MindfulWatch
Hao Tian, Chongguang Bi, Guoliang Xing, Roxane Raffin Chan, Linlin Tu
Proceedings of the ACM on Interactive Mobile Wearable and Ubiquitous Technologies September 11, 2017 DOI: 10.1145/3130922 (opens in new tab)
Study at a glance
AI-extracted from the abstract| Characteristics | Observational feasibility study Longitudinal Peer reviewed |
|---|---|
| Sample size | 11 |
| Population | Individuals during real-world meditation sessions |
| Duration | 36 real-world meditation sessions totaling 8.7 hours |
| Citations | 62 |
| Key findings | MindfulWatch, a smartwatch-based system using motion sensors to detect respiration-induced wrist rotation, measured respiratory timing with 70% of errors under 0.5 seconds across 36 meditation sessions (8.7 hours, 11 subjects), suggesting reliable real-time respiratory timing measurement. |
Abstract
With a wealth of scientifically proven health benefits, meditation was enjoyed by about 18 million people in the U.S. alone, as of 2012. Yet, there remains a stunning lack of convenient tools for promoting long-term and effective meditation practice. In this paper, we present MindfulWatch, a practical smartwatch-based sensing system that monitors respiration in real-time during meditation -- offering essential biosignals that can potentially be used to empower various future applications such as tracking changes in breathing pattern, offering real-time guidance, and providing an accurate bio-marker for meditation research. To this end, MindfulWatch is designed to be convenient for everyday use with no training required. Operating solely on a smartwatch, MindfulWatch can immediately reach the growing population of smartwatch users, making it ideal for longitudinal data collection for meditation studies. Specifically, it utilizes motion sensors to sense the subtle “micro” wrist rotation (0.01 rad/s) induced by respiration. To accurately capture breathing, we developed a novel self-adaptive model that tracks changes in both breathing pattern and meditation posture over time. MindfulWatch was evaluated based on data from 36 real-world meditation sessions (8.7 hours, 11 subjects). The results suggest that MindfulWatch offers reliable real-time respiratory timing measurement (70% errors under 0.5 seconds).