With the wave of quantified self devices popping up over the last few years, the ŌURA ring has always been one that is hyped by the biohacker community - and I believe for good reason. Having used my ŌURA V2 consistently for over eight months, I confidently say that this has been a great tool to gage sleeping habits, and the effects of various supplements and practices upon them. The ŌURA has no shortage of reviews as well, and in these you would stumble upon many a reference to a study that the Stanford Sleep Institute performed on the device.
Curious to check out the study for myself, I gave it a read over, well… a few read overs. It looked like the perfect article to pose for a case study, and so here I have prepared just that!
Hopefully this can share some third party insight into the ŌURA , and also introduce the approach and challenges of clinical trials.
A breakdown of the article:
Below I’ve provided a breakdown of what the study article reported. Much like the original article is a simplification of the study conducted, so this is my simplification of the article. Do bear that in mind as there is undoubtedly information loss along the way.
To evaluate the performance of a multi-sensor sleep-tracker (the ŌURA ring) against polysomnography (PSG) in measuring sleep and sleep stages.
Sleep stages using the ŌURA appear to be primarily determined by Pulse Waveform using Infrared Sensors measured from the two arteries in a finger. This is supported with a 3D accelerometer, gyroscope, and body temperature sensor. You can find out more information on the ŌURA device, from the creators themselves - click through here.
A polysomnography system uses numerous inputs to monitor many different body functions during sleep. These include brainactivity with a electroencephalogram (EEG), eye movements with a electrooculogram (EOG), muscle activity or skeletal muscleactivation with electromyography (EMG), and heart rhythm with a electrocardiogram (ECG). The stages of sleep are then interpreted by a technician. This form of sleep analysis is considered the “gold standard” by the scientific and medical community. You can read more information on PSG systems here.
41 participants (28 Male: 13 Female: 35 Caucasian: 6 Non-Caucasian);
Average BMI: 21.6±3.5 kg.m−2.
Participants were recruited from the San Francisco Bay Area as part of a long-term multi-site sleep study funded by the National Consortium on Alcohol and NeuroDevelopment in Adolescence.
Each of the participants sleep was recorded using both the first generation ŌURA (with software version 1) and polysomnography (PSG) for one night only. The participants had the choice of two ŌURA of ring size US11 & US7, therefore it was necessary that the finger which wore the ring varied to fit.
Evaluations of Sleep, Wake, Deep Sleep, REM Sleep, & Light Sleep were then compared using Epoch-by-Epoch (EBE) analysis and Bland-Altman plots. For more information on these analysis techniques, see the articles here and here respectively.
Results demonstrated agreement between the PSG & ŌURA for Sleep Onset Latency (SOL), Total Sleep Time (TST), and Wake After Sleep Onset (WASO).
From Epoch-by-Epoch (EBE) analysis, the ŌURA had a 96% sensitivity to detect sleep, agreement of 65% to detect light sleep, 51% to detect deep sleep, and 61% to detect REM sleep. Specificity in detecting wake was 48%.
Below is an example of the sleep stages plotted as a function of time in the night (a hypnogram), obtained from a participants recordings showing typical PSG-ŌURA discrepancies.
Some sources of discrepancy could be identified by the finger on which the ŌURA was worn. It can be seen from the figure 2 (below) that a significantly higher number of the Light and REM sleep categorisation differences occurred when the ŌURA was worn on the ring finger, as compared to when worn on index or other fingers.
The ŌURA did not present systemic over/under-estimation, of which has had significant impact in the findings of the other fitness tracker investigations. But like other trackers the ŌURA had very high accuracy for detecting sleep, and then comparably low for detecting wake. This is understandable in trackers dependant on movement measures (actigraphy), however is unclear why the ŌURA exhibits this trait as Heart Rate Variability (HRV) demonstrates significant changes on waking state, which the ŌURA tracks.
It was also investigated if the ŌURA was able to detect a well-established literature finding, that being the decline in slow wave sleep (N3) with advancing age during adolescence. The ŌURA was successful in demonstrating this trend, as shown in figure 3 below.
It was also recognised that only two sizes of the ŌURA ring were available, and so perhaps with better fit, better accuracy could be obtained. Lesser performance on the ring finger may be explained by lesser blood supply. Previously it has been found that there is a finger dependant difference in pulse oximetry signal per finger, due to differing amounts of oxygen in the blood. 
The study concluded that “...first results of the ability of the ŌURA ring to distinguish sleep stages could be viewed as promising, however, future development and validation is needed.”
My personal take-homes
As a personal quantification tools enthusiast, I really appreciated this study setup. For me, this premise of comparing a consumer technology, to a widely renown standard provides real useful insight for my life.
To my judgement, this study presented very mixed outcomes. Through its results we can be quite confident in the ŌURA’s ability to track total sleep times to an accuracy applicable to daily living. But determination of the actual sleep stage (light, REM or deep sleep) was below what we might hope. In the light of the fact that an overwhelming contributor to inaccuracy occurred when worn on the ring finger, I think it is a safe assumption that we would be presented with far better results had the study been repeated when excluding ring finger use. In finding this issue, the early study has paved the way for the future, by filling the pitfalls with its own sacrifice. Remember, one study never paints the full picture, and this is a clear example of this fact. It also presents a great example to me on how what might seem like very small differences in use of a device, might lead to very different outcomes.
Also as a learning, this paper was a valuable reminder of why a thorough review and consideration of any paper is necessary for before judgement, for I can honestly say that my thoughts on the findings flip-flopped so many times I felt like a fish. In the process, finding many of the logical flaws in my previous thinking.
Wear the ŌURA on a non ring finger - *because users are not queried which hand/finger the ŌURA is mounted on prior to use, I assume the finger difference has not been resolved.
Ensure good fit according to ŌURA guidelines, not just before purchase, but also as you use it on a day to day.
Consider potential variance in accuracy for any reason before coming to conclusions on any nights sleep.
I also passionately encourage every reader to look through the article itself. Then meditate on the information. And read over it again!
It will no doubt be a valuable exercise in thought recognition and lead to more valuable understanding.
NOTE: The ŌURA has been developed significantly since the study, having released a new hardware model and numerous software updates.
Other Valuable Reading
VitaKea Founder & Caretaker
Key players in my current health regime include meditation, regular exercise (type depending on current fixation), early to bed, early to rise, packing in as many fungi & leafy greens as possible.
My supplementation varies as I test out new protocols. That considered, I prioritise anything that can be utilised for long-term benefits. Of course, for special cases I can definitely appreciate a good stimulant, or perspective modifier.
I am not an unbiased source. As a fond user of the second generation ŌURA device, and although I have no direct financial ties to the ŌURA, I do place good faith toward them and their mission.