Every day, millions of members rely on Oura Ring to better understand their rest, readiness, and overall health. We’ve always provided our community clarity about how our technology works—what Oura measures, how PPG technology works, and how Oura’s algorithms perform against clinical standards like polysomnography (PSG). 

A recent baseless, opportunistic legal allegation seeks to discredit how consumer wearables—including Oura Ring—track and estimate sleep stages. We dispute these allegations and stand firmly behind our research and accuracy. 

We are proud of the work our team of more than 50 MDs and PhDs have undertaken to ensure that Oura is scientifically sound and drives health outcomes, and we will defend our work in the appropriate legal forums. We have a history of being transparent and open, regularly publishing research and scientific validation to educate anyone who is curious about our technology. 

Recent research validating the accuracy of our sleep staging algorithm includes the following peer-reviewed studies (additional details below):

“At Oura, scientific and clinical rigor are core to how we build our products and features,” Shyamal Patel, PhD, SVP of Science at Oura says. “We hold ourselves to a high bar so we can earn the trust of our members as well as the scientific and clinical community—and because transparent, rigorous science can advance our understanding of sleep and health, from an individual level to the population level.”

RELATED: Inside the Ring: Developing Oura’s Latest Sleep Staging Algorithm

How Does Oura Track Sleep?

From the beginning, Oura was designed to prioritize accurate sleep data using photoplethysmography (PPG) (which measures the volumetric changes in your arteries using light reflection), temperature sensors, and motion sensors (accelerometer). 

“Sleep is not solely a simple state of brain activation,” says Max de Zambotti, PhD, Director of Health Science at Oura. “Sleep is also a whole-body biobehavioral state, and Oura captures multiple physiological signals simultaneously, such as heart rate, movement, and body temperature trends, to determine when someone falls asleep and which sleep stage they’re in.” 

These signals include: 

  • Heart Rate (HR): Beats per minute across the night.
  • Heart Rate Variability (HRV): Beat-to-beat variation in time intervals, reflecting autonomic nervous system activity (sympathetic vs. parasympathetic tone).
  • Respiratory Rate: Minute variations in heart rate caused by breathing (respiratory sinus arrhythmia).
  • Body Temperature & Movement: Skin temperature deviations and 3D accelerometer signals tracking physical restless periods.

RELATED: Independent Study Finds Oura Ring Most Accurate Wearable for HRV & RHR

Each sleep stage (awake, light sleep, REM, and deep sleep) is characterized by distinct biosignals. Sleep stages are not only reflected in electroencephalography (EEG) signals (which measure brain waves), but they also affect multiple signals throughout the entire body. 

1. During deep sleep, or slow-wave sleep, autonomic tone shifts toward parasympathetic dominance. Heart rate reaches its lowest points, HRV stabilizes, and breathing becomes highly rhythmic.

2. In REM (rapid eye movement) sleep, autonomic activity becomes erratic. Heart rate and breathing become variable, mimicking wakefulness, while muscle tone drops.

3. Light sleep and wakefulness are characterized by shifting heart rate patterns, micro-arousal movements, and temperature changes.

By continuously analyzing these physiological markers together, Oura’s algorithms map these physiological signals to sleep stages with precision.

READ MORE: The Accuracy Advantages of Finger-Worn Wearable Devices 

How Oura Compares to Polysomnography (PSG)  

Polysomnography (PSG) is the clinical standard for sleep assessment. Performed in a sleep lab, PSG uses head-attached electroencephalography (EEG) electrodes to measure brainwaves directly, alongside eye movement (EOG), electrical activity of the heart (ECG), breathing patterns and muscle activity (EMG).

In comparison, Oura Ring provides a consumer-friendly, unobtrusive, and non-diagnostic method of measuring sleep. Rather than measuring brain activity directly, Oura measures behavioral and peripheral physiological signals that exhibit a strong relation with cortical brain activity during sleep.

In a foundational 2021 study published in Sensors, researchers evaluated 440 nights of simultaneous polysomnography (PSG) and Oura Ring sensor data from 106 individuals (over 3,400 combined hours). The findings: while a simple motion/accelerometer model achieved only 57% accuracy for 4-stage sleep classification, adding PPG-derived autonomic nervous system (ANS) signals and temperature features boosted 4-stage sleep classification accuracy to 79% and 2-stage (sleep/wake) accuracy to 96%.

