Characteristics of children's physical activity during active play.

BACKGROUND: Emerging evidence suggests that oxygen consumption (VO2) for self-paced physical activity (PA) is underestimated when applying laboratory-based treadmill (TM)-derived regression equations. This study examines the accelerometer (ACC) characteristics for paced TM PA and self-paced children's PA to ascertain if the predictable regulated patterns of paced PA are implicated in the inferior estimates of VO2.

METHODS: Children's (9.3±1.2 years) (N.=21) VO2 (portable oxygen analyzer) and PA (accelerometry) were measured for paced treadmill PA (4-10 km/h). Active playing of children's games in a camp setting was used for self-paced PA. Treadmill and self-paced PA were compared by linear regression and Bland-Altman analysis. Relative contribution of each axes (% axis difference) for paced and self-paced PA were assessed (N.=21).

RESULTS: The VO2 responses during paced treadmill exercise was linear with ACC quantified PA (for vertical axis r=0.95±0.03 and for vector magnitude [VM] r=0.95±0.05, P>0.05). During self-paced PA, the VO2 responses for ACC quantified PA were not linear (for vertical axis r=0.20±0.11 and for VM r=0.25±0.09) over the same range of ACC PA (0-1500 counts/10 s). VO2 estimates for self-paced PA (using TM-derived equations) were underestimated (P<0.05) across the range of intensities, which increased as the intensity of PA increased (>6 METs). Comparing paced versus self-paced PA the % axis dominance (i.e., difference between the highest and lowest axis) contribution to PA was 41±14% for paced and 3±2% for self-paced PA (P<0.05).

CONCLUSIONS: This study reveals that the inferior estimates of VO2 for self-paced PA (using-derived linear equations) is attributable to the presence of a dominant axis with paced TM PA, which inflates the calculation of VM and the predicted VO2 for self-paced PA where no % axis difference exists.