REF-1360 · Psilander et al., 2019 · Journal of Applied Physiology · Paper · PubMed

If I lifted before, will my muscles come back faster?

In 19 sedentary adults a leg trained, then rested 20 weeks, regained size and strength no faster than an untrained one; training added no nuclei, so muscle memory went untested.

Main caveat: The test failed at its first step. No myonuclei were added in T1 and fibre volume did not change, so the trial could not test whether retained nuclei speed…
limited evidenceMeasured on non-athletes

Psilander, Gundersen and colleagues asked whether a previously strength-trained human leg responds better to a new training period than a previously untrained leg, as mouse studies of myonuclear "muscle memory" predict.

Nineteen healthy, previously inactive young adults (9 men, 10 women) trained one randomly chosen leg for 10 weeks (T1: leg press and leg extension three times a week at 70-85% of 1RM, undulating, plus two weeks of blood-flow-restricted low-load work), stopped all training for 20 weeks, then trained both legs for 5 weeks (T2); they took 25 g of whey protein after each session.

In T1 the trained leg's leg-extension 1RM rose 20% (40 ± 12 to 48 ± 14 kg), vastus lateralis thickness rose from 20.4 ± 3.5 to 22.4 ± 3.7 mm and total fibre cross-section by about 17%, but neither myonuclear number (by two methods) nor single-fibre volume changed.

After 20 weeks off, thickness had returned to the level of the untrained leg while about 60% of the strength gain was still present. In T2 both legs gained strength and about 10% in thickness to the same extent; fibre cross-section, fibre volume and myonuclear number did not change in either leg.

The authors conclude that previously trained human muscle did not respond differently to a second training period, but that this does not rule out muscle memory, since the first period added no nuclei; they attribute the retained strength to motor learning and call it misleading to name that "muscle memory".

Our reading: in untrained adults a short block of strength training left no detectable cellular advantage after 20 weeks, so this trial gives no support for a faster comeback — but because the cellular memory was never created, it does not refute the myonuclear model of REF-1359 either. The two are positions in a live dispute.

Protocol

Design: within-subject, two-leg design over 35 weeks; legs randomly assigned and counterbalanced to "mem-leg" (trained in T1 and T2) and "con-leg" (trained only in T2); approved by the Regional Ethics Committee of Stockholm.
Participants: n = 19 healthy, inactive volunteers (25 ± 1 years, 71 ± 4 kg, 175 ± 8 cm; 10 women, 9 men), none with regular sport or physical activity; one of 10 men recruited dropped out before T1.
T1 (10 weeks, mem-leg only): three sessions a week — Monday 70-75% 1RM (~10-12 reps) and Friday 80-85% 1RM (~5-7 reps) to volitional failure, Wednesday at 10% lower load, not to failure; three sets of leg press then three of leg extension; in weeks 4 and 8 instead five sessions of blood-flow-restricted training (~100 mmHg cuff; ~20% 1RM leg extension, ~30% leg press). Supervised; 25 g whey protein (One Whey, Fitnessguru Sweden AB) after each session in both periods.
Detraining (~20 weeks): no training allowed; daily activity not controlled.
T2 (5 weeks, both legs): the same programme, legs trained one at a time, alternating sets.
Measures: unilateral leg-extension 1RM (two familiarisation sessions); vastus lateralis thickness by ultrasound at 25, 50 and 75% of thigh length; biopsies of vastus lateralis (mem-leg before and after T1, both legs before and after T2) for fibre cross-section by fibre type, myonuclei per fibre cross-section and per sarcomere in isolated single fibres (blinded analyst), and myonuclear domain. Cross-section analyses n = 17.
Analysis: data normalised to each leg's first measurement; repeated-measures ANOVA with Greenhouse-Geisser correction and Sidak's comparisons; mean ± SD; p ≤ 0.05.
Funding and interests: "supported by grants from the Swedish National Centre for Research in Sports (P2016-0134), the Swedish Military Forces Research Authority (Grant #AF9220915), the Research Council of Norway (240374), and the Swedish School of Sport and Health Sciences" (as printed in the accepted manuscript; Crossref's funder list adds Mid Sweden University, MM10); "The authors have no conflicts of interest, financial, or otherwise, to declare".The whey protein is named with its maker; the maker is not listed as a funder.

