University of Southern California researchers have developed an AI-powered ring that uses near-infrared light to track sweat activity beneath the skin as a possible measure of stress. The team, led by assistant professor Yasser Khan and PhD student Farhad Hassan, published its findings in Science Advances.

Wearables such as the Fitbit Sense typically measure electrodermal activity (EDA), changes in skin conductance caused by sweat glands. The method requires electrodes to maintain steady contact with the skin; conductive gels can dry out, while movement can affect readings over time.

The ring uses two near-infrared LEDs to detect subtle changes beneath the skin that occur when stress activates sweat glands. “Our idea is very simple. We want to look at this small sweat under the skin with light,” Khan said. One LED shines at 960 nanometers and responds to changes in water content linked to sweat. The other operates at 1,250 nanometers as a stable reference. The researchers call the method optodermal activity, or ODA.

“This is the first time that someone is showing that stress can be measured with light,” Khan said. In experiments with four participants, the optical signals showed an average 86 percent agreement over time with a commercial gel-based EDA sensor. The system also measures heart rate and heart rate variability. A temperature sensor and accelerometer help distinguish psychological stress from movement or heat-induced sweating.

In short-term tests involving six participants, the ring distinguished stress from relaxation 92.3 percent of the time, compared with 77.95 percent for the gel-based EDA sensor. When researchers tested four conditions—relaxation, physical movement and heat-induced sweating among them—the ring reached 100 percent accuracy, compared with 74.58 percent for EDA.

After 12 hours of wear, tested with two participants, the gel-based sensor’s accuracy in distinguishing stress from relaxation fell to 57.35 percent, while the ODA ring recorded 79.55 percent.

The machine-learning model runs on the ring, so raw biometric data does not have to be sent to an outside server. According to the USC announcement, this reduces delays and keeps sensitive personal data on the device.

Khan plans a yearlong follow-up study with 40 participants in an office setting. The researchers are also working to reduce interference from heavy surface sweat and to test a broader mix of cognitive and psychosocial stressors. The technology is in the patent process and may later be offered for commercial licensing through USC’s Stevens Center for Innovation.