A smartwatch application that uses machine learning to detect social interactions through environmental sounds may help improve treatment and recovery for stroke survivors, according to preliminary research to be presented at the American Stroke Association’s International Stroke Conference 2026. The app, called SocialBit, was found to be 94% as accurate as human observers in recognizing social interactions among hospitalized patients, maintaining 93% accuracy even in patients with aphasia, a language disorder common after stroke.
Researchers developed SocialBit to address a critical gap in post-stroke care. Social interaction is known to support brain health and recovery after neurological injury, yet tools to track such engagement have largely focused on people without disabilities. "My previous research has demonstrated that stroke survivors who are socially isolated or have a smaller circle of friends and family have worse physical outcomes at 3 and 6 months after a stroke," said study lead author Amar Dhand, M.D., D.Phil., an associate professor of neurology at Mass General Brigham. The app captures acoustic patterns of human speech to quantify social engagement without recording specific words, protecting privacy while being effective for those with language difficulties.
The study involved 153 adults hospitalized for ischemic stroke. Participants wore an Android smartwatch with the SocialBit app in their rooms between 9 a.m. and 5 p.m. daily for up to eight days. The app's minute-by-minute logging of socialization time was compared against ratings from research team members watching a livestream video. The technology performed consistently despite background noise from televisions, side conversations, different hospital environments, and across various smartwatch models. The findings also revealed that participants with more severe strokes, as measured by the NIH Stroke Scale, had less social interaction.
The implications of this technology are significant for stroke rehabilitation. By identifying social isolation in real-time, the app could alert patients, families, caregivers, and healthcare professionals, enabling timely interventions. "Tracking human engagement is crucial, and social isolation can now be identified in real-world situations," Dhand noted. He suggested the app could support existing therapies like speech, occupational, and exercise therapy. Future research may use SocialBit to measure the risk of social isolation during and after hospitalization and explore its relationship with post-stroke depression and other mental health changes. Researchers also envision applications for other brain injuries and healthy aging to maintain cognitive health.
Cheryl Bushnell, M.D., M.H.S., FAHA, chair of the American Heart Association Stroke Council, who was not involved in the study, called the research "fascinating" and noted multiple potential applications. "There are multiple interesting ways this app could be used in future studies, including measures of quality of hospital care and social interactions at rehab facilities and nursing homes," said Bushnell, who also chaired the writing group for the Association’s 2024 Guideline for the Primary Prevention of Stroke. She highlighted the importance of distinguishing between interactions with hospital personnel and non-hospital visitors, as factors like care team size or family proximity could influence social engagement metrics.
The study has limitations, including its confinement to hospital and rehabilitation settings. SocialBit is currently available only for research purposes. The research abstract is part of the American Stroke Association International Stroke Conference 2026 Online Program Planner. Findings presented at the association's meetings are considered preliminary until published as full manuscripts in peer-reviewed journals. The American Stroke Association provides extensive resources on stroke, available through its Stroke Hub. According to the American Heart Association’s 2026 Heart Disease and Stroke Statistics, stroke is now the fourth leading cause of death in the United States, underscoring the need for innovative recovery tools like SocialBit.
