UCLA

Purpose: Accelerometry data from 2 physical activity monitors (PAM) paired with 3 physiological measures (heart rate, oxygen uptake, and energy expenditure) across 3 exercise modes (constant treadmill load, stepping, and incremental treadmill exercise).

Methods: 10 males performed 9 intensities each of steady-state treadmill and stepping exercise with wrist- and ankle-mounted PAMs. Last 2 minutes of each exercise was defined as steady-state, where energy expenditure (EE), derived by indirect calorimetry, and vector magnitude units (VMU), derived from accelerometer vectors, were averaged. Seven combinations of PAM configuration were compared. 2-factor ANOVA determined the effect of speed, grade, and their interaction for treadmill exercise, and the effect of step height, step rate, and their interaction for stepping exercise. 10 different, equally healthy subjects performed 3 incremental treadmill exercise tests with a chest-mounted PAM. Slopes from HR/V ̇O_2 and HR/VMU were compared, with repeatability assessed by intraclass correlation coefficients (ICC).

Results: Mean EE (Kcal/min) for all steady-state exercises ranged between 3 and 18 kcal/min 2-factor ANOVA revealed no effect of grade on all PAM configurations during treadmill exercise. ICC for HR/V ̇O_2 slopes = 0.80 ICC for HR/VMU slopes = 0.31

Discussion: PAM output was not influenced by grade for any configuration, which can be problematic for predicting EE. Slopes and intercepts of HR/V ̇O_2 was deemed more repeatable across 3 incremental exercise tests than HR/VMU.