The Research Behind Oura’s Staging Algorithm  

Oura’s sleep staging algorithms are developed using advanced machine learning models trained on one of the largest datasets in sleep research—incorporating over 1,200 nights of clinical PSG data across diverse demographic populations, ages, and skin tones.

Researchers and top academic institutions worldwide have rigorously evaluated Oura’s performance against PSG.

Peer-Reviewed Studies

Below is a select list of recent publications that have evaluated Oura’s latest sleep staging algorithm, released in 2023:

  • Brigham and Women’s Hospital Study, 2024: In an external validation study conducted by researchers at Brigham and Women’s Hospital comparing major consumer sleep trackers against PSG, Oura Ring was found to be the most accurate consumer sleep tracker tested for 4-stage sleep classification (Light, Deep, REM, and Wake). Oura exhibited substantial agreement in the determination of specific sleep stages (Kappa > 0.61). The other wearable devices demonstrated moderate agreement (Kappa statistic < 0.61). Epoch-by-epoch analyses demonstrated sensitivity between PSG and the Oura Ring > 75% for all the sleep stages (range: 76.0–79.5%). Oura achieved 76.3% agreement for 4-stage classification and 92% for 2-stage (sleep/wake) classification.
  • National University of Singapore Study, 2022: Researchers evaluated Oura’s algorithm across 157 nights of at-home and laboratory sleep data. Oura achieved 76.4% agreement with clinical PSG for 4-stage sleep classification across four nights and 92–93% agreement for 2-stage (sleep vs. wake) classification.
  • National University of Singapore Multi-Device Comparison, 2023: Researchers compared six wearables—an EEG headband, research-grade actigraphy, and four consumer trackers—against in-lab PSG scored by three-reader consensus in 60 adults aged 18 to 70. Oura was the best-performing non-EEG device for 2-stage (sleep vs. wake) classification, with 91.1% accuracy and kappa of 0.64, ahead of Fitbit Sense (89.4%, 0.58) and research-grade actigraphy (87.4%, 0.47), and it had the highest 4-stage kappa range among consumer trackers (0.55–0.70). The authors concluded that „only the Dreem headband and Oura ring achieved kappa values indicating substantial agreement” with PSG.
  • University of Tokyo Validation Study, 2024: One of the largest wearable validation studies to date, evaluating Oura’s Sleep Staging Algorithm 2.0 against ambulatory PSG across 96 participants aged 20–70 and 421,045 30-second epochs, with up to three nights per participant. For 2-stage (sleep vs. wake) classification, Oura achieved 91.7–91.8% accuracy, 94.4–94.5% sensitivity, and 73.0–74.6% specificity, with prevalence- and bias-adjusted kappa of 0.83–0.84 — near-perfect agreement with PSG. Sleep staging accuracy ranged from 75.5% for light sleep to 90.6% for REM sleep. Oura’s measurements did not significantly differ from PSG for time in bed, total sleep time, sleep onset latency, wake after sleep onset, light sleep, or deep sleep.
  • MDPI Sensors Validation Study, 2021: This foundational study evaluated 440 nights of simultaneous polysomnography (PSG) and Oura Ring sensor data from 106 individuals (over 3,400 combined hours). The findings demonstrated the crucial power of multi-sensor fusion: while a simple motion/accelerometer model achieved only 57% accuracy for 4-stage sleep classification, adding PPG-derived autonomic nervous system (ANS) signals and temperature features boosted 4-stage sleep classification accuracy to 79% and 2-stage (sleep/wake) accuracy to 96%.

Committed to Transparency

While Oura is not a medical device intended to diagnose medical conditions like sleep apnea or insomnia, it provides an unprecedented, accessible lens into night-to-night sleep architecture and long-term health trends.

We will continue to publish our findings, collaborate with leading academic institutions, and transparently communicate what our device measures, how our algorithms operate, and how members can use this data to improve their daily lives. We remain confident in our science, proud of our engineering, and committed to defending our technology in court.

FAQ: Oura’s Sleep Tracking Science & Accuracy

Q: Is it accurate that Oura cannot track sleep cycles because „sleep happens in the brain, not on the finger?”  

