Bioelectrical Responses to Resistance Training Performed to Momentary Failure or with Repetitions in Reserve: A Within-Subject Analysis of Phase Angle and Impedance Components

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

This manuscript compares 8 weeks of unilateral resistance training performed to momentary failure or with repetitions in reserve on segmental phase angle, resistance, reactance, and derived capacitance.  I recommend rejection in its current form because of following reasons: 

My main reason is that the study does not provide sufficient methodological strength or clear scientific contribution to support publication. The final sample is small, with only 19 participants, including 11 men and 8 women, and the sex-specific analysis is therefore very underpowered. The wide credible intervals show that the study mainly provides uncertain estimates rather than strong evidence of similarity between failure and RIR training.

A second major concern is the interpretation of the bioelectrical outcomes. Segmental BIA values obtained from the InBody S10 are indirect, device-dependent estimates. They should not be interpreted as direct indicators of muscle hypertrophy, intracellular water, membrane integrity, or cellular health. This is especially important because the study did not include direct measures of muscle adaptation, such as ultrasound, MRI, DXA segmental lean mass, or muscle thickness.

The BIA methodology also has important limitations. Hydration status was not objectively assessed, and menstrual cycle phase was not controlled in female participants. Since BIA parameters are sensitive to fluid distribution, these uncontrolled factors may strongly affect the measured outcomes.

In addition, capacitance was derived from reactance and was not directly measured. Therefore, CAP is not an independent physiological outcome, and its interpretation adds limited value beyond Xc.

Overall, the manuscript concludes that failure and RIR training produce similar bioelectrical responses, but the data only show no clear evidence of difference under a small and methodologically limited design. I do not think the current study provides enough robust or novel evidence for publication. Therefore, I recommend rejection.

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 2 Report

Comments and Suggestions for Authors

The manuscript addresses a relevant and relatively unexplored question regarding whether resistance training performed to momentary muscular failure or with repetitions in reserve produces different longitudinal bioelectrical responses. The unilateral within-subject design, supervised intervention, standardized bioimpedance assessments, and Bayesian modeling are important strengths. However, the manuscript currently interprets highly imprecise estimates as evidence of similarity, despite limited statistical sensitivity, uncertain measurement repeatability, possible interference between contralateral conditions, and insufficient confirmation that the two training protocols produced a clearly differentiated exercise dose. These issues require revision before the conclusions can be fully supported.

