From Movement to METs: A Validation of ActTrust® for Energy Expenditure Estimation and Physical Activity Classification in Young Adults

Abstract Estimating physical activity (PA) levels is a challenging and expensive task. An alternative could be the use of actigraphy devices to estimate PA. This has been previously done to a number of devices, including ActiGraph® GT3X+. In this study, we validated ActTrust® against the widely used GT3X+ and compared activity counts to metabolic equivalents (METs) derived from indirect calorimetry during treadmill walking and running. Fifty-six young adults (34 men, 22 women) participated in controlled effort exercises including light, moderate, vigorous, and very vigorous activity intensities. We developed a linear model to estimate energy expenditure (EE) from movement count of combinations of devices placed at hip or wrist. We then estimated cut-off points for each intensity range. Our results showed correlations between treadmill speed and both METs (r = 0.95, p < 0.05) and movement counts from both GT3X+ and ActTrust devices placed either on the hip (r = 0.94, p < 0.05; r = 0.93, p < 0.05) or on the wrist (r = 0.88, p < 0.05; r = 0.88, p < 0.05), respectively. Our proposed model performed well with balanced accuracies above 0.77 for all intensity ranges and over 0.9 for light and moderate activity. This is the first study to model estimate and validate PA intensity thresholds on ActTrust® devices. Our findings support the use of ActTrust® devices as simple, cost-effective tool for 24-hour assessments of EE. Competing Interest Statement The authors have declared no competing interest. Footnotes First revision (v2): This revised version addresses comments from the first round of peer review at PLOS ONE. Changes relative to the preprint are as follows. Methodological transparency was enhanced, including justification for epoch selection, synchronisation details between devices and calorimetry, and rationale for the square root transformation of METs and activity counts. The fixed progressive protocol order was acknowledged as a potential limitation and added to the limitations section. Regression assumptions were verified through residual and Q-Q plots, included as supplementary material. The discussion was expanded to include direct comparison of cut-points and AUC values with Santos-Lozano et al. (2013) and Sasaki et al. (2011), and claims regarding clinical applicability were appropriately softened. The conclusion was revised to reflect feasibility rather than confirmed clinical utility. The introduction was expanded to articulate the specific advantages of ActTrust over existing devices, including cost, battery life, and integrated sensors. The limitations section was expanded to address generalisability to older adults and clinical populations, and to acknowledge potential order effects from the fixed progressive protocol. Sex-stratified confusion matrices and AUC values, Bland-Altman plots, and regression diagnostics were added as supplementary material. Cut-point estimates now include 95% confidence intervals, and device-specific regression coefficients are reported with 95% CI. Repository documentation was updated to include a revised README.md, data dictionary, commented analysis scripts, and a requirements file with version numbers for all software dependencies. Data availability and funding statements were corrected and matched between the submission form and manuscript. Second revision (v3): Figure 1 has been updated to encode sex via point shape (squares: male, circles: female) across both panels, in response to a reviewer request to distinguish sex in the main figures and provide insight into the clustering observed at 7 km h^-1.