Description

For both a hip- and back-worn ActiGraph monitor, three regression equations were created for predicting energy expenditure using either count-based or acceleration-based input features (six total models). Root Mean Square Error (RMSE) was 0.88-1.02 kcal/min. Cut-points to classify activity intensity from each metric were determined with a sensitivity of 0.89-1.00 and a specificity of 0.79-0.99.

Development/Validation

Sample: 43 (11 males) adults with class 2 or 3 obesity, 42.6±9.2 years of age

Setting: Laboratory

Activities: Stationary, treadmill walking

Criterion: Oxycon Pro (VO₂)

Accelerometer(s): ActiGraph GT9X on right hip and lower back

Validation approach: Leave-one-out cross-validation

Phase Designation

(What's this?)

This model is in Phase 1.

Instructions

ActiGraph counts were used to calculate the ‘activity counts’ (AC) metric as the resultant vector over a 5-s epoch (this is not further defined but is typically the square root of the sum of the squared counts in each axis, i.e., vector magnitude). Euclidean Norm Minus One (ENMO) and Mean Amplitude Deviation (MAD) were calculated for each 5-s epoch using the R package GGIR. Six regression equations were developed and are provided below- one for each of the three metrics at each of the two monitor locations (back, hip). Cut-points are also provided for determining activity intensity for each 5-s epoch from activity counts, MAD, or ENMO for each monitor wear location.

Source Information

Diniz-Sousa, F., Veras, L., Ribeiro, J. C., Boppre, G., Devezas, V., Santos-Sousa, H., ... & Fonseca, H. (2020). Accelerometry calibration in people with class II-III obesity: energy expenditure prediction and physical activity intensity identification. Gait & Posture, 76, 104-109. https://doi.org/10.1016/j.gaitpost.2019.11.008 Link to Paper

Corresponding author: Florêncio Diniz-Sousa, joseflorenciosousa@gmail.com

Contact

Kimberly Clevenger at accelerometerrepository@gmail.com