Validation of Accelerometer-Based Estimations of Energy Expenditure During High-Intensity Interval Training

Accelerometers are used to assess free-living physical activity (PA) and energy expenditure (EE). Energy expenditure estimation algorithms have been calibrated using steady-state exercise. However, most free-living PA is not steady-state. Objective: The purpose of this study was to discern the differences between criterion-measured and accelerometer-estimated EE (kCals) during a non-steady-state High-Intensity Interval Training (HIIT) session. Methods: Recreationally active adults (N=29, 18-30 years) completed one of two HIIT protocols. Each participant wore ActiGraph GT3X+ accelerometers on the right hip and non-dominant wrist while EE was measured using portable indirect calorimetry. Data analysis was conducted using custom R scripts and bias [95% CIs] to determine significant differences between indirect calorimetry and EE estimates using previously developed algorithms. Results: All accelerometer algorithms underestimated EE during recovery intervals (range; -4.31 to -6.55 kCals) and overestimated EE during work intervals (0.57 to 5.70 kcals). Over the whole HIIT session, only the Hildebrand wrist method was not significantly different from the criterion measured EE. Conclusion: Current ActiGraph EE estimations based on steady-state activities underestimate EE during recovery periods of treadmill HIIT sessions. Future studies should investigate accelerometer signals immediately after high-intensity bouts to more accurately predict EE of the subsequent recovery period. KEYWORDS: ActiGraph; Accelerometer; HIIT; Indirect calorimetry; EPOC; Energy expenditure; Non-steady state; Calories