Review Reports

Round 1

Reviewer 1 Report (Previous Reviewer 1)

Comments and Suggestions for Authors

1. Clarify the interpretation of the comparison between ET and HIIT

The manuscript appropriately notes that the two protocols differed not only in intensity but also in duration, exercise modality, and work-rest structure. Because several variables differed simultaneously, the observed differences in HRV responses cannot be attributed exclusively to exercise intensity.

I recommend emphasizing this point more explicitly in both the Discussion and Conclusions sections. The current wording occasionally implies that HIIT itself was responsible for the greater autonomic disturbance, whereas the study design only supports the conclusion that the specific HIIT protocol used in this investigation elicited greater HRV suppression than the specific ET protocol examined.

2. Further clarify the implications of the fixed intervention sequence

The authors appropriately acknowledge that ET was always performed before HIIT and that randomization was not feasible within the elite training environment. However, this design feature remains one of the most important methodological limitations.

I recommend expanding the discussion of potential sequence effects, including the possibility that accumulated training adaptations or residual fatigue across the camp may have influenced responses independently of training modality.

3. Moderate several statements regarding practical applications

The manuscript contains several statements suggesting that HRV can be used as a real-time tool for quantifying training load and guiding recovery strategies. While the presented data support the sensitivity of HRV to acute exercise stress, the study did not directly evaluate training prescription, performance outcomes, injury prevention, or decision-making processes.

I recommend slightly tempering these statements so that practical implications remain fully supported by the evidence presented in the study.

Minor Points

4. Clarify the calculation and terminology of the “overall HRV” variable

The manuscript explains that the overall HRV score was derived from lnRMSSD and scaled by a factor of 20 using the Elite HRV application. Since this variable represents the primary outcome showing a significant interaction effect, I suggest clearly identifying it throughout the manuscript as an lnRMSSD-derived composite score to avoid confusion with the broader concept of “heart rate variability.”

5. Provide additional descriptive information for the TRAINING measurements

The manuscript states that HRV values recorded during exercise were averaged to represent the training response. It would be helpful to clarify whether these averages included the entire duration of each session and whether any preprocessing or artifact correction procedures were applied before averaging.

6. Improve consistency of terminology

Several terms are used interchangeably throughout the manuscript, including “overall HRV,” “HRV score,” “total HRV,” and “HRV index.” Using a single term consistently throughout the manuscript would improve readability and reduce ambiguity.

7. Consider adding exact estimated marginal means to the Results

The statistical analyses are reported appropriately; however, presentation of estimated marginal means (or a supplementary table containing mean values with confidence intervals for each time point and protocol) would further improve interpretability and allow readers to better evaluate the magnitude of observed effects.

Overall Assessment

This is a well-written and practically relevant study examining acute autonomic responses to ET and HIIT in an elite wrestling population. The primary findings are clearly presented and supported by the reported analyses. The remaining issues primarily concern clarification of methodological limitations and refinement of interpretation rather than fundamental flaws in the study.

Comments on the Quality of English Language

The manuscript is written in clear and generally fluent English. The text is well organized and the scientific content is communicated effectively. Only minor language editing is recommended to improve conciseness, reduce occasional repetition, and ensure consistency in terminology throughout the manuscript. Overall, the quality of English is suitable for publication after minor editorial revision.

Author Response

Reviewer 1

We thank Reviewer 1 for a thorough and constructive review, and in particular for distinguishing major from minor points, which helped us prioritize the revision.

Comment 1: The interpretation of the comparison between the ET and HIIT protocols should be clarified, given that the two protocols differed not only in intensity but also in duration, modality, and work-rest structure. We recommend that the authors more explicitly emphasize this distinction in both the Discussion and the Conclusions, with wording such as: ‘the specific HIIT protocol used in this investigation elicited greater HRV suppression than the specific ET protocol examined.’

