What a training stimulus actually is
A training stimulus is not one thing. A bout of contraction imposes mechanical tension on the fibre, draws down stored energy, floods the cell with calcium and shifts the concentration of dozens of metabolites — and the muscle has sensors for all of it. What arrives at the cell is a package of mechanical, metabolic, neuronal and hormonal signals, and adaptation is the muscle’s reading of that package.
The consequence that matters is that skeletal muscle is a plastic tissue. Its size, its force, its endurance and its contractile speed are all revisable, and the reviews describe them as being revised continuously in response to the functional demands placed on the tissue. Nothing about a muscle’s current state is settled.
This is also why the field studies the signal rather than the exercise. Two sessions that look nothing alike from the outside can deliver overlapping signals inside the fibre, and two that look similar can fail to.
One session is a response. Many sessions are an adaptation.
The most useful idea in exercise physiology is the split between what one session does and what many sessions do. A bout of exercise produces a large but transient molecular response: signalling cascades activate, transcription of particular genes rises, the balance of protein synthesis and degradation shifts — and over the following hours most of it subsides. The muscle returns close to where it started.
Training adaptation is what accumulates when that transient response is triggered repeatedly. Each session adds to the abundance of the proteins the previous responses transcribed, and the durable phenotype — more mitochondria, more contractile protein, more capillaries — is the summed residue of many incomplete returns to baseline.
That is the mechanistic reason a programme is not the same object as a session, and it is why the literature treats “molecular responses to acute exercise” and “adaptations to exercise training” as two distinct things to measure rather than one.
Adaptation is specific to the stimulus that caused it
Adaptation is specific. The reviews treat endurance and resistance exercise as divergent stimuli — one biased toward oxidative adaptation, the other toward hypertrophy — because the signals they impose on the fibre differ: mechanical tension routes toward the protein-synthesis machinery, energetic disturbance routes toward mitochondrial building. The separation is a matter of emphasis rather than a clean division, since most sessions activate both, but it is why the field pairs a stimulus with the outcome it produces.
Intensity is its own axis of specificity. In work-matched comparisons within the same individual, high-intensity interval work produces greater mitochondrial adaptation than moderate continuous work, and sprint interval training raises mitochondrial content to a similar extent as continuous training despite substantially less total exercise. Intensity, not merely time spent, is carrying part of the effect.
The same review is direct about what has not been established. Far less evidence exists on how intensity mediates capillary density, maximal stroke volume, cardiac output and blood volume, and the interactions between intensity, duration and frequency have not been thoroughly explored — so the intensity story is well supported for mitochondria and thin everywhere else.
What actually remodels: mitochondria and capillaries
The two structural changes best documented in human muscle are mitochondrial and capillary growth, and a 2025 systematic review and meta-regression pooled 353 studies and 5,973 participants to size them. After adjustment for covariates, mitochondrial content rose by a similar amount across training types — 23 ± 5% with endurance training, 27 ± 5% with high-intensity interval training and 27 ± 7% with sprint interval training — with no significant difference between them.
Capillaries followed a similar but not identical pattern. Capillaries per fibre increased comparably across types (endurance 15 ± 3%, high-intensity interval 13 ± 4%, sprint interval 10 ± 11%), while capillaries per square millimetre rose only with endurance training (13 ± 3%) and high-intensity interval training (7 ± 4%), the larger gains attaching to continuous work. VO₂max improved similarly across all three.
Efficiency is where the training types separate. Per hour of exercise performed, sprint interval training was estimated at roughly 2.3 times the effect of high-intensity interval training and 3.9 times that of endurance training for mitochondrial content. The pooled data also report larger changes at higher weekly session counts, larger relative changes in those who began less fit, and most of the change occurring within roughly the first four weeks in untrained participants.
Does overload require the load to go up?
Progressive overload is the field’s oldest organising principle: the stimulus must keep increasing for adaptation to keep occurring. Whether the load itself has to be the thing that increases is a separate question, and it has been tested directly. A randomised trial assigned 43 resistance-trained participants to eight weeks of either load progression — adding weight while holding the repetition range — or repetition progression, holding the load constant and adding repetitions.
The two strategies produced close to the same result. Rectus femoris growth modestly favoured the repetition group (2.8 mm summed across measurement sites) and maximal strength slightly favoured the load group (2.0 kg), while every remaining outcome differed by a trivial margin: under a millimetre of muscle thickness elsewhere, under 1% for endurance, a tenth of a centimetre of jump height, a tenth of a kilogram of leg lean mass. The authors concluded that both progressions are viable routes to muscular adaptation across a block of that length.
It is one trial, of eight weeks, in trained participants, so it constrains the principle rather than overturning it. But it points at what the muscle is responding to: the increase in demand, not the particular variable used to deliver it.
Training two things at once
The concurrent-training question is whether adding aerobic work to strength training compromises the strength adaptation. A 2022 systematic review and meta-analysis pooled 43 studies to answer it, comparing concurrent training against strength training alone.
On the two outcomes the question was originally about, the answer was no. Hypertrophy showed a standardised mean difference of −0.01 (95% CI −0.16 to 0.18) and maximal strength −0.06 (95% CI −0.20 to 0.09), both indistinguishable from no effect. Explosive strength was the single attenuated outcome at −0.28 (95% CI −0.48 to −0.08), a modest reduction.
That attenuation concentrated where the two kinds of training were performed within the same session; when they were separated by at least three hours it was no longer evident. Aerobic modality, training frequency, training status and age made no significant difference to the result — the interference effect, where it exists at all, is narrower and smaller than its reputation.