A: No. Oura tracks sleep through the body-wide physiology that goes hand in hand with what the brain is doing. This claim rests on the fact that sleep has traditionally been defined and scored using EEG—brain activity. But sleep is not only EEG, and EEG is not a direct measure of sleep either; it is one window into it. Sleep is a whole-body physiological state, and the science is clear that during sleep, there is a tight coupling between central and autonomic nervous system activity. 

The autonomic nervous system modulates heart rate, heart rate variability (HRV), peripheral blood flow, respiration, and body temperature in distinct, reproducible patterns during each sleep stage. The stage transitions that EEG detects in the brain are accompanied by corresponding, measurable changes in peripheral physiology. That autonomic signal is among the main signals wearables use to infer sleep stages. It is not a coin flip; it is physiology, measured through a different—and legitimate—window.

By taking measurements from the digital arteries in the finger—where PPG signal strength is superior to wrist-based sensors—Oura tracks these physiological manifestations. The algorithm evaluates multi-sensor inputs to map peripheral body signals to sleep stages.

Q: Is it accurate that Oura’s sleep staging relies on „AI guesswork” that is no better than a „coin flip”? 

A: No. Assertions framing Oura’s sleep staging as a „coin flip” mischaracterize peer-reviewed science. 

Multiple peer-reviewed studies evaluate Oura’s actual accuracy against gold-standard polysomnography far higher:

  • MDPI Sensors (2021): Demonstrated 79% agreement for 4-stage classification and 96% agreement for 2-stage (sleep/wake) classification using PPG + multi-sensor fusion.
  • Brigham and Women’s Hospital (Sensors, 2024): Demonstrated 76.3% agreement for 4-stage classification and 92% for 2-stage (sleep/wake) classification. This study rated Oura Ring as the top-performing consumer tracker tested for 4-stage classification accuracy against PSG.
  • National University of Singapore (Nature & Science of Sleep, 2022): Showed 76.4% 4-stage agreement with clinical PSG across 157 test nights.

Furthermore, in multi-stage classification (where a device must distinguish between 4 distinct categories: Light, Deep, REM, and Wake), a random guess or „coin flip” would yield an expected baseline accuracy of 25%—not 50%. 

Q: Oura claims „95% Sleep Staging Accuracy compared to a clinical sleep lab.” What does this percentage actually measure? 

A: In a validation study, you sleep one or more nights wearing both the ring and full clinical polysomnography (PSG), the sleep-lab gold standard. The night is then chopped into 30-second segments called epochs—roughly 900 of them in an 8-hour night. A trained technician labels each epoch from the PSG, the ring produces its own label for the same epoch, and the two are compared one by one. „Accuracy” is simply the share of epochs where the ring and the technician agree. This is why the numbers are demanding: getting most of the night right still leaves plenty of room to disagree on individual 30-second windows.

Accuracy is reported across a few different dimensions:

  1. Sleep vs. wake detection: whether you were asleep or awake in each epoch. Oura’s algorithms consistently reach 90–96% agreement with PSG. This is what the 95% figure refers to.*
  2. Whole-night summaries: total sleep time, sleep efficiency, and when you fell asleep and woke up. These aren’t epoch-by-epoch measures — they compare the ring’s total for the night against the lab’s total, in minutes. Independent studies have repeatedly found no significant difference between Oura and PSG for total sleep duration.
  3. Four-stage classification: telling Light, Deep, REM, and Wake apart in each epoch. This is a harder problem, with Oura reaching roughly 76–79% agreement with human-scored PSG in healthy adults across peer-reviewed studies.

*Across multiple peer-reviewed studies comparing the Oura Ring against clinical sleep-lab recordings (polysomnography), the ring correctly detected sleep 94–98% of the time, with overall sleep/wake agreement of roughly 90–96%: 98% sensitivity and 96% overall accuracy for the full multisensor model (Altini & Kinnunen, 2021); 95% sensitivity and 93% accuracy across three nights at home (Ghorbani et al., 2022); 95% sensitivity and 91% accuracy (Ong et al., 2023); 94% sensitivity and 92% accuracy across more than 400,000 epochs of sleep (Svensson et al., 2024); and 95% sensitivity in an overnight clinical study (Robbins et al., 2024).

