Monitoring Tai Chi Training in Older Adults with Sensors and Machine Learning
Author Information
Author(s): Corniani Giulia, Sapienza Stefano, Vergara-Diaz Gloria, Valerio Andrea, Vaziri Ashkan, Bonato Paolo, Wayne Peter
Primary Institution: Harvard Medical School, Spaulding Rehabilitation Hospital
Hypothesis
Can wearable sensors and machine learning effectively monitor Tai Chi training adherence and proficiency in older adults?
Conclusion
The study demonstrates that wearable sensors and machine learning can accurately assess Tai Chi movement adherence and proficiency.
Supporting Evidence
- The movement identification model achieved a high accuracy of 90.05%.
- Proficiency assessment models achieved a mean accuracy of 78.64%.
- Tai Chi is endorsed by health organizations for fall prevention.
Takeaway
This study shows that we can use special sensors to help older people practice Tai Chi better and safer, even at home.
Methodology
Participants performed six Tai Chi movements while wearing sensors that recorded their movements, which were then analyzed using machine learning algorithms.
Potential Biases
Potential bias in expert scoring of Tai Chi proficiency could affect the accuracy of the results.
Limitations
The study's findings may not generalize to all older adults, as participants were healthy and had varying Tai Chi experience.
Participant Demographics
32 healthy participants aged 60 to 85 years, including 20 females.
Digital Object Identifier (DOI)
Want to read the original?
Access the complete publication on the publisher's website