Frequency domain analysis to extract dynamic response characteristics for oxygen uptake during transitions to moderate- and heavy-intensity exercises.

At the onset of an exercise transition, exponential modeling to calculate a time constant (τ) is the conventional method to analyze pulmonary oxygen uptake (V̇O2p) kinetics for moderate and heavy exercises. A new frequency domain analysis technique, mean normalized gain (MNG), has been used to analyze V̇O2p kinetics during moderate exercise, but has not been evaluated for its ability to detect differences in kinetics between moderate and heavy exercises. This study tested the hypothesis that MNG would detect smaller amplitude V̇O2p responses in the heavy-exercise domain compared with moderate-exercise domain. Eight young healthy adults (3 female; age: 27 ± 6 yr; peak V̇O2p: 43 ± 6 mL·min-1·kg-1; means ± SD) performed three bouts of pseudorandom binary sequence (PRBS) exercise for frequency analysis, with the work rate (WR) changing between 25 W and 90% ventilatory threshold (VT; L → MPRBS), 25 W and 50% of the difference between VT and peak V̇O2p (Δ50%; L → HPRBS), and VT to Δ50% (VT → HPRBS). Step exercise tests with equivalent changes in WR to the PRBS tests were performed to facilitate the comparison between MNG and τ. MNG was the highest for L → MPRBS (59 ± 7%), then L → HPRBS (52 ± 6%), and the lowest for VT → HPRBS (38 ± 7%, F(2,14) = 129.755, P < 0.001) exercise conditions indicating slower kinetics with increasing exercise intensity that correlated strongly in repeated measures with τ from step transitions (rrm = -0.893). These results indicate that frequency domain analysis and MNG reliably detect differences in V̇O2p kinetics observed across exercise intensity domains.NEW & NOTEWORTHY Mean normalized gain is able to detect differences in V̇O2p kinetics between moderate-, heavy-, and heavy-intensity exercises from a raised WR within the same individuals. This new method of kinetic analysis may be advantageous compared with conventional V̇O2p curve fitting, as it is less sensitive to breath-by-breath noise, it can provide useful information from a single exercise testing session, and it can be applied to nonconstant work rate exercise situations.