Beyond a Single Time Constant: Evaluating the Cross- Protocol Transferability of τW′ in Cyclists

The quantification of W′ reconstitution kinetics (τW′) is central to W′ balance (W′BAL) modelling, yet whether an individualized τW′ can be transferred across different intermittent exercise structures remains unclear. This study examined the cross-protocol transferability of a reference τW′ derived from a 60 s - 30 s high intensity interval training (HIIT) protocol. Six trained male cyclists completed baseline testing and four exhaustive HIIT protocols: 60 s–30 s HIIT, short HIIT (S-HIIT), long HIIT (L-HIIT), and decreasing power output HIIT (DPO-HIIT). The τW′ required to yield W′BAL = 0 kJ at task failure during 60 s - 30 s HIIT was defined as Ref-τW′ and subsequently applied to the other protocols. End-exercise W′BAL differed across protocols (p<0.001). Ref-τW′ produced the smallest bias and root mean square error during S-HIIT (3.3 and 4.3 kJ), whereas larger errors were observed during L-HIIT (6.5 and 7.0 kJ) and DPO-HIIT (6.8 and 7.4 kJ). Protocol-specific τW′ values also differed (p<0.001), with no differences between 60 s - 30 s and S-HIIT or between L-HIIT and DPO-HIIT. These findings indicate that τW′ is protocol dependent and that a single individualised value may not accurately transfer across distinct HIIT structures. Transferability therefore appears greater between protocols with similar work and recovery characteristics.