A preliminary study will present findings at the American Stroke Association's International Stroke Conference 2026 regarding a smartwatch application designed to measure social interactions among hospitalized stroke survivors. This technology may contribute to new treatments aimed at preserving or enhancing cognition, social engagement, and quality of life post-stroke.
SocialBit App Development
Researchers developed SocialBit, a machine learning app compatible with Android smartwatches. The app is designed to identify social interactions in individuals with and without neurological conditions, distinguishing it from other devices focused solely on people without disabilities. SocialBit is currently restricted to research projects.
The American Stroke Association indicates that speech and language impairments, such as dysarthria and aphasia, significantly impact the social lives of stroke survivors. However, research suggests that social engagement is crucial for maximizing post-stroke recovery.
Previous research by study lead author Dr. Amar Dhand highlighted that socially isolated stroke survivors or those with smaller social circles experience poorer physical outcomes three and six months after a stroke.
Dr. Dhand stated that tracking human engagement is vital, and identifying social isolation in real-world settings could facilitate intervention by notifying patients, family, caregivers, and healthcare professionals.
Study Methodology
The study recruited 153 adults hospitalized for an ischemic stroke. Participants wore a smartwatch equipped with the SocialBit app daily from 9 a.m. to 5 p.m. for up to eight days, including time spent in rehabilitation hospitals. The app recorded socialization time based on acoustic patterns of conversation. Simultaneously, research team members livestreamed video of participants and logged minute-by-minute social interactions for comparison.
Key Findings
- SocialBit demonstrated 94% accuracy in recognizing social interactions when compared to human observers.
- In patients with aphasia, the app maintained an accuracy of 93%.
- The app's performance remained consistent despite environmental factors such as TV noise, side conversations, different hospital settings (rehabilitation unit versus general hospital), and across various Android smartwatch models.
- A correlation was observed between stroke severity and social interaction: participants with more severe strokes showed approximately a 1% decrease in total social interaction minutes for each 1-point increase on the NIH Stroke Scale.
Dr. Dhand expressed surprise regarding the app's effectiveness in individuals with aphasia.
He noted that capturing sounds instead of words, a feature implemented for privacy, proved beneficial for those with limited language skills.
He suggested SocialBit could also support recovery from other brain injuries and aid therapies like speech, occupational, and exercise therapy.
Future Implications and Limitations
Future research could leverage SocialBit to assess the risk of social isolation during hospitalization and after discharge, and to investigate its relationship with mental health changes post-stroke, such as depression. The app may also be tested for use in other brain injuries and in healthy aging to support brain health.
A limitation of the study was that detailed evaluations of social interactions were conducted exclusively within hospital or rehabilitation settings.
Dr. Cheryl Bushnell, who was not involved in the study, commented on the research's ability to capture social interactions. She raised questions regarding whether the app can differentiate between conversations with hospital personnel and non-hospital individuals, which could provide further insights into factors influencing social interaction and the quality of hospital care.