OpenMHC: Accelerating the Science of Wearable Foundation Models

Summary

OpenMHC (OpenMyHeartCounts) is the largest open-access wearable health dataset published to date, released together with open-source implementations of recent wearable foundation models and a unified evaluation benchmark. The dataset is derived from over a decade of passive and active data collection through the Stanford My Heart Counts study app, and contains more than 60 million hours of wearable data spanning 19 sensor channels — including heart rate, step count, sleep, and workouts — from 11,894 consenting participants. Each participant record can be linked to up to 169 additional variables covering health status, lifestyle, mood, and behavior, which makes the corpus unusually rich on the label side relative to other large wearable datasets.