AI-enabled wearable sensor uses vibration to curb nocturnal scratching

Northwestern University researchers have tested a method to reduce the dreaded nocturnal itch-scratch cycle. Using an AI-trained wearable sensor with vibrating feedback, the device significantly reduced nocturnal scratching in adults with mild atopic dermatitis. The small trial demonstrated a decrease in scratching events and duration without impacting total sleep opportunity.

Atopic dermatitis is characterized by chronic pruritus and the itch-scratch cycle. Scratching at an itchy site creates more inflammation, which increases the itchiness and leads to more scratching. This cycle can severely disrupt sleep, causing additional fatigue and stress in those affected.

Previous research validated an AI-driven wearable sensor (ADAM sensor, Sibel Health) for detecting nocturnal scratching. The current study incorporated a vibratory feedback triggered by the sensor to assess its potential as a nonpharmacological intervention.

In the study, “Artificial Intelligence–Enabled Wearable Devices and Nocturnal Scratching in Mild Atopic Dermatitis,” published in JAMA Dermatology, researchers conducted a single-arm, two-stage cohort trial at Northwestern University’s Department of Dermatology.

Ten participants aged 18 or older with mild atopic dermatitis and moderate to severe scratching were enrolled. Disease severity was assessed using the Validated Investigator Global Assessment (vIGA) score.

The AI-sensor and feedback unit is described as a soft, flexible, dorsal hand-mounted device made of medical-grade silicone with wireless capabilities and a haptic motor that delivers 1.4G vibratory feedback at 10,000 RPM.

Participants wore the feedback-enabled sensor on the back of their hands for two weeks. The first week recorded baseline nocturnal scratching with the feedback disabled.

In the second week, the feedback mechanism was activated, delivering a vibratory alert upon detected scratching. Scratch events and durations were measured using a validated AI algorithm. A total of 104 nights and 831 hours of monitoring data were collected.

Statistical analysis demonstrated a 28% reduction in mean nightly scratch events (45.6 to 32.8) and a 50% reduction in per-hour scratch duration (15.8 seconds to 7.9 seconds). There were no significant changes to total sleep opportunities.

Most participants reported the device as easy to use and expressed interest in the methods for reducing scratching.

The act of scratching can be very automated, and scratchers can be unaware that they are doing it, acting subconsciously or even while asleep. Feedback from the device may disrupt this automatic action. It is also possible that the vibration itself might provide a mild counter-stimulus, reducing the perceived need to scratch.

Future randomized studies with larger cohorts, a longer follow-up, an assessment of potential impacts on skin, and daytime scratching behaviors are needed to confirm these results.

© 2025 Science X Network