects, which is why the same person can be tested over and over without the numbers drifting for reasons that have nothing to do with sleep (Basner, Mollicone and Dinges, Acta Astronautica, 2011).

The headline number is not your average, which is a surprisingly poor summary of what sleep loss does to a brain. The number that matters is the lapse: a trial where your response takes longer than 500 milliseconds. A lapse is not you being a bit slow. Half a second on a task this trivial means attention dropped out entirely and then came back. The stimulus was there, your eyes were open, and for a moment nobody was home.

Rested adults produce very few of them. That is the baseline the morning has to be compared against.

What the numbers actually say about 7 AM

Here is the measurement that gets closest to your kitchen.

Santhi and colleagues gave eleven participants a 6.5 hour sleep opportunity and then started testing within three minutes of lights-on, continuing for four hours. Mean reaction time was significantly impaired immediately on waking, and lapses rose to roughly two and a half times baseline, from about 3 to about 7.7 per test (Santhi et al., PLOS ONE, 2013). Speed measures took the damage; accuracy held up better. Sustained attention and alertness were hit hardest, cognitive throughput least.

Read that as a sentence about your morning: the part of you that reacts is impaired well before the part of you that thinks notices anything is wrong.

It gets less flattering. Wertz and colleagues compared performance straight out of a full night’s sleep against performance after extended sleep deprivation, and found that a simple addition test taken immediately on waking was more impaired than the same test after a night without sleep (Wertz et al., JAMA, 2006). The full mechanism behind that is in sleep inertia and how to wake up sharp.

And it does not clear on a switch. Jewett and colleagues tracked alertness and cognitive throughput across the four hours after habitual wake time and found the recovery followed a saturating exponential, with time constants of roughly 40 minutes for subjective alertness and roughly 70 minutes for cognitive performance, taking two to four hours to approach its asymptote (Jewett et al., Journal of Sleep Research, 1999). A review by Hilditch and McHill puts the practical version plainly: performance typically returns to pre-sleep levels within about 30 minutes, but full recovery may not be complete until at least an hour after waking (Nature and Science of Sleep, 2019).

Thirty minutes to be functional. An hour or more to be yourself.

“Awake but not sharp” is a real category, and you are bad at detecting it

The uncomfortable finding is not that morning reaction time is slow. It is that your internal report on it is unreliable.

In the chronic restriction arm of Van Dongen and colleagues’ study, participants held to 6 hours in bed for 14 nights showed cumulative, dose-dependent deficits on the vigilance task that kept getting worse day after day. Their subjective sleepiness ratings did not. Those responded once, early, and then flattened out (Van Dongen et al., Sleep, 2003). People got measurably worse and stopped noticing.

The morning version of this is the same failure over a shorter window. Grogginess arrives as a general fog, so you assume the impairment is a general fog too: everything slightly dimmed, uniformly. What the lapse data describes is different and worse. Most of your responses are close to normal, and then, unpredictably, one is not there at all. You cannot feel the gaps, because feeling them would require the attention that just went missing.

What half a second is worth

Milliseconds are hard to care about, so translate them.

At 100 km/h you cover about 28 metres every second. An extra 300 milliseconds before your foot moves is roughly 8 metres of road that you drove through while unavailable. That is not a study, it is arithmetic, but it is the kind of arithmetic that turns a number on a screen into a bumper.

For scale on the effects themselves: Dawson and Reid put sustained wakefulness and alcohol on the same axis and found that after 17 to 24 hours awake, psychomotor impairment reached levels equivalent to a blood alcohol concentration of roughly 0.05 to 0.10 percent (Dawson and Reid, Nature, 1997). Sleep inertia is a different physiological state from prolonged wakefulness, so that is not a like-for-like swap. The useful point is that impairments of this general size are the kind we write laws about, and we do not usually apply the same seriousness to the first twenty minutes of a Tuesday.

Five things that move the number

1. Put a 30 minute buffer in front of anything irreversible. Not important things, irreversible ones: the reply you cannot unsend, the merge, the transfer, the yes. The Jewett curve says the steepest recovery is in that first half hour, so spending it on laundry rather than judgement is close to free.
2. Protect the wake time before you optimise anything else. How deep the hole goes depends heavily on what stage you were dragged out of and how much sleep debt you carried in. A stable schedule is a cheaper intervention than anything you can buy.
3. Stop re-entering sleep after the first alarm. Every snooze cycle gives you a few minutes of fragmented, low-value sleep and then a fresh awakening to be impaired from. The full accounting is in what the snooze button costs.
4. Do something with a response requirement in the first five minutes. Not a screen you scroll. A task that fails visibly if you get it wrong: a stretch of arithmetic, making the bed properly, a cold rinse. See the first 30 minutes after waking for what is happening chemically while you do it.
5. Measure instead of asking yourself. Given how badly self-report tracked performance in the Van Dongen data, “I feel fine” is the one input you should weight least. A rough number you actually collect beats a confident feeling.

Where WakeSharp fits, and where it honestly does not

That last point is the one the app is built around, so here is the specific claim and its limits.

Reaction Tap is one of WakeSharp’s warm-up games. It runs after you have already dismissed the alarm, and it is not a mission. Missions are the tasks that dismiss an alarm (on the free tier, Mind Games and Photo Proof). Reaction Tap cannot dismiss anything, and a slow tap has no effect on whether your alarm is over. The warm-up suite is a short set of games you play once the morning has actually started, and it produces a Sharpness Score.

The score is deliberately scored against your own rolling baseline, not against other users and not against some ideal. That design choice is not modesty, it is the only defensible reading of the data.

Because a phone tap test is not a lab psychomotor vigilance task, and it would be dishonest to imply otherwise:

• Touchscreen latency inflates every number. Researchers examining touchscreen versions have flagged differences on the order of 50 to 100 milliseconds in response speed compared with purpose-built hardware, along with more lapses per minute (Sleep Health, 2018). Your phone’s absolute milliseconds are not comparable to a published figure.
• The trial count is small. A lab session is ten minutes of stimuli. A warm-up game is a handful. Fewer trials means noisier estimates, full stop.
• Games have practice effects. The original task was engineered to avoid them. A game you enjoy is not.

What survives all of that is the trend. When Arsintescu and colleagues compared a touchscreen task against the standard device across extended wakefulness, both tracked the same direction: reaction times slower and lapses higher as sleep loss accumulated (Accident Analysis and Prevention, 2019). Absolute values shift with the hardware. The shape of the curve holds. That is precisely why your score is compared with your own last two weeks and nothing else.

Two more honest notes. WakeSharp does not track your sleep, so it knows nothing about what stage you were pulled from or how long you slept. And a Sharpness Score is not a fitness-to-drive test, a diagnostic, or medical advice. It is a number that is worth glancing at on a morning when you were about to tell yourself you felt fine.

Your alarm rings free, forever. No ads.

FAQ

How long does it take for reaction time to return to normal after waking?

Most of the recovery happens in the first 15 to 30 minutes, but the tail is long. Jewett and colleagues found dissipation followed a saturating exponential with time constants of about 40 minutes for alertness and about 70 minutes for cognitive performance, approaching its asymptote only after two to four hours (Journal of Sleep Research, 1999). Hilditch and McHill’s review puts return to pre-sleep levels at around 30 minutes, with full recovery possibly taking an hour or more.

What counts as a slow reaction time in the morning?

In the research literature the meaningful threshold is not an average but a lapse: any response slower than 500 milliseconds on the vigilance task. Rested adults produce very few. Santhi and colleagues recorded roughly two and a half times as many lapses immediately on waking as at baseline (PLOS ONE, 2013). Absolute millisecond figures from your phone will read higher than lab numbers because of touchscreen latency, so compare them only with your own.

Can a phone app measure reaction time accurately?

Accurately in absolute terms, no. Touchscreen implementations have been shown to add tens of milliseconds and produce more lapses than dedicated hardware. Usefully, yes, as a relative measure: touchscreen versions track sleep loss in the same direction as the standard device. Treat the trend against your own baseline as the signal and ignore the raw number.

Why do I feel awake but keep making mistakes?

Because alertness and performance are separate things that recover on different schedules, and self-report is the less reliable of the two. Participants under chronic sleep restriction kept getting objectively worse while their subjective sleepiness ratings plateaued (Van Dongen et al., Sleep, 2003). Feeling fine is weak evidence. If persistent daytime sleepiness or unexplained slowness continues past the first hour of your day, that is a conversation for a clinician, not an app.