Key Results

Measure (mean ± SD)Mem-legCon-leg
1RM leg extension, T140 ± 12 → 48 ± 14 kg (+20%, p < 0.01)40 ± 12 → 42 ± 13 kg (+5%, p = 0.06; not trained)
Strength kept after 20 weeks off~60% of the T1 gain—
1RM, T245 ± 14 → 49 ± 13 kg (p < 0.01)41 ± 13 → 47 ± 13 kg (p < 0.01) — "improved to the same extent" (normalised to pre-T1)
Muscle thickness, T120.4 ± 3.5 → 22.4 ± 3.7 mm (p < 0.01)21.0 ± 3.5 → 21.2 ± 3.5 mm
Muscle thickness after 20 weeks offgain lost; both legs similar at the start of T2—
Muscle thickness, T220.5 ± 3.3 → 22.8 ± 3.6 mm (p < 0.01)20.8 ± 3.7 → 22.7 ± 3.5 mm (p < 0.01) — both ~10%, in half the time of T1
Fibre CSA (all fibres), T14020 ± 1086 → 4718 ± 1145 µm² (~17%, p < 0.01); type 2 ~18% (p < 0.01), type 1 ~13% (p = 0.08)—
Fibre CSA in detraining and T2unchanged (both legs)unchanged
Myonuclei (cross-sections and single fibres)no significant change in T1, detraining or T2no change
Fibre volume, myonuclear domainno significant changeno change
Sexessame responses; men larger and stronger at most time points

Mechanisms

Measured: 10 weeks of training raised strength, muscle thickness and fibre cross-section but not myonuclear number or single-fibre volume; thickness gains reversed in 20 weeks off while most strength gain did not; on retraining the previously trained leg gained no faster.
Authors' interpretation: the hypertrophy (under 20% in cross-section, none in fibre volume) was probably too small to require new nuclei (the myonuclear domain ceiling); the strength retained after detraining is motor learning in the nervous system, which should not be called "muscle memory"; the mismatch between thickness and fibre size may come from oedema, different scan and biopsy sites, or damage from repeated biopsies; epigenetic memory was not measured; retained fibre size during detraining may reflect more daily activity, which was not controlled.
Authors' interpretation (the dispute): the result "does not rule out the existence of human muscle memory", because the mouse studies tie the memory to added myonuclei, which did not occur here.
Inference: the trial tests the practical claim — a muscle trained once comes back faster — in the setting most athletes ask about (a block of training, a long break, a return), and finds no advantage at this dose in untrained adults; it does not test the cellular claim. Both statements are needed to describe it honestly.
Inference: the control leg gained 5% in strength without training (p = 0.06), the cross-education the authors describe, so the con-leg was not fully naive at T2; this would, if anything, shrink a true difference between legs.

Practical Application

If I lifted before, will my muscles come back faster? In 19 sedentary adults a leg trained, then rested 20 weeks, regained size and strength no faster than an untrained one; training added no nuclei, so muscle memory went untested.
Strength outlasts muscle thickness in a break: after 20 weeks without training about 60% of a 10-week strength gain remained while the gain in muscle thickness was gone (fibre cross-section, by the results, stayed up) (B, 19 untrained adults); the authors attribute the kept strength to motor learning. This fits the base's line that strength is lost more slowly than aerobic fitness (REF-374, REF-375, SYNTH-36).
Do not promise a faster comeback: this trial found no faster regain, and the myonuclear model that predicts one rests on animal work (REF-1359). The trial's introduction cites one study in which young men gained muscle mass faster in a second identical 7-week block after 7 weeks off (12% vs 7%), with no untrained comparison limb; that study is not in the base. Not from this paper; the base's own practice for returns is in SYNTH-36.
What it is good for as a source: the human counterweight to the myonuclear memory model, and a measured time course of what is kept (strength) and lost (muscle thickness) after 20 weeks off in untrained adults. What it does not show: anything for trained athletes, for endurance fitness, or for a programme that does add myonuclei.

