A proposed class action filed in federal court on August 20 accuses Oura of selling clinical confidence through a consumer ring. The complaint targets claims including “79% agreement” with polysomnography and a newer “95% Sleep Staging Accuracy” line, arguing that a finger sensor cannot directly observe the brain activity, eye movement, and muscle tone used to define sleep stages in a lab.[1]
Oura disputes the allegations. Its response says the ring estimates stages from heart rate, heart-rate variability, movement, breathing patterns, and temperature, and that independent studies have compared those estimates favorably with polysomnography.[6] The scientific record supports a messier picture than either side’s cleanest sentence.
The ring can infer sleep stages. Its accuracy changes with the population, the metric, the stage, and the difference between a cohort average and one person’s strange Tuesday night. The product turns that conditional estimate into a polished score with no visible uncertainty.
The lawsuit attacks the sensor
The complaint’s sharpest technical argument is also its weakest. It says sleep happens in the brain and the ring therefore cannot measure sleep stages.[1] Direct clinical measurement does require polysomnography, which combines signals such as EEG, eye movement, and muscle activity. Inference from correlated peripheral signals remains possible. Weather stations infer conditions they do not touch. Spam filters infer intent they cannot read. The serious dispute concerns accuracy, disclosure, and use.
Oura’s own technical page says its algorithm combines movement, temperature, heart rate, and heart-rate variability. It reports 79% agreement with polysomnography across wake, light, deep, and REM classifications, based on a 2021 study and followed by later validation work.[2] A 2024 multi-night study of 96 generally healthy adults reported sleep-stage accuracy ranging from 75.5% for light sleep to 90.6% for REM, with overall sleep/wake accuracy above 91%.[5]
Another 2024 study compared Oura, Fitbit, and Apple devices against a single night of lab polysomnography in 35 healthy adults. Oura’s stage sensitivities landed between 76% and 79.5%, and its estimates for wake, light, deep, and REM duration did not differ significantly from the lab measurements at the group level.[4]
Those results kill the complaint’s categorical claim that peripheral sensing cannot produce useful sleep-stage estimates. They do not settle the product question.
Accuracy changes when the sleeper changes
A 2025 clinical study tested Oura and two other rings in a university sleep-lab population with sleep disorders and other medical conditions. Oura’s average total-sleep-time difference from polysomnography stayed below 12 minutes, which sounds excellent. Individual errors remained large. Four-stage accuracy for Oura reached 53.18%, and the researchers concluded that group-level agreement masked individual-level inaccuracies severe enough to prevent clinical use.[3]
The number shifted because the operating environment shifted. Healthy-adult validation, clinical cohorts, stage-by-stage sensitivity, total-duration error, and epoch accuracy answer different questions. Marketing compresses them into one word: accurate.
The lawsuit cites the clinical result as evidence of deception.[1] Oura cites favorable studies as evidence of validity.[2][6] Both can point at real numbers because “accuracy” has been carrying five jobs at once.
A score changes behavior
The interface matters because sleep trackers do not end at measurement. They interpret the night, rank it, compare it with prior nights, and suggest behavior. A raw pulse trace looks provisional. A score of 84 looks adjudicated.
That design converts a model output into a small authority system. The user wakes with one subjective experience and one numerical verdict. When they conflict, the screen has decimals, charts, historical baselines, and the prestige of a laboratory comparison. The body has a feeling.
The complaint describes consumers organizing their days and interpreting their health around those outputs.[1] Its legal claims remain allegations, and the court has not ruled on the merits. The behavioral mechanism is already visible in the product category: sensors collect proxies, a model assigns stages, stages feed a composite score, and the score becomes advice.
A serious uncertainty interface could expose four things without turning breakfast into a statistics seminar:
- confidence for each stage estimate and the composite score
- a plain distinction between total-sleep estimates and stage classification
- warnings when signals, motion, illness, medication, or unusual physiology reduce reliability
- validation ranges tied to cohorts instead of one universal accuracy badge
Consumer health companies usually bury those limits in research pages because uncertainty weakens the sales pitch. That choice also weakens the instrument. A score that admits when it is guessing gives the user a way to arbitrate between model, memory, and symptoms.
The error bar belongs on the screen
Oura has a defensible scientific response. Multiple studies show that its ring extracts real sleep information from peripheral signals.[2][4][5] The clinical evidence also shows why an aggregate validation number cannot guarantee a trustworthy classification for one person on one night.[3]
The lawsuit may succeed or fail on advertising language, consumer expectations, procedural issues, and evidence that has yet to be tested. Its durable contribution is forcing a product-design problem into public view. Consumer health inference has matured past novelty. Its interfaces still present probability as fact.
The next credible wearable will treat uncertainty as user-facing information. Confidence belongs beside the sleep score, cohort limits belong beside accuracy claims, and disagreement with symptoms should trigger caution instead of another optimization prompt. A ring can estimate the night. The screen should say how sure it is.
Sources
[1] https://clarksonlawfirm.com/wp-content/uploads/2026/08/COMPLAINT-26-cv-08686-Surber-v.-Oura-Inc.-et-al.pdf [2] https://ouraring.com/blog/new-sleep-staging-algorithm [3] https://pubmed.ncbi.nlm.nih.gov/40108409 [4] https://pubmed.ncbi.nlm.nih.gov/39460013 [5] https://pubmed.ncbi.nlm.nih.gov/38382312 [6] https://techcrunch.com/2026/08/21/oura-faces-lawsuit-accusing-it-of-misleading-consumers-about-sleep-tracking-accuracy