Response: We agree, and have revised the Discussion and Conclusions accordingly, adopting language close to the Reviewer’s suggested wording. The Discussion now states that “the specific HIIT protocol used in this investigation may have elicited a somewhat greater suppression of the lnRMSSD-derived HRV Score than the specific ET protocol examined, a pattern the present sample was not powered to confirm statistically,” and clarifies that “because the protocols differed simultaneously in intensity, duration, exercise modality, and work-rest structure, the observed differences should not be interpreted as being attributable to exercise intensity alone.” The Conclusions were revised in parallel: the findings are now explicitly described as specific “to the sessions examined rather than as an isolated effect of exercise intensity.”

Comment 2: The implications of the fixed intervention sequence should be further clarified, including potential sequence effects such as accumulated training adaptations or residual fatigue.

Response: We have expanded the discussion of order effects in both the Methods and the Limitations. Methods Section 2.1 now states that “randomization or counterbalancing was not feasible given the operational constraints of the elite training environment,” that the study “should be considered a field-based repeated-measures investigation rather than a randomized crossover trial,” and that “potential order effects related to training camp progression cannot be completely excluded.” The Limitations section further notes that “the fixed protocol sequence (ET preceding HIIT) represents a design limitation, as potential cumulative fatigue, training adaptations, or other time-dependent effects cannot be fully excluded despite the 15-day interval between sessions.”

Comment 3: The practical-application statements should be moderated, since the study did not directly evaluate training prescription, performance outcomes, injury prevention, or athlete-monitoring decision-making.

Response: We agree, and have rewritten the Practical Applications paragraph to remove the “real-time” and decision-making framing. It now reads: “HRV monitoring, particularly during and immediately after sessions, may serve as a practical, non-invasive tool for quantifying acute physiological stress. These autonomic markers may complement other training-load indicators already used by coaches and practitioners, although the present study did not directly evaluate their use for real-time decision-making or load prescription.”

Minor Points

Comment 4: The terminology used to describe ‘overall HRV’ should be clarified; it should be identified as the lnRMSSD-derived composite score throughout.

Response: Addressed. “Overall HRV” has been replaced throughout with “lnRMSSD-derived HRV Score,” and the derivation is explained in Methods Section 2.5 (natural logarithm of RMSSD, rescaled 0–100 by the Elite HRV application’s proprietary algorithm).

Comment 5: Additional descriptive information should be provided regarding TRAINING-phase measurements, including whether the averages included the entire session duration and whether artifact correction was applied.

Response: Addressed. Methods Section 2.4 now clarifies that “HRV values obtained during exercise represent average values derived from the complete training session” and that “abnormal beats and artefacts identified by the software’s automatic filtering algorithm were corrected through interpolation according to the manufacturer’s procedures,” with exclusion only where “signal loss prevented reliable RR interval acquisition.” Because the ET and HIIT sessions differed in total duration, we have also added a caveat in the Limitations noting that the resulting TRAINING-phase averages “do not represent comparable time windows.”

Comment 6: The consistency of terminology should be improved; ‘overall HRV,’ ‘HRV score,’ ‘total HRV,’ and ‘HRV index’ are currently used interchangeably.

Response: Addressed. The manuscript now uses “lnRMSSD-derived HRV Score” consistently throughout the Abstract, Results, Discussion, and Conclusions.

Comment 7: Consider adding the exact estimated marginal means (EMMs) to the Results, perhaps as a supplementary table with means and confidence intervals for each timepoint and training protocol.

Response: Partially addressed. Table 1 now reports, for every parameter and timepoint, the raw within-cell mean and SD for each protocol together with the model-based mean difference, its 95% confidence interval, the Bonferroni-corrected p-value, and Cohen’s d. We have not added a separate supplementary table of per-cell EMMs with their own confidence intervals, since the reported pairwise contrasts already convey the model-based estimates relevant to the comparisons of interest; we would be glad to add this as a supplementary table if the Editor considers it necessary.

Overall Assessment: We thank Reviewer 1 for the positive overall assessment and for framing the remaining issues as “clarification of methodological limitations and refinement of interpretation rather than fundamental flaws,” which guided the scope of this revision.