Limitations

The test failed at its first step. No myonuclei were added in T1 and fibre volume did not change, so the trial could not test whether retained nuclei speed regrowth (the authors' own conclusion).
Inconsistent reporting of fibre size. The abstract says muscle thickness and fibre cross-section "returned to baseline values" during detraining; the results say fibre cross-section "remained unchanged during DT and T2" and the discussion that participants "retained muscle fiber hypertrophy" during detraining; the abstract even contradicts itself, ending that "no clear detraining effect was observed for cell size during DT".Only muscle thickness clearly returned to baseline.
Small, untrained and short. 19 sedentary adults (fibre cross-section n = 17), one 10-week block; no endurance or trained participants.
Not a pure control. The con-leg gained strength by cross-education (+5%, p = 0.06); thickness and fibre-size results disagreed, possibly from biopsy sites and oedema (the authors' explanations).
Uncontrolled detraining. Daily activity during the 20 weeks off was not measured; epigenetic changes were not studied.
Source text. Read from the accepted manuscript, which "may contain minor differences" from the published version.
Funding and interests: Swedish National Centre for Research in Sports, Swedish Military Forces Research Authority, Research Council of Norway and the Swedish School of Sport and Health Sciences (as printed; Crossref adds Mid Sweden University); no conflicts declared; the whey supplement's maker is named in the methods, not as a funder. The senior author is the author of the model under test (REF-1359).
How this entry was checked and changed
Why this level

a within-person comparison in 19 adults, which by design would sit with a controlled crossover of 15 or more (B+), held a step lower because the manipulation the hypothesis needs failed (no myonuclei were added in T1, so by the authors' own account it was not a rigorous test), the abstract's statement that fibre cross-section returned to baseline in detraining conflicts with the results, and the control leg was not untouched (a near-significant 5% strength gain from cross-education). Read from the full text (the accepted manuscript in the Norwegian research archive; the publisher's free PDF sits behind a bot check and was not used)

Full text (checked 2026-10-28)
Study participants included 19 healthy, inactive volunteers (25 ± 1 years old, 71 ± 4 kg, 175 ± 8 cm, 10 women and nine men), who had never engaged in any regular sport or physical activity (full text)
Maximal strength improved by 20% in the mem-leg after T1 (40 ± 12 to 48 ± 14 kg, p < 0.01), and 60% of this improvement was still present after the DT period (full text)
A trend for a 5% strength improvement in the con-leg after T1 (40 ± 12 kg to 42 ± 13 kg, p = 0.06) (full text)
After T2, the strength in both the mem-leg (45 ± 14 to 49 ± 13 kg, p < 0.01) and the con-leg (41 ± 13 to 47 ± 13 kg, p < 0.01) improved to the same extent (full text)
After T1, an increased MT was observed in the mem-leg (20.4 ± 3.5 to 22.4 ± 3.7 mm, p < 0.01), whereas the con-leg was unaffected (21.0 ± 3.5 mm to 21.2 ± 3.5 mm) (full text)
After T2, both the mem-leg (20.5 ± 3.3 to 22.8 ± 3.6 mm, p < 0.01) and the con-leg (20.8 ± 3.7 to 22.7 ± 3.5 mm, p < 0.01) showed a similar increase in muscle thickness (~10%) (full text)
After T1, an increase of ~17% in total fiber CSA of the mem-leg (4020 ± 1086 to 4718 ± 1145 µm2, p < 0.01); both legs CSA remained unchanged during DT and T2 (full text)
Neither method showed any significant effect of training or detraining on the number of myonuclei (full text)
Grants from the Swedish National Centre for Research in Sports (P2016-0134), the Swedish Military Forces Research Authority (Grant #AF9220915), the Research Council of Norway (240374); the authors have no conflicts of interest (full text)