A Comparison of Modelling Techniques used to Characterise Oxygen Uptake Kinetics During the on‐Transient of Exercise

Abstract

The on-transient V O 2 response shows three functionally distinct phases as previously described (Whipp & Ward, 1990; see Fig. 1). The first phase, lasting approximately 20 s, has been termed the 'cardio-dynamic' phase (Whipp et al. 1982) as the increase in V O 2 is due primarily to an increase in pulmonary blood flow (Cummin et al. 1986). During phase 2, V O 2 increases towards a steady-state (phase 3) value in a manner which is exponential in nature, and is temporally aligned with the kinetics of phosphocreatine breakdown (the surrogate of adenosine triphosphate production) and thus is believed to reflect muscle O 2 consumption Whipp et al. 1999). If the work rate is of sufficient intensity to induce a sustained increase in blood lactate concentration, then a steady state is delayed or not seen, and a slow component of V O 2 becomes evident (Casaburi et al. 1989). This slow component, which is of delayed onset, will produce an end-exercise ∆V O 2 /∆WR relationship greater than that predicted from the V O 2 -power output relationship for moderate-intensity exercise (Poole et al. 1988). The purpose of this investigation was to examine differences in parameter estimates derived from different mathematical models and fitting procedures used for the analysis of V O 2 kinetics following the onset of moderate-and heavy-intensity cycle exercise. Exponential models with one or more components were used to describe the dynamics of gas exchange. Estimates of the parameters of the response, specifically the phase 2 time constant (r), and the slow component of V O 2 , were compared, together with a statistical analysis of how well each model fitted the data. We

Regulation of V̇o2 kinetics by O2 delivery: insights from acute hypoxia and heavy-intensity priming exercise in young men

“…After exclusion of the initial 20 s of data from the model [which, while not necessarily reflecting the exact duration of phase I V O 2p in each subject, is most likely to avoid inclusion of data points from phase I V O2p in the fitting of phase II V O2p (25)], while still allowing TD to vary freely (to optimize accuracy of parameter estimates), V O 2p data were modeled to 4 min (240 s) of the step transition; this ensured that each subject had attained a V O2p steady state yet did not bias the model fit during the on-transient (4). HR data were modeled from exercise onset to the end of the 6-min exercise transition, with TD constrained to Ն0 s and baseline fixed as described above (i.e., 60-s average during the 75-to 15-s window prior to a transition).…”

Noninvasive estimation of microvascular O2 provision during exercise on-transients in healthy young males

“…After excluding the initial 20 s of data from the model, while still allowing TD to vary freely [to optimize accuracy of parameter estimates (17)], V O2p data were modeled to 4 min (240 s) of the step-transition; this ensured that each subject had attained a V O2p steady-state, without biasing the model fit during the on-transient (3,17). HR data were modeled from the first datum after a transition to the end of the 6-min exercise transition with TD constrained to Ն0 s. The model parameters were estimated by least-squares nonlinear regression (Origin, OriginLab, Northampton, MA), in which the best fit was defined by minimization of the residual sum of squares and minimal variation of residuals around the y-axis (y ϭ 0).…”

Modeling energy expenditure and oxygen consumption in human exposure models: accounting for fatigue and EPOC

“…These methods allow for parametrization of the oxygen deficit over a range of work intensities using data from the published literature. Explicit mathematical models of oxygen uptake kinetics are available, many of which use sums of exponentials to model the fast and slow components of oxygen deficit (Bearden and Moffatt, 2000;Bell et al, 2001;Stirling et al, 2005). Such models can be used to derive explicit expressions for the slow and fast oxygen deficits in terms of time constants and delays, but determining appropriate parametrization of such models over a wide range of intensities currently is very difficult.…”