1. The title and conclusions overstate the evidence by stating that bioelectrical responses were similar between protocols. The reported credible intervals are wide and remain compatible with potentially meaningful differences in both directions. No equivalence framework, region of practical equivalence, or predefined smallest effect size of interest was used. The findings should therefore be described as showing no clear evidence of protocol-specific differences, rather than demonstrating similarity or equivalence. The title should be revised accordingly.
2. The within-subject design reduces between-person confounding but also introduces the possibility of interference between conditions. Both contralateral limbs were trained within the same session, and systemic responses related to hydration, glycogen storage, inflammation, and recovery may affect both limbs. Neural cross-education and fatigue transferred from the limb trained first may also attenuate differences between protocols. These limitations should be discussed, particularly because the bioimpedance outcomes are sensitive to systemic fluid changes.
3. The analyzed segmental bioimpedance values are not localized measurements of the trained elbow flexors or knee extensors. The current path for an arm or leg includes tissues beyond the specifically trained muscle group. Consequently, small local adaptations may be diluted within the segmental estimate. The authors should provide evidence that the InBody S10 segmental outputs are sufficiently sensitive to detect unilateral changes in these specific muscle groups or acknowledge that the measurement approach may not adequately test the proposed biological question.
4. The random-effects structure should be determined primarily by the experimental design rather than selected separately for each outcome through predictive model comparison. Measurements were repeated within limbs, which were nested within participants, and protocol was assigned at the limb level. Omitting the limb-within-participant random effect in some models may not adequately represent the correlation of repeated observations from the same limb. The authors should justify this decision and consider using a consistent hierarchical structure that reflects the design across all outcomes.
5. The sample-size justification remains insufficient. The study was designed pragmatically to retain 19 participants and could detect only approximately moderate standardized differences. This provides limited sensitivity for the small effects hypothesized by the authors and is particularly problematic for the three-way protocol-by-time-by-sex interactions based on eight women and eleven men. The sex-specific analyses should be clearly identified as highly exploratory and may be better removed from the primary models or presented only in supplementary analyses.
6. Given the small sample and complex interaction structure, the Bayesian results may be sensitive to prior specification. The authors should report complete prior distributions for every model, provide prior predictive checks, and present sensitivity analyses using alternative reasonable priors. Posterior predictive checks should also be reported for each outcome. This is particularly important for capacitance, which appears highly variable and may not be adequately represented by a Gaussian likelihood.
7. The final analysis excluded three participants after the intervention began, including two for insufficient attendance and one because of injury. This represents a per-protocol analysis and may compromise the original randomized allocation. A participant flow diagram should report recruitment, allocation, follow-up, and exclusions. The authors should clarify whether all available repeated observations from participants who began the intervention could be incorporated into the mixed models and provide an intention-to-treat or available-data sensitivity analysis where possible.
8. Intervention fidelity requires more complete reporting. The manuscript should provide attendance, actual repetitions completed, training volume, loads, progression, and achieved repetitions in reserve for each condition. The RIR condition permitted a range of one to three repetitions in reserve, and the accuracy of RIR estimation during the intervention was not reported. Without these data, it is unclear whether the protocols produced a sufficiently distinct proximity-to-failure exposure or whether differences in total work may have influenced the results.
9. The measurement-error analysis is based on only six participant-date groups and includes a sensitivity-cleaned dataset after excluding flagged pairs. The criteria used to identify and remove these observations should be reported transparently and justified as predefined rather than data driven. The leg phase-angle coefficient of variation of 9.50 percent is especially concerning because the observed protocol contrasts were much smaller than this apparent measurement variability. The magnitude of all longitudinal effects should be explicitly compared with typical error or minimal detectable change.
10. Capacitance was not independently measured but mathematically derived from reactance at a fixed frequency. Therefore, it represents a deterministic inverse transformation of reactance and does not provide an independent outcome. Phase angle is also mathematically determined by resistance and reactance. The manuscript should acknowledge the dependence among these outcomes, reconsider whether capacitance adds meaningful information, and avoid interpreting agreement across mathematically related variables as independent physiological confirmation.
11. The physiological interpretation of the bioelectrical variables should be moderated. Resistance, reactance, phase angle, and derived capacitance cannot directly demonstrate changes in intracellular water, muscle cell mass, membrane integrity, or hypertrophy. No direct measure of muscle size, body-water compartments, glycogen, or cellular properties was included. Statements suggesting that the protocols produced similar physiological adaptations or tissue-hydration changes therefore extend beyond the outcomes measured.
12. The assessment standardization is generally well described, but additional information is needed regarding the exact interval between the final training session and each measurement. A minimum of 48 hours may not ensure equivalent recovery or glycogen restoration if the actual interval varied between 48 and 96 hours. Dietary intake, carbohydrate intake, fluid intake, creatine use, and recent habitual training may also influence bioimpedance. The authors should report how consistently these factors were controlled across baseline, week 4, and week 8 assessments.
13. The primary inferential outcomes and analysis plan should be clearly prespecified. The manuscript evaluates four related outcomes, two body regions, three time points, two protocols, and sex interactions, producing a large number of posterior estimates. Although Bayesian analyses do not require conventional multiplicity correction in the same manner as frequentist testing, extensive model exploration increases the risk of selective interpretation. The authors should identify the primary outcome and contrast, provide trial registration or a protocol if available, and distinguish confirmatory from exploratory analyses.
14. The main manuscript should provide sufficient descriptive information to understand the findings without relying predominantly on supplementary material. Baseline and follow-up values for resistance, reactance, and capacitance are not clearly presented in the main tables. Participant-level or limb-level descriptive values, longitudinal changes, and direct protocol contrasts with credible intervals should be reported for all primary outcomes.
15. The practical conclusion that practitioners may select either approach without expecting meaningful differences in bioelectrical adaptations is stronger than the evidence permits. The wide credible intervals, limited measurement reliability, modest sample, and lack of an equivalence analysis prevent ruling out potentially relevant differences. In addition, recommending raw bioelectrical parameters for monitoring training-induced changes is not supported without demonstrating their longitudinal reliability, sensitivity, and relationship with direct physiological adaptations.
16. Ethical reporting should include the approval date in addition to the full official name of the ethics committee and approval number. The manuscript should also ensure consistent use of the institutional committee name between the Methods and final declarations.

