Related Experiment Video
Updated: Sep 18, 2025
08:45
Assessment of Physical Activity Intensity with Accelerometers and Oxygen Consumption
Published on: June 20, 2025
194
Developing and comparing a new BMI inclusive energy expenditure algorithm on wrist-worn wearables
Boyang Wei1,2, Christopher Romano1, Mahdi Pedram3
1Department of Preventive Medicine, Northwestern University, Chicago, 60611, USA.
Scientific Reports
|June 19, 2025
Summary
This study developed a new algorithm using smartwatch data to accurately estimate energy expenditure (EE) in individuals with obesity. This method offers a more inclusive and reliable alternative to traditional actigraphy for EE measurement.
Area of Science:
- Biomedical Engineering
- Wearable Technology
- Human Physiology
Background:
- Accurate estimation of energy expenditure (EE) is vital for understanding human behavior and energy balance.
- Actigraphy, a common method for EE estimation using wrist-worn inertial measurement units (IMU), has known accuracy limitations, especially in individuals with obesity.
- Existing EE estimation algorithms are often validated in non-obese populations, limiting their applicability.
Purpose of the Study:
- To develop and validate a novel algorithm for estimating EE from commercial smartwatch sensor data.
- To compare the accuracy of the new algorithm against actigraphy-based estimates in people with obesity.
- To assess the reliability and inclusivity of smartwatch-based EE measures for diverse populations.
Main Methods:
- Developed a machine learning model utilizing accelerometer and gyroscope data from a commercial smartwatch (Fossil Sport).
- Validated the algorithm against a metabolic cart in an in-lab study with 27 participants performing varied activities.
- Assessed algorithm performance in a 2-day free-living study with 25 participants with obesity.
Main Results:
- The machine learning model demonstrated lower root mean square error (0.28-0.32) compared to a metabolic cart across different sliding windows.
- In free-living conditions, the algorithm's EE estimates were within 1 SD of the best actigraphy-based estimates for 95.03% of minutes.
- The proposed method showed superior accuracy compared to 11 other algorithms, particularly those validated in non-obese cohorts.
Conclusions:
- Commercial smartwatches, when used with the developed algorithm, can provide accurate and reliable energy expenditure estimates.
- This approach offers a more inclusive and accessible method for EE measurement, especially for individuals with obesity.
- The findings suggest a significant advancement in wearable technology for personalized health monitoring and research.