Q: How does Oura differentiate itself from medical devices? 

A: Oura has consistently maintained that the Oura Ring is a consumer wellness product designed for continuous, non-invasive health tracking—not a medical device intended to diagnose conditions like sleep apnea or clinical insomnia.

Medical polysomnography requires patients to sleep in an unfamiliar laboratory wired to over a dozen sensors, which often disrupts natural sleep patterns. Oura provides long-term, nightly trend analysis in a user’s natural home environment. Validating a consumer device against PSG as a benchmark does not imply it replaces diagnostic clinical equipment, but rather demonstrates the high fidelity of its non-invasive measurements.

Oura Ring is not intended to replace a clinical sleep assessment, but to complement it—offering accurate, meaningful, and actionable data that brings new insight into sleep during daily life and how it evolves over time. That longitudinal view is already enabling discovery at both ends of the scale: from what individuals notice in their own data, as documented in a case report in JAMA Neurology, to population-level findings such as cross-country differences in nocturnal sleep patterns drawn from large-scale wearable data.

Q: What is PPG, and why is the finger a better location for PPG than the wrist? 

A: Photoplethysmography (PPG) is an optical technology that uses LED light and photodetectors to measure arterial pulse waves and blood volume changes.

The finger is an ideal location for PPG because digital arteries run close to the surface on the palmar side. Compared to the back of the wrist—where skin is thicker, bone structures interfere, and blood vessels are smaller. This signal quality allows Oura to capture subtle beat-to-beat changes in Heart Rate Variability (HRV) and respiratory variations essential for sleep-stage modeling.

Q: What is the „gold standard” for sleep tracking, and how close can a wearable get to it? 

A: The gold standard is polysomnography (PSG), which combines scalp EEG (brainwaves), EOG (eye movements), and EMG (muscle tone).

It’s worth understanding how PSG results are actually produced. A certified technician reviews the recording and assigns every 30-second segment of the night a single label — Wake, N1, N2, N3, or REM — following standardized clinical rules. Two things follow from this. First, real sleep doesn’t change state neatly every 30 seconds; a segment can contain features of more than one stage, and the technician has to pick the one that predominates. Boundaries between stages are especially ambiguous, and light sleep (N1 in particular) is the hardest to pin down. Second, because the labels reflect expert judgment applied to ambiguous signals, two technicians scoring the same recording won’t fully agree. Peer-reviewed studies put inter-scorer agreement at around 83%, and it drops further when scoring people with sleep disorders.

This matters for interpreting wearable accuracy: the benchmark a ring is measured against is itself imperfect, so no device, including another PSG scorer, would reach 100%. Against that backdrop, achieving ~78–80% 4-stage agreement with PSG from an unobtrusive finger-worn ring is a high level of performance. In a head-to-head comparison of six wearables against three-reader consensus PSG, researchers found that „only Dreem and Oura achieved kappa values indicating substantial agreement with PSG.” (Dreem is an EEG headband, and Oura was the only non-EEG device to reach that threshold.)

Q: Oura says I slept well, but I don’t feel rested. What causes this discrepancy, and does it mean the algorithm is inaccurate? 

A: Subjective experience is often conflated with objective accuracy—when in reality, they measure two entirely different things.

In sleep science, accuracy refers strictly to how closely a device’s measured output reflects a user’s true underlying physiology when benchmarked against the clinical gold standard, Polysomnography (PSG). A member’s subjective perception of their sleep is real and meaningful, but it reflects a different dimension of health altogether.

It is well documented in sleep literature that an individual can experience a physiologically sound, well-structured night of sleep and still wake up feeling tired—or conversely, feel refreshed after a night of fragmented sleep. This mismatch is not evidence of device inaccuracy. Subjective sleep quality is influenced by a complex web of external and psychological variables, including mental stress, mood, caffeine timing, early-stage illness, circadian alignment, and waking expectations.

Treating a gap between how you feel and what your physiology showed as proof of sensor inaccuracy misunderstands what objective physiological measurement actually evaluates. Oura provides an objective, repeatable baseline of your body’s physical sleep architecture to help uncover those broader patterns over time.