SocialBit2: High-fidelity Mobile Sensing of Social Connection after Stroke

In the epidemic of social isolation affecting today’s society, stroke survivors are particularly vulnerable. After stroke, social isolation has adverse biological and psycho-social effects, which can be prevented. Therefore, increasing social connection has the potential to be a powerful strategy for enhancing stroke recovery outcomes. Developing effective interventions, however, necessitate accurate and high-fidelity measurement of social life changes after stroke, and knowledge of how these changes relate to patient characteristics and recovery trajectories. Traditional research methods rely on self-report surveys, which are subjective, retrospective, and cannot be completed in up to 45% of patients due to cognitive and language deficits. Therefore, there is gap in measurement of social connection that limits the design of interventions. In response, we developed SocialBit, an artificial intelligence algorithm that operates on a smartwatch and utilizes ambient audio features to detect minutes of social interaction without storing any raw audio data. It provides an objective, continuous, inclusive, and high-resolution picture of social life changes after stroke, which can inform decision-making around when and how to intervene in this at-risk population. We created and validated SocialBit specifically for stroke survivors, taking into account cognitive, physical, and language variations. Our preliminary data show that SocialBit measures an important aspect of patients’ social connection: time spent interacting. However, our studies to date focus on the context of the hospital. The goal now is to deploy SocialBit in patients’ natural, everyday environments, which will be the first study of objective social connection of stroke survivors. We propose a study of two hundred (N = 200) patients with ischemic stroke who will wear a SocialBit smartwatch for 3 months after hospitalization. Our central hypothesis is that there will be discernable trajectories of post-stroke social life related to stroke and contextual factors that predict cognitive and physical outcomes, which are ideal targets for future interventions. The study aims are to 1) Characterize patients’ social interaction trajectories for the first 3 months after ischemic stroke in relation to important stroke characteristics and contextual factors; 2) Examine the bidirectional associations between participants’ social interaction, cognitive, and physical functioning over the first 3 months following ischemic stroke; and 3) Through qualitative inquiry, determine the social interaction mechanisms that may be leveraged by SocialBit to improve stroke recovery. We have assembled a multidisciplinary team with expertise in stroke, artificial intelligence, mobile sensing, mixed methods, and social connectedness to execute this project. Achieving these aims will provide foundational insights into quantifying social connection for stroke survivors. Moreover, this study represents a significant advancement by employing a direct, continuous, and scalable in-vivo mobile sensing method for measuring social connection. The acquired data will set the stage for testing "pro- connection technology" to improve quality of life for stroke survivors. Project Number: 1R01NS142764-01A1 | Fiscal Year: 2026 | NIH Institute/Center: National Institute of Neurological Disorders and Stroke (NINDS) | Principal Investigator: Amar Dhand | Institution: BRIGHAM AND WOMEN'S HOSPITAL, BOSTON, MA | Award Amount: $716,816 | Activity Code: R01 | Study Section: Biobehavioral Medicine and Health Outcomes Study Section[BMHO] View on NIH RePORTER: https://reporter.nih.gov/project-details/11295617