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 3 Report

Comments and Suggestions for Authors

This study employed a unilateral withinsubject controlled design to compare the effects of an 8week resistance training program performed to momentary muscular failure (FAIL) versus with repetitions in reserve (RIR) on segmental bioelectrical parameters (phase angle, resistance, reactance, and capacitance). Nineteen trained participants completed the intervention, and Bayesian analyses revealed small and imprecise betweenprotocol differences across all outcomes, with no clear protocolspecific or sexspecific effects.

Major Concerns

1.The sample size was only 19 participants, and the authors acknowledge that it was determined based on “pragmatic considerations” rather than a prospective power calculation targeting a primary bioelectrical outcome. The posthoc sensitivity analysis (d = 0.59) cannot compensate for this limitation. Given the considerable interindividual variability in parameters such as PhA, R, and Xc, the wide 95% credible intervals observed more likely reflect insufficient statistical precision than true equivalence. I recommend that the authors consistently use cautious wording such as “no clear differences were detected” throughout the manuscript, avoiding implicit claims of equivalence.

2.No formal testretest reliability analysis was conducted for the bioelectrical measurements, yet the authors report ancillary withinday coefficients of variation (CV) based on six duplicate measurements in the Results section. This dataset was not prespecified in the Methods, is derived from a very small number of observations, and lacks description of whether the duplicates covered different conditions, time points, or participants. As such, it cannot substitute for a proper reliability assessment. With small effect estimates and wide confidence intervals, the absence of reliability boundaries makes it difficult to distinguish true physiological changes from measurement noise. The authors should explicitly discuss this limitation and its implications for the interpretation of their findings.

3.Menstrual cycle phase was not tracked or hormonally verified in female participants, despite the fact that fluctuations in estrogen and progesterone can significantly affect fluid distribution and cell membrane properties – which are the core physiological substrates reflected by bioelectrical impedance. With only eight women in the sample, cyclerelated variability may confound the estimation of sexinteraction terms. The authors should expand their discussion on how this limitation might have influenced the direction and precision of the sexspecific estimates.

4.The RIR condition allowed a tolerance range of 1–3 repetitions in reserve, with a target of 2 RIR. If participants frequently performed closer to 1 RIR, the actual training stimulus may not have differed substantially from the FAIL condition, potentially explaining the lack of divergence in bioelectrical responses. The authors should consider this design feature in their discussion, rather than attributing the null findings solely to equivalence between strategies.

Minor Concerns

1.The derivation of R and CAP is described across multiple sections of the Methods. For clarity and reproducibility, I suggest that the authors consolidate all relevant formulas (e.g., R = √(Z² − Xc²), and CAP derived from Xc at 50 kHz) in a single location within the “Bioelectrical impedance assessment” subsection, clearly indicating which parameters were directly exported by the device and which were calculated.

2.Information regarding the allocation of FAIL and RIR to each limb in the final sample of 19 participants, as well as the baseline training loads and training volumes of the three excluded participants (two males, one female) and their initial condition assignment, is not reported. Providing these details would help readers assess betweengroup balance and the representativeness of the analytical sample.

3.In the reference list (page 17), the journal name for reference #10 (Santisteban et al., 2024) is given only as the abbreviation “EJHM”. The full journal title should be provided in accordance with standard formatting requirements, to facilitate reader access.

