Smartwatch Detects Social Life in Stroke Survivors

Stroke survivors facing speech difficulties now have a powerful ally in their corner: a smartwatch that monitors their social connections and could transform how they heal.

Researchers at Mass General Brigham in Boston developed SocialBit, a machine learning app that tracks social interactions through sounds in hospital rooms. The app proved 94% accurate at detecting when stroke patients engaged with visitors, caregivers, and loved ones, matching the precision of trained human observers.

The breakthrough matters because socializing is one of the best ways to recover from a stroke. Yet speech and language changes make social connection incredibly difficult for many survivors, pushing them toward isolation at the exact moment they need human contact most.

Dr. Amar Dhand, who led the study, found in previous research that stroke survivors with smaller social circles had worse physical outcomes three and six months after their stroke. He created SocialBit to identify who's at risk before isolation takes its toll.

The app tested remarkably well with 153 stroke patients during their hospital stays. Participants wore Android smartwatches from 9 a.m. to 5 p.m. daily for up to eight days while the app listened to acoustic patterns indicating conversation.

The Ripple Effect

What makes SocialBit special is its accuracy for people with aphasia, a language disorder that affects many stroke survivors. The app maintained 93% accuracy even when patients couldn't speak clearly or form words, because it detects sounds of interaction rather than specific words.

This privacy-protecting feature became an unexpected advantage. The app worked consistently despite TV noise, side conversations, different hospital environments, and across various smartwatch models.

The data revealed a concerning pattern: patients with more severe strokes had about 1% less social interaction for each point increase on the stroke severity scale. That means the people who need connection most are getting it least.

The technology could alert patients, families, caregivers, and medical teams when someone is becoming isolated. It might support speech therapy, occupational therapy, and exercise programs by measuring real-world social engagement beyond the therapy room.

Future research will explore using SocialBit after patients leave the hospital, tracking connections to depression and mental health changes that often follow strokes. The app could also help people with other brain injuries and support healthy aging to preserve brain health over time.

Right now, SocialBit is only available for research projects, but it represents a new way to measure something doctors have long known matters: human connection heals.