Study: Galaxy Watch6 biosignals predicted vasovagal fainting up to five minutes — not a device feature

Samsung says a peer-reviewed clinical study found that biosignals collected by the Galaxy Watch6 could predict vasovagal fainting up to five minutes before it happened during hospital testing. But the company also says the findings do not reflect a feature currently available on the watch.

The research, published online April 5 in European Heart Journal — Digital Health, followed 132 patients at Chung-Ang University Gwangmyeong Hospital in South Korea whose symptoms suggested neurally mediated syncope, another term for reflex fainting. Led by Jun Hwan Cho, the study prospectively enrolled patients between March 2023 and July 2025 and monitored them during head-up tilt testing, a standard hospital test used to provoke fainting episodes under controlled conditions. Participants wore a Samsung Galaxy Watch6 that collected photoplethysmography, or PPG, signals. Of the 132 patients, 63 fainted during testing and 69 did not. Using watch-derived heart-rate-variability data and a machine-learning model, the researchers reported prediction of impending vasovagal syncope up to five minutes in advance. For that five-minute window, the paper reported an AUROC of 0.91, specificity of 0.64 at a fixed sensitivity of 90%, and accuracy of 84.6%.

Vasovagal syncope is a common type of fainting caused by a sudden drop in heart rate and blood pressure. While many episodes are not dangerous in themselves, the loss of consciousness can lead to falls and injuries. A warning even a few minutes ahead could give someone time to sit or lie down, alert another person or move away from a risky situation. In Samsung’s press release, Cho said, “Up to 40% of people experience vasovagal syncope over their lifetime, with one-third experiencing recurrent episodes.”

The paper and Samsung describe the work as the first prospective study to show presyncope prediction using PPG data from a commercial smartwatch. That claim should be read alongside the study’s limits, which the paper spells out clearly. It was a relatively small, single-center study using one device, and the model was built around fainting episodes triggered during tilt-table testing in a hospital, not spontaneous events in daily life. The authors said the findings still need external validation and real-world testing. The reported specificity also suggests false alarms would be a practical issue if such a system were ever turned into a real-world alert tool, because moderate specificity means some warnings would happen when fainting did not.

The study’s disclosures also warrant attention. Several co-authors were employees and shareholders of Medical AI Co., Ltd., according to the paper’s conflict-of-interest section. The paper reported no funding.

Samsung announced the findings in a newsroom post dated May 7 in Korea, with a closely matching U.S. version posted May 6. The company framed the research as part of a broader push toward preventive health monitoring. “This study is an example of how wearable technology can help shift healthcare from being designed for ‘post-care’ to a model of ‘preventive care,’” said Jongmin Choi, head of the Health R&D Group in Samsung Electronics’ Mobile eXperience business. But Samsung’s U.S. press page also includes the key caveat: “Findings are based on the research study and do not reflect a feature currently available on the device.”