4.In the first sentence of the “5. Conclusions” section (page 15), the abbreviation “RIR” appears incorrectly as “RiR”. This should be corrected to the allcaps format used consistently elsewhere in the manuscript, and the authors are advised to check for any similar typographical errors throughout the text.

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 4 Report

Comments and Suggestions for Authors

GENERAL COMMENTS

This manuscript examines whether eight weeks of unilateral resistance training performed to momentary failure (FAIL) or with repetitions in reserve (RIR) produces different segmental bioelectrical responses in resistance-trained adults. The topic is relevant because raw BIA variables are increasingly used as practical monitoring markers, and the within-subject unilateral design is a strength because each participant serves as their own control. The standardized pre-test conditions, supervised training, and Bayesian mixed-effects modeling are also positive features.

However, several issues should be addressed before the manuscript is suitable for publication. The most important concern is interpretation: the study is not powered to establish equivalence between FAIL and RIR, especially for sex-specific effects, and the authors appropriately acknowledge uncertainty in places. This caution should be applied consistently in the abstract, discussion, and conclusion. The conclusion should emphasize “no clear evidence of protocol-specific differences” rather than implying that the two methods definitively produce the same bioelectrical adaptations.

A second concern is measurement specificity. The manuscript clearly states that the InBody S10 outputs are segmental limb-level estimates rather than localized measurements over the trained elbow flexors or knee extensors, and that R and CAP were derived rather than directly measured. This is important, but the physiological discussion sometimes moves beyond what the data can support. Statements linking CAP, Xc, or PhA to intracellular volume, cellular integrity, or muscle cell mass should be softened unless supported by direct validation measures.

Finally, main-text reporting should be improved. PhA receives tables and figures, but R, Xc, and CAP are primary outcomes and are mostly summarized briefly or relegated to supplementary materials. Training exposure and 1RM outcomes are also insufficiently reported, even though the intervention and 1RM testing are described in detail. A compact main-text table reporting actual training volume, load progression, adherence, and pre-post 1RM by condition would strengthen interpretation substantially.

SPECIFIC COMMENTS

Page 1, Lines 13-28: The abstract is clear, but the statement that responses are “similar” should be phrased more cautiously because the credible intervals are wide and equivalence was not formally tested. Please report the final sample size and indicate that sex-specific analyses were exploratory.

Page 2, Lines 61-65: The caution that segmental estimates should not be interpreted as whole-body or localized values is helpful. Please ensure the same caution appears in the abstract and conclusion.

Page 3, Lines 126-140: The sample size justification is transparent, but the study was not powered for small effects or sex interactions. Please avoid overemphasizing sex-specific conclusions, or move them clearly into exploratory secondary analyses.

Page 4, Lines 148-170: Please report actual completed repetitions, loads, total volume-load, load progression, and RIR accuracy for FAIL and RIR limbs. Without these data, readers cannot judge whether the protocols were comparable beyond set number and target endpoint.

Page 5, Lines 183-198: 1RM testing is described in detail, but 1RM results are not presented. Please add pre-post 1RM changes by condition and limb, or explain why strength outcomes were not analyzed.

Page 5-6, Lines 223-257: The description of electrode placement, segmental outputs, and derived R/CAP is partly duplicated. Please consolidate this section and state once, clearly, which variables were device-exported and which were calculated.

Page 6, Lines 258-288: The Bayesian model description is detailed but dense. Add a plain-language statement of the main estimand: the protocol difference in change from baseline to week 4 and week 8. Baseline FAIL-RIR contrasts are less relevant than longitudinal change contrasts.

Page 8-12, Tables/Figures: Because R, Xc, and CAP are listed as primary outcomes, include a compact main-text table for their week-4 and week-8 FAIL-RIR change contrasts, not only PhA. Figure 1 and Figure 2 are useful, but the other outcomes need similar visibility.

Page 12, Lines 367-377: The statement about the absence of formal within-study test-retest reliability should be reconciled with the same-day repeatability analysis. Clarify that the duplicate data are pragmatic short-term repeatability descriptors, not full reliability validation.

Page 13, Lines 391-418: The physiological interpretation of R, Xc, and CAP is useful but should be softened. These variables are indirect and analyzer-specific; they should not be described as direct measures of fluid compartments, muscle cell mass, or membrane integrity.

Page 15, Lines 526-536: The practical conclusion should be more cautious. The data suggest no clear protocol-specific divergence under the present conditions, but they do not prove that FAIL and RIR produce equivalent bioelectrical adaptations.

Minor issues: Page 13, Line 432: correct “elicitedsimilar.” Page 15, Line 537: remove the unused “Patents” heading. Page 16, Lines 571-573: the Conflicts of Interest statement currently describes funder role but does not explicitly declare whether conflicts exist.

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Round 2

Reviewer 1 Report

Comments and Suggestions for Authors

The revised manuscript has improved substantially, and the authors have adequately addressed the main concerns raised in the previous review. The response letter directly addresses the small sample size, the exploratory nature of sex-specific analyses, the indirect and device-dependent nature of segmental BIA outcomes, the lack of objective hydration and menstrual-cycle control, and the fact that CAP is derived from Xc rather than directly measured.

The manuscript is now more appropriately framed as a small, hypothesis-generating within-subject comparison, rather than as evidence of equivalence between training to failure and training with repetitions in reserve. The revised text also clearly states that the estimates are imprecise and that the findings provide no clear evidence of protocol-specific or sex-specific divergence, while meaningful differences cannot be excluded. This cautious interpretation is appropriate.

I think there are only minor editorial issues remain. The final clean version should be checked carefully to remove tracked-change artifacts, duplicated or merged sentences, and any remaining wording that may imply equivalence or direct physiological interpretation of BIA variables. In particular, phrases suggesting “similar physiological adaptations” or direct monitoring of “tissue hydration and cellular properties” should be softened, because the study measured segmental bioelectrical descriptors rather than direct tissue-level adaptations.

Author Response

Please see the attachment. 

Author Response File: Author Response.docx

Reviewer 2 Report

Comments and Suggestions for Authors

The revised manuscript has improved substantially, and the authors have satisfactorily addressed most of the methodological and statistical concerns raised in the previous review. The additional sensitivity analyses, intervention-fidelity reporting, clarification of the hierarchical Bayesian models, measurement-error contextualization, and more cautious interpretation of the bioelectrical outcomes strengthen the manuscript considerably. At this stage, the remaining issues are primarily related to ensuring that the revised manuscript fully and consistently implements the changes described in the response letter.

1. The manuscript should consistently avoid interpreting the absence of clear protocol-specific differences as evidence of similarity or equivalence. Although the response letter explicitly acknowledges this distinction, residual statements remain in the title, Discussion, and Conclusions describing FAIL and RIR as producing “similar” bioelectrical or physiological adaptations. These statements should be replaced with the more appropriate interpretation that no clear or consistent protocol-specific pattern was identified and that the wide credible intervals preclude conclusions of equivalence.
2. A complete editorial audit of the revised manuscript is required. Several sections currently contain duplicated old and revised text, occasionally resulting in contradictory statements. This is particularly evident in the Introduction, bioelectrical assessment methods, statistical analysis, Results, Discussion, and Conclusions. Figure numbering and several sentences also contain residual revision artifacts. Only the final intended version of each passage should remain.
3. Please ensure that all changes described in the response letter are fully reflected in the final manuscript. In particular, CAP should no longer be presented as a primary outcome in the Abstract or main inferential narrative, and the ethics committee name, approval date, and approval number should be reported consistently in both the Methods and Institutional Review Board Statement.

Author Response

Please see the attachment. 

Author Response File: Author Response.docx

Reviewer 3 Report

Comments and Suggestions for Authors

 Accept in present form

Author Response

Thank you

Reviewer 4 Report

Comments and Suggestions for Authors

The author has addressed my questions.

Author Response

Thank you