Detection of a lactate threshold during incremental exercise?

An important question in the study of the exercise response is the real or imaginary nature of the anaerobic threshold, and mathematical modeling techniques have been invoked to assist in resolving this issue. Two opposing views with competing data models recently published in this journal are criticized. One view suggests a segmented model with a discontinuous first derivative at the threshold. The other suggests a continuous model over the whole work load range, implying the anaerobic threshold to be imaginary. However, neither group of authors has devoted proper rigorous attention to the models they use. Had this been done, some of the divergence of opinion may have been avoided. Ideal data from an alternate segmented model that has a continuous first derivative at the threshold are considered for comparative purposes. This suggests that the log-log transformation method may well lead to improved detection of a threshold when one exists, although the estimates of the threshold value obtained are unreliable. Modeling methodology is a useful approach to the resolution of scientific issues, but there exist fundamental implications and alternatives that must be fully recognized.