Importance of Biosignal Tracking with Wearable Devices
Biosignal tracking through wearable devices is important for keeping an eye on wellness and catching potential health issues early. With the use of wearable devices, it’s easy to monitor various biosignals and track health status without interrupting daily activities.
Challenges in Developing Bio-Markers
Despite the popularity of wearable devices and digital biomarkers, the lack of well-curated data with medical labels makes it difficult to develop new biomarkers for common health conditions. Medical datasets are often small when compared to other areas, posing a hurdle in creating neural network models for biosignals.
Innovative Solution: Self-Supervised Learning with Data from Apple Heart and Movement Study
To overcome this challenge, a new approach has been implemented using self-supervised learning with large, unlabeled sensor data from the Apple Heart and Movement Study. The data collected from approximately 141,000 participants over a period of three years has been used to train foundation models for two common biosignals: photoplethysmography (PPG) and electrocardiogram (ECG) recorded on Apple Watch.
Positive Results: Encoded Information and Potential for Future Enhancements
The self-supervised learning framework has shown promising results, as the pre-trained foundation models readily encode information regarding participants’ demographics and health conditions. This study is the first to build foundation models using large-scale PPG and ECG data collected via wearable consumer devices, setting it apart from prior works which commonly used smaller-sized datasets from clinical and experimental settings.
Implications for Future Health Monitoring
The use of PPG and ECG foundation models has the potential to improve future wearable devices by reducing the reliance on labeled data. Ultimately, this advancement can help users enhance their health and well-being.