The science of sleep is still full of mysteries, and it's part of the reason why, in 2017, nobody "owns" sleep tech yet. Sure, plenty have taken a stab: Hello, Withings, Fitbit, Beddit. But progress beyond basic tracking has been slow, and compared to lab conditions, wearables and even dedicated devices can only go skin deep.
Rythm, a French-American startup, claims to have made a wearable not only capable of lab-condition sleep tracking, but able to improve the quality of our slumber using stimulations. We first spoke to Rythm a year ago, and since then it's launched a beta trial for its Dreem headband and begun work on an AI platform for diagnosing sleep disorders. It's now preparing for the full launch of Dreem, a first-of-its-kind wearable it claims is 95% accurate when held to medical data. Maybe you find it hard to fall asleep; maybe you struggle in the mornings; or perhaps your sleep just isn't efficient. Rythm claims it is about to solve the science of sleep, no lab required.
Read next: The best sleep trackers and monitors
These are big claims, but Rythm can't be accused of not doing its homework. The company studied the sleep behaviour of 500 people up close during its beta trial. "We have almost 30,000 nights in our database, which is one of the largest sleep databases in history," Rythm CEO and co-founder Hugo Mercier tells us.
We've talked a lot about the different stages of sleep, and Rythm's goal with Dreem is to maximise the amount of deep sleep people get each night. Important physiological processes take place during deep sleep; it's where our brain and muscles are restored, and where memories are consolidated. Hugo says they saw a 30-40% improvement in deep sleep efficiency and duration during testing, which analysed participants ranging in age and levels of health.
This is what Rythm claims it can do
"We have approximately the same efficiency across all the segments - young, old, sick, healthy - because of the personalisation of the algorithm," he says. It's hard not to be skeptical of any consumer tech that claims it can match the power of an electroencephalogram (EEG), but Mercier insists that Rythm has managed to achieve just that. "We compared our data with medical data and we have more than 95% correlation. So it's the same quality of what you would find in a sleep centre." However, he doesn't claim Dreem will be a "medical device" per se.
Hugo claims Dreem can boost the quality of sleep without modifying the amount of sleep needed, and possibly even decrease one's sleep duration - all of which is done using sound stimulations. If Rythm can make good on these promises, it would put medical-grade technology into the hands of consumers for a fraction of the cost and on a daily basis. An overnight polysomnogram test can cost thousands of dollars, the final Dreem device should cost somewhere in the $350 area. Hugo says it won't just "show raw data" either, but try to provide contextual, actionable feedback.
Fitbit recently steered this way with its Sleep Insights feature, which provides more actionable insights based on your sleep patterns and daily activity; rather than just telling you how well you slept, it tries to teach you why.
Mercier won't reveal much more about the new product, but says it will include a broader spectrum of features, focusing on more use cases (such as falling asleep faster, tackling insomnia, having more energy through the day). He also says the design has been improved to make it lighter, smaller and more comfortable to wear - which is important for anything you have to wear it in bed each night.
There's a lot to live up to, but it's what people do with this data that could be most interesting. Rythm has just announced plans to roll out a collaborative open-source platform, named Morpheo, that will use machine learning to predict and diagnose sleep disorders. It's built to use AI to data from wearables and traditional medical polysomnography to seek out patterns, and potentially pave the way for more preventative healthcare. Could, for example, your sleep patterns tell you that you're at risk of certain diseases?
"Basically you have a lot of correlation between bad sleep and certain disease," he says. "You have neurodegenerative diseases like Alzheimer's and Parkinson's, you have cardiovascular disease, high blood pressure, etc. You have mental disorders as well, and diabetes, and all these diseases are correlated with sleep.
"We will be able to, in a few years, by looking at the data of the users, say, 'Be careful, you're going to have Alzheimer's in five years from now because we are going to be able to identify patterns in the data."
That might be unsettling, but what if we could use sleep data to prevent health conditions down the line? Mercier says that correlations between sleep patterns and some of these diseases can already be picked out, but predictive analysis will take longer. "I think we need three to five years for that - that's why it's more of a long term moonshot project for us, and we made [Morpheo] open source."
"In the future, healthcare and the consumer tech world are going to merge a lot," says Hugo. "It's inevitable. Healthcare is too costly, you have a bottleneck on human resources. You can wait up to six months to get a sleep test. It's costing you, the insurance, the healthcare system a lot."
Rythm is planning for a launch this summer. If it can live up to the promises above, 2017 might see the arrival of a big player in sleep technology.
How we test