Description

Five prediction models for energy expenditure were developed using ActiGraph GT1M monitors worn on the hip, wrist, and ankle while adolescents performed a variety of activities in a room calorimeter. Root Mean Square Error (RMSE) was 143.0 (hip + wrist + ankle), 155.8 (hip + wrist), 163.6 (hip), 180.6 (wrist), and 151.2 (ankle) kcal/day. When used to classify physical activity intensity, classification accuracy was 63 to 75% (sedentary), 68 to 79% (light), and 34 to 80% (moderate-to-vigorous) depending on the prediction model used.

Development/Validation

Sample: 76 (29 males) healthy adolescents, 10-17 years of age

Setting: Room calorimeter

Activities: Activities of daily living, cycling, stationary, stepping, treadmill running, walking

Criterion: Room calorimeter (VO₂)

Accelerometer(s): ActiGraph GT1M on dominant hip, wrist, and/or ankle

Validation approach: Leave-one-out cross-validation

Phase Designation

(What's this?)

This model is in Phase 0 and 2.

TEE coefficients and equations

PAEE coefficients and equations

METs coefficients and equations

Source Information

Reference:

Choi, L., Chen, K. Y., Acra, S. A., & Buchowski, M. S. (2010). Distributed lag and spline modeling for predicting energy expenditure from accelerometry in youth. Journal of Applied Physiology, 108(2), 314-327. https://doi.org/ 10.1152/japplphysiol.00374.2009 Link to Paper

Corresponding author: Leena Choi, leena.choi@vanderbilt.edu

Contact

Kimberly Clevenger at accelerometerrepository@gmail.com