Reanalysis of Hickson et al. (1977) weekly VO2max data from eight participants with mixed-effects models showed curvilinear trajectories with diminishing returns rather than linear increases, and systematic interindividual variability in response persisted even under a normalized, maximal sustainable training stimulus.
Key Findings
Results
Curvilinear mixed-effects models with random slopes described the weekly VO2max trajectories better than linear models, capturing a plateauing effect.
Data were weekly VO2max from eight participants (31.5 ± 7.3 yr, 182.0 ± 8.6 cm, 79.2 ± 17.1 kg) over a 10-wk high-intensity training program.
Candidate mixed-effects models were linear, quadratic, cubic, logarithmic, rectangular hyperbolic, and monoexponential.
Model selection used AICc, preferring the most parsimonious model when AICc differences were small.
The abstract states curvilinear models captured plateauing effects not represented by linear models.
Discussion
The reanalysis challenges the original interpretation that VO2max rose linearly across the 10-week strenuous program.
Adaptations were better described by curvilinear trajectories with diminishing returns.
The pattern is described as consistent with VO2max plateauing as individuals approach a 'genetic ceiling'.
Results
A quadratic model with a random slope fit significantly better than its fixed-slope equivalent, indicating systematic interindividual variability in VO2max trajectories.
Likelihood-ratio comparison gave χ2 = 26.1, P < 0.001.
The estimated slope standard deviation was 0.029 L·min-1·wk-1 (95% CI: 0.010 to 0.042).
Variability in response remained after accounting for residual variability.
Discussion
Heterogeneity in adaptive response persisted even though the training stimulus was normalized to a maximal sustainable effort.
The training program was strenuous and based on a maximal sustainable effort.
Despite this normalized stimulus, individual response magnitude differed systematically between participants.
Conclusions
The study argues that legacy datasets can be revisited with contemporary analytical approaches to better characterize human adaptation to exercise training.
The analysis used weekly individual-level data and mixed-effects modeling rather than group-level analysis.
The findings are presented as relevant to individualized exercise prescription.
What This Means
This study took a classic 1977 exercise experiment, in which eight people did very hard endurance training for 10 weeks, and re-examined the weekly fitness measurements (maximal oxygen uptake, a key measure of aerobic fitness) using modern statistical methods. The original study described fitness as rising in a roughly straight line. The reanalysis found that curved patterns fit better: gains were larger early on and then leveled off, consistent with diminishing returns as people approach their personal upper limit.
The study also found that people differed systematically in how quickly their fitness changed over time. This variation was beyond random measurement noise. It remained even though the training was scaled to each person's maximal sustainable effort, so a standardized hard stimulus did not make everyone respond the same way.
This research suggests that fitness gains from hard training may not continue at a constant rate, and that individuals can follow different trajectories. The sample was only eight participants, so the findings are best viewed as a demonstration of how old datasets can be re-analyzed rather than as definitive evidence. The work supports the idea that training programs may benefit from being tailored to individuals.
Check Your Own Numbers
Upload your bloodwork. We'll cross-reference your results against this study and 4,700 others.
Bossi A. (2026). Adaptive response heterogeneity induced by a strenuous program of endurance exercise: a reanalysis of Hickson et al. (1977).. Journal of applied physiology (Bethesda, Md. : 1985). https://doi.org/10.1152/japplphysiol.90248.2026