Reviewer 2 Report (New Reviewer)

Comments and Suggestions for Authors

The manuscript addresses a relevant applied question concerning the acute effects of two training modalities on heart rate variability (HRV) in elite Greco-Roman wrestlers. The inclusion of national-level elite athletes is a clear strength, and the ecological setting of a real training camp increases the practical relevance of the study. The topic is suitable for readers interested in athlete monitoring, recovery, internal load, and combat sports performance. However, the manuscript requires substantial revision before it can be considered scientifically robust.

The main limitation concerns the study design. Although the study is described as a crossover design, the order of the protocols was fixed, with endurance training performed on Day 7 and HIIT on Day 22. This creates a relevant risk of order effects, cumulative fatigue, adaptation to the training camp, changes in recovery status, and potential influence of the training load performed between the two experimental sessions. The absence of randomization or counterbalancing should be treated as a major design limitation and should be emphasized more clearly throughout the manuscript.

A second major concern is that the two protocols differed across multiple dimensions simultaneously. The endurance training session consisted of a prolonged cycling bout, whereas the HIIT protocol combined cycling and running intervals and had a markedly different duration, structure, modality, and work-rest organization. Therefore, the study does not isolate the effect of training intensity. The results should be interpreted as a comparison between two complex field-based conditioning sessions, rather than as a direct comparison between endurance training and HIIT based only on intensity. The title, abstract, discussion, and conclusion should be revised to reflect this limitation.

The interpretation of HRV during exercise also requires greater caution. HRV metrics obtained during high-intensity exercise may be affected by signal non-stationarity, respiratory frequency, movement artifacts, changes in heart rate, and differences in recording duration. The manuscript acknowledges limitations regarding the LF/HF ratio, but similar caution should be extended to the broader interpretation of HRV during exercise. The authors should provide more details regarding artifact correction, signal quality criteria, removal or interpolation of abnormal RR intervals, segmentation procedures, missing data, and whether comparable time windows were analyzed across protocols.

The methods section should be expanded to improve reproducibility. Important details are missing or insufficiently described, including inclusion and exclusion criteria, control of caffeine, alcohol, medication, sleep, hydration, nutrition, psychological stress, and training load before each protocol. Because HRV is highly sensitive to these factors, their absence limits interpretation, especially in a small sample of elite athletes. Even if these variables were not objectively measured, this should be acknowledged as a limitation.

The statistical approach is generally appropriate in principle, as linear mixed models are suitable for repeated-measures data. However, statistical reporting should be improved substantially. The manuscript should include a complete table with means, standard deviations, estimated marginal means, mean differences, 95% confidence intervals, p values, and effect sizes for each outcome, time point, and training condition. Given the small sample size and elite-athlete variability, individual response plots would also be highly informative. The authors should also report model diagnostics, residual assumptions, treatment of outliers, handling of missing values, and the exact random-effects structure.

The use and interpretation of the “overall HRV index” require clarification. This index is described as lnRMSSD multiplied by 20 and exported from the Elite HRV application. Since RMSSD itself did not show significant between-protocol differences, the authors should explain why the derived HRV score showed a significant interaction and avoid presenting it as an independent physiological construct. The main conclusion that HIIT induced greater autonomic suppression appears to be driven mainly by this derived score rather than consistently by all HRV indices. This point should be discussed more transparently.

The results are generally understandable, but their presentation should be improved. The figures are simple and readable, but they are not sufficient for full interpretation. Tables with numerical values and confidence intervals are needed. In addition, the figure legends are too interpretive and partially repetitive. They should describe the data more objectively and leave interpretation for the results and discussion sections.

The discussion contains useful interpretation, but some claims are overstated. The manuscript suggests that HRV may serve as a real-time tool for training prescription, recovery management, overreaching prevention, and performance optimization. These are plausible practical applications, but the present study did not directly assess performance outcomes, injury risk, overreaching, neuromuscular fatigue, hormonal markers, perceptual recovery, or long-term adaptation. Therefore, these claims should be presented as preliminary applied implications rather than demonstrated conclusions.

