Estimating biological age on consumer-grade wearables usually relies on long-term signals such as heart rate variability, resting heart rate, sleep structure and activity volume, and then compares them with population baseline models.