Evaluation of moderate to vigorous boundary under free-living conditions with accelerometry

Daily moderate-to-vigorous physical activity (MVPA) evaluations based on absolute cutoffs from accelerometer data are inappropriate for characterising physical capacities, given inter-individual variability. We developed a new method that evaluates an individual's MVPA boundary without difficult exhaustion tests by estimating the critical intensity in daily activities using accelerometers. Thirty-five participants with varying fitness levels wore a 3-axis accelerometer for 3-6 weeks. They performed a sub-maximal incremental walking test on a 15%-inclined treadmill, where the speed at 4 mmol·L^-1 (S_{4 mM}) was computed. A two-parameter critical intensity model was fitted to Dynamic Body Acceleration (DBA) records using quantile regression. The critical DBA (DBA_C) was computed from this model and compared with the traditional absolute cutoff for MVPA. DBA_C was a strong predictor of S_{4 mM} (r² = 0.788) and, compared with individual characterisation, the traditional MVPA analysis resulted in a mean absolute error of 10:37 min/day (118%) and 12:07 min/day (26%) for time spent in vigorous and moderate activity, respectively. In conclusion, the individual DBA_C computed from daily life activities without controlled laboratory tests accurately represents the severe intensity domain boundary. This method enhances the current absolute cutoff MVPA classification, which fails to accurately represent the time spent in intensity domains with individual physiological meanings.