The conclusion should be substantially toned down. The data support the statement that both protocols reduced HRV during exercise and that the derived HRV score showed greater suppression during HIIT. However, the study does not provide strong evidence that HIIT generally imposes greater autonomic stress than endurance training, because the protocols were not matched for duration, modality, volume, or order. The conclusion should explicitly state that the findings are specific to the two field-based conditioning protocols tested.

The English language is understandable, but the manuscript would benefit from technical editing. Several sections are repetitive, some sentences are overly long, and terminology is not always consistent. Terms such as “overall HRV,” “HRV score,” “total HRV index,” and “HRV” should be standardized throughout the manuscript. The disclosure of language editing using ChatGPT is transparent and acceptable, but the final manuscript still requires careful scientific and technical editing.

Overall, this is a potentially useful applied study in a difficult-to-access elite population, but the current version requires major revision. The authors should substantially improve methodological transparency, statistical reporting, interpretation of HRV indices, and the strength of the conclusions.

Author Response

Reviewer 2

We thank Reviewer 2 for a detailed and rigorous critique. The comments prompted substantial improvements to the manuscript’s methodological transparency and to the calibration of our claims; we address each point below. As Reviewer 2’s comments were submitted as continuous prose rather than as a numbered list, we have numbered each distinct point in the order in which it appeared, for clarity.

Comment 1: The fixed order of the two protocols (ET on Day 7, HIIT on Day 22) creates a risk of order effects, including cumulative fatigue, adaptation, recovery-status changes, and training-load influence. The absence of randomization or counterbalancing should be treated as a major design limitation and should be emphasized more clearly throughout the manuscript.

Response: Addressed — the same revisions described in our response to Reviewer 1, Comment 2 above (Methods Section 2.1 and Limitations).

Comment 2: The two protocols differ across multiple dimensions simultaneously (cycling-only vs. cycling and running, different duration, structure, and modality); this design does not isolate the effect of training intensity. The title, abstract, discussion, and conclusion should be revised to reflect this limitation.

Response: We have revised the Abstract, Discussion, and Conclusions to state explicitly that the protocols “differed in intensity, duration, modality, and work-rest structure” and that observed differences “should not be interpreted as being attributable to exercise intensity alone” (see response to Reviewer 1, Comment 1). Regarding the title, we reviewed it carefully and believe that “Investigating the Acute Effect of Different Training Protocols on Heart Rate Variability” does not claim an intensity-isolated comparison — it refers neutrally to “different training protocols” — so we have left it unchanged. We remain open to revising it further if the Editor judges this necessary.

Comment 3: Interpretation of HRV recorded during exercise itself requires greater caution given signal non-stationarity, respiratory frequency, movement artifacts, and changes in heart rate. More detail should be provided on artifact correction, signal quality criteria, removal or interpolation of abnormal RR intervals, segmentation procedures, missing data, and whether comparable time windows were analyzed across protocols.

Response: Methods Section 2.4 now describes artifact correction (interpolation per manufacturer procedure) and the signal-quality criterion governing inclusion, and confirms that exercise-phase values are averaged across the complete session. We have added an explicit caveat in the Limitations that, because session durations differed between ET and HIIT, the TRAINING-phase comparison does not reflect comparable time windows. On missing data: the Table 1 note confirms n = 13 per cell at every timepoint for every parameter, i.e., a complete, balanced dataset with no missing values; we are happy to add an explicit sentence restating this if the Editor prefers. No statistical outlier removal was performed; data inclusion was governed solely by the pre-specified signal-quality criterion described above.

Comment 4: The Methods section should be expanded to improve reproducibility. Missing information includes inclusion/exclusion criteria and control of factors such as caffeine intake, alcohol consumption, medication use, sleep, hydration status, nutritional intake, and psychological stress. Even if these variables were not objectively measured, this should be acknowledged as a limitation.

Response: Inclusion and exclusion criteria are detailed in Methods Section 2.2. Regarding the unmeasured lifestyle variables, we have added this explicitly as a limitation: “while the camp environment ensured similar training and living conditions, factors known to influence HRV, including sleep quality, caffeine intake, hydration status, nutritional intake, and psychological stress, were not objectively measured. Therefore, their potential influence on autonomic responses cannot be excluded.”

Comment 5: Statistical reporting should be improved substantially, with a complete table reporting means, standard deviations, estimated marginal means, mean differences, 95% confidence intervals, p-values, and effect sizes for each outcome variable across timepoints and conditions. Individual response plots would also be highly informative given the small sample and elite-athlete variability. The authors should also report model diagnostics, residual assumptions, treatment of outliers, handling of missing values, and the exact random-effects structure.

Response: Table 1 now reports, for each parameter and timepoint, the within-cell M (SD) per protocol, the mixed-model mean difference (ET − HIIT) with its 95% CI margin, the Bonferroni-corrected p-value, and Cohen’s d. The random-effects structure is stated explicitly in Methods Section 2.6 (random intercept per participant) and repeated in the Table 1 note. We have added a note in the Limitations addressing one diagnostic concern visible in the data: within-cell variance is markedly compressed at the TRAINING timepoint relative to other timepoints, and because the model assumes homogeneous residual variance, we flag that the TRAINING-phase confidence intervals in Table 1 should be interpreted with this in mind. All five mixed models converged, and the proportion of variance attributable to between-participant differences (the intraclass correlation) ranged from .19 (SDNN) to .40 (LF/HF ratio). Individual response plots are now provided directly in the main text: Figures 2–6 have each been redesigned as two-panel figures, with Panel A showing every participant’s individual trajectory across the four timepoints (n = 13 per protocol) alongside the protocol mean ± SE, and Panel B showing the full distribution of values at each timepoint/protocol combination as box plots with every individual data point overlaid. We also re-examined model residuals for normality (Shapiro-Wilk test) and for outliers (standardized residuals): SDNN, RMSSD, the LF/HF ratio, and the lnRMSSD-derived HRV Score showed no significant departure from normality (all p > .19) and no residuals beyond |z| = 3. Heart rate was the exception: its residuals departed significantly from normality (p < .001), which we attribute to the same TRAINING-phase compression of within-condition variability noted above, and we have flagged this as an added caveat on the heart rate findings. The complete diagnostic output (convergence status, ICC, Shapiro-Wilk statistics, and outlier counts for all five parameters) is available as supplementary material should the Editor or Reviewer wish to see it in full.

Comment 6: The ‘overall HRV index’ (lnRMSSD × a proprietary scale via the Elite HRV app) needs clarification. Since the raw RMSSD value showed no significant between-protocol difference, the authors should explain why the derived score showed a significant interaction, and should avoid presenting this composite as if it were an independent physiological construct. This point should be discussed more transparently.

Response: Addressed. The Discussion now explains this divergence directly: “the Timepoint × Training Type interaction was significant for the lnRMSSD-derived HRV Score (p = .008) but not for raw RMSSD (F(3,84) = 0.905, p = .446), even though the two metrics are mathematically related. This divergence is consistent with the fact that the Elite HRV application does not apply a simple fixed multiplier to lnRMSSD but rather a proprietary, non-linear, database-calibrated scaling, which can amplify or attenuate between-condition differences relative to the raw RMSSD signal.”

Comment 7: The presentation of results should be improved. Figures are simple and readable but not sufficient on their own; tables with numerical values and confidence intervals are needed. Figure legends are too interpretive and partially repetitive and should describe the data more objectively, leaving interpretation for the Results and Discussion sections.

Response: Addressed. Table 1 now provides full numerical detail (means, SDs, mean differences, CIs, p-values, effect sizes) for every parameter and timepoint. All figure legends (Figures 2–6) have been rewritten to report only objective statistical information — main effects, interaction p-values, and post hoc results — with interpretive language removed; interpretation is now confined to the Results and Discussion text.

Comment 8: The Discussion contains some overstated claims. HRV is suggested as a real-time tool for training prescription, recovery management, overreaching prevention, and performance optimization, yet the study did not directly assess performance outcomes, injury risk, overreaching, neuromuscular fatigue, hormonal markers, perceptual recovery, or long-term adaptation. These claims should be presented as preliminary applied implications rather than demonstrated conclusions.

Response: Agreed, and addressed; the Practical Applications paragraph has been rewritten (see response to Reviewer 1, Comment 3) to remove the real-time/decision-making framing and to state explicitly that the study “did not directly evaluate their use for real-time decision-making or load prescription.”

Comment 9: The conclusion should be substantially toned down. The data support that both protocols reduced HRV and that the derived score showed greater suppression in HIIT, but this does not support a general claim that HIIT generally imposes greater autonomic stress than endurance training, since the protocols were not matched for duration, modality, volume, or order. The conclusion should explicitly state that the findings are specific to the two field-based conditioning protocols tested.

Response: Addressed. The Conclusions now state that the present findings “should be regarded as exploratory, given the small sample size and the use of two training protocols that differed simultaneously in intensity, duration, modality, and structure,” and that they are specific “to the sessions examined rather than as an isolated effect of exercise intensity.”

Comment 10: The English is generally understandable but requires technical editing; some sections are repetitive, with overly long sentences and inconsistent terminology. The disclosure of language editing using ChatGPT is transparent and acceptable, but the final manuscript still requires careful scientific and technical editing.

Response: We have standardized terminology throughout (see response to Comment 6 above) and performed an additional editing pass to shorten overly long sentences and reduce repetition, particularly in the Discussion. The ChatGPT-4o disclosure in the Acknowledgments has been retained.

We thank Reviewer 2 again for recognizing the value of studying HRV in a difficult-to-access elite population, and we believe the revisions above substantially improve the methodological transparency, statistical reporting, and the calibration of our conclusions, as requested.

Reviewer 3 Report (New Reviewer)

Comments and Suggestions for Authors

Comments and suggestions are uploaded.

Comments for author File: Comments.pdf

Author Response

We thank Reviewer 3 for the encouraging assessment and the specific, section-by-section suggestions, which we address below.

Comment 1 (Abstract): Nicely written and accurately reflects the content. I suggest to also report actual HRV values or percentage changes because it would provide better understanding of the practical significance.

Response: Addressed. The Abstract Results now report percentage changes, e.g., “SDNN decreased by approximately 73% in ET and 82% in HIIT; the lnRMSSD-derived HRV Score decreased by approximately 52% in ET and 57% in HIIT.”

Comment 2 (Introduction): Very good. It would be useful that authors also state the hypotheses.

Response: Addressed. The Introduction now states: “it was hypothesized that both protocols would reduce HRV during exercise, with the HIIT protocol producing a greater acute suppression of HRV indices than the ET protocol. A secondary hypothesis was that HRV would demonstrate partial recovery within 24 hours following both exercise sessions.”

Comment 3 (Materials and Methods): The fixed order of intervention is a major limitation and should be mentioned in the Discussion. Please provide a reference that the instrumentation used is valid and reliable for HRV sampling.

Response: Addressed. The order/randomization limitation is now stated in Methods Section 2.1 and reiterated in the Limitations (see response to Reviewer 1, Comment 2). A validation reference for the Polar H10 chest strap has been added in Methods Section 2.4: “The Polar H10 has been validated for accurate RR interval acquisition and HRV analysis at rest and during incremental exercise [23].”

Comment 4 (Results): Reported in adequate manner. No further comment.

Response: We thank the Reviewer; no changes were required.

Comment 5 (Discussion): Nicely written. I suggest to elaborate in greater detail that the protocols differed in multiple characteristics beyond intensity. Other than that, Discussion is excellent.

Response: Addressed (see response to Reviewer 1, Comment 1); the Discussion now explicitly elaborates that the protocols “differed simultaneously in intensity, duration, exercise modality, and work-rest structure.”

Comment 6 (Conclusions and Practical Implications): Would be strengthened by emphasizing the exploratory nature of the findings and the need for confirmation in larger cohorts and wrestling-specific training settings.

Response: Addressed. The Conclusions now state that “the present findings should be regarded as exploratory” and that “future studies should include larger, wrestling-specific cohorts, randomized and protocol-matched designs, and longer recovery monitoring periods to confirm and extend the present findings.”

Round 2

Reviewer 2 Report (New Reviewer)

Comments and Suggestions for Authors

The revised manuscript is substantially improved compared with the previous version. The authors have corrected several important issues, including the overstatement of between-protocol differences, the inappropriate framing of the design as a randomized crossover study, and the interpretation of the lnRMSSD-derived HRV Score. The addition of inclusion and exclusion criteria, a clearer description of the fixed protocol order, expanded statistical reporting, estimated marginal means, confidence intervals, effect sizes, and individual-participant visualizations also strengthens the manuscript.

Nevertheless, several important issues remain before the manuscript can be considered for publication.

First, the study design must be framed even more cautiously throughout the manuscript. The protocols were not randomized or counterbalanced, and ET was always performed on Day 7 whereas HIIT was always performed on Day 22. Therefore, order effects, cumulative fatigue, adaptation across the training camp, and time-dependent changes cannot be excluded. In addition, the two protocols differed simultaneously in intensity, duration, modality, and work-rest structure. As a result, the findings should not be interpreted as isolating the effect of exercise intensity or as a pure comparison between HIIT and endurance training. They should be interpreted as a comparison between two specific field-based conditioning sessions used in an elite wrestling camp.

Second, the validity and interpretation of HRV during exercise require further methodological clarification. HRV measured during high-intensity, non-stationary exercise is sensitive to movement artefacts, respiratory frequency, signal loss, and non-stationarity of RR intervals. Averaging HRV across an entire endurance session and an interval session of very different duration and structure may obscure segment-specific autonomic responses and may not produce directly comparable physiological estimates. The authors should report signal-quality criteria, artefact correction thresholds, the proportion of corrected beats, whether any recordings or segments were excluded, and whether results change when only standardized segments are analyzed. The LF/HF ratio during exercise should be treated as descriptive only, not as a direct measure of sympathovagal balance.

Third, the lnRMSSD-derived HRV Score requires more cautious handling. The revised manuscript appropriately states that this score is derived from a proprietary, non-linear, database-calibrated algorithm rather than a fixed transformation of lnRMSSD. However, this creates a reproducibility issue. Raw RMSSD and/or lnRMSSD should be treated as the primary HRV outcomes, whereas the proprietary HRV Score should be secondary or exploratory. Claims based on the proprietary score should be softened, especially where raw RMSSD does not show the same pattern.

Fourth, the statistical analysis section should be made more transparent. The authors should provide the exact model specification, method for estimating degrees of freedom, covariance/random-effects structure, residual diagnostics, outlier criteria, and a clearer justification for the Bonferroni correction across 28 comparisons per parameter. Some discrepancies between large effect sizes and non-significant adjusted p-values should be explained clearly so readers understand the influence of small sample size, wide confidence intervals, and multiplicity correction. The authors should also avoid stating that the study had “adequate statistical power” for all relevant inferences, because the sample is very small and the study is underpowered for many protocol-specific and interaction effects.

Fifth, the manuscript should clarify the role of artificial intelligence tools. The current statistical methods mention that Python analyses were replicated with assistance from Claude, and the acknowledgments mention language editing with ChatGPT-4o. AI tools may assist with editing or code generation, but they do not independently validate statistical results. The authors should provide reproducible analysis code or supplementary statistical output if they wish to claim independent verification. The AI-use statement should comply with the journal’s policy and should clearly distinguish language editