Sleep research

935 sleepers, 474,852 nights, and one number that decided who gained sleep

The peer reviewed Apollo and Oura analysis found real gains in total sleep time. They scaled with a setting you control yourself.

Published August 18, 2026 · 8 minute read

Most sleep research on consumer wearables is small, short and done in a laboratory. This one was none of those things. It followed 935 people through 474,852 nights of their own lives, in their own beds, over three years, and it measured them with a ring most of them had already bought for their own reasons.

The result is worth reading carefully, because the number that travels is not the number that matters. What the analysis really found is that the benefit had a price of admission, and the price was time.

The setup

What was actually measured

Researchers looked at a community of people who were wearing two devices at once: the Apollo wearable, which delivers what the paper calls transcutaneous vibratory stimulation, and the Oura Ring, which was doing the measuring. The window ran from January 2019 to May 2022.

Two things about that design are unusual enough to name. The first is that nobody was assigned to anything. People used their devices the way they wanted to, which is exactly why the usage varied so much, and exactly why the researchers could see what varying it did. The second is that the ruler was independent of the thing being tested. The Apollo does not measure sleep. The Oura does. That separation is the reason the numbers are worth discussing at all.

Finding one

The odds of a short night fell by 77 percent

The headline result concerns people the paper calls short sleepers, meaning those whose nights ran under six hours. In that group, using the device overnight was associated with a 77 percent reduction in the odds of recording a short total sleep time.

It is worth being precise about what an odds reduction is, because it is easy to hear it as something larger than it is. It does not mean anybody slept 77 percent longer. It means that on nights with stimulation, a short night was substantially less likely to happen to the people who were prone to them. The paper is describing frequency, not duration.

Finding two

The gain was dose dependent, and the published dose is larger than most people run

Here is the sentence that changes how you would actually use the thing. The abstract reports the increase in total sleep time as, in its own words, adding 36 minutes to the baseline sleep of short sleepers who used less than 6 hours of sleep and 240 or more minutes of stimulation, and states that the relationship was dose dependent.

Read the condition again. Two hundred and forty minutes. That is four hours of stimulation across the night. Not a wind down session before bed. Not twenty minutes while you fall asleep. Four hours, running while you are asleep, is the usage tier that the published figure belongs to.

This is the part that quietly rewrites the instructions. If the benefit scales with hours of stimulation during the night, then where you wear it and whether it keeps running after you fall asleep stop being comfort preferences. They become the dose. A device sitting on a nightstand at 2am is delivering nothing, and the published number was never about that device.

Finding three

The extra sleep was not made of the cheap kind

Extra minutes are only interesting if they are good minutes. A longer night made entirely of light sleep is a longer night spent barely asleep. The paper addresses this directly: extended sleep showed a higher proportion of time spent in rapid eye movement sleep, at the expense of light sleep.

The mechanism

Why vibration, of all things

The premise behind the device is not that a buzzing wristband makes you drowsy. It is that the body decides when it is safe to let go, and that decision is made below the level of thinking about it. David Rabin, the physician who co-founded the company and who is one of the five authors on the paper, put the idea this way when the research was announced:

Sometimes the brain isn't keeping you awake. Sometimes the nervous system simply doesn't believe it's safe enough to let you sleep. David M. L. Rabin, MD, PhD, co-founder and chief medical officer, writing on apolloneuro.com, July 17, 2026

The proposed route is touch. Gentle vibration through the skin is a sensory signal, and sensory signals of that kind are read by the autonomic nervous system, the part that also runs heart rate and recovery. The company's stated goal is not to override that system but to change what it concludes.

Rather than overriding the nervous system, the goal is to remind it that the environment is safe enough to rest. David M. L. Rabin, MD, PhD, on apolloneuro.com

The fine print

What this study does not say

A paper is more useful when you know its edges, and this one has several worth stating out loud.

One of my favorite aspects of this study is that it didn't happen inside a sleep laboratory. These data came from people living their everyday lives. David M. L. Rabin, MD, PhD, on apolloneuro.com

In practice

What owners describe, in their own words

The published reviews are worth reading next to the dose finding, because the people describing the clearest effects tend to be describing overnight use rather than a session before bed.

Two of those four describe the same specific thing, which is getting back to sleep after waking rather than getting to sleep in the first place. That is the part of the night the dose finding is about, and it is the part a bedtime routine cannot reach.

Wearing it

Three places, one of which matters at night

The device is worn on the wrist, on the ankle, or clipped to clothing. During the day the choice is comfort. At night it is not, because the only version that delivers hours of stimulation is the one that is still on your body at four in the morning and is not waking you up by being there.

Editorial visual to be produced: the device worn on the ankle with the sleep band, in a dark bedroom

The money

What it costs, and what comes with it

The version the company sells as its starter kit pairs the wearable with a year of the software that decides when to run overnight, which is the piece the dose finding actually depends on.

The product bullets above are the company's own, reproduced from its product page. The 60 minute figure it quotes there is a company claim drawn from its wider research programme, not the figure published in the study discussed on this page, which is the 36 minute result at the highest stimulation tier.

Before you decide

Who this is not for

In one line

The finding, restated plainly

Across 474,852 nights, people who slept badly and ran the device for hours while they slept recorded more total sleep, a sharply lower chance of another short night, and a shift in that added time toward REM. The relationship was dose dependent, which is both the most encouraging thing in the paper and the most demanding one. The result is available, and it is available at four hours.

See the Apollo wearable $448 · HSA and FSA eligible · 30 day guarantee

Study: Mandala M, Krishnan S, Weathington N, Breus M, Rabin D. Association Between Duration of Transcutaneous Vibratory Stimulation Delivered by the Apollo Neuro Device and Extension of Total Sleep Time. JMIR mHealth and uHealth, May 19, 2026, online ahead of print. doi:10.2196/79588, PMID 42163014. Company commentary on the study: apolloneuro.com, July 17, 2026. Product details and reviews: apolloneuro.com, read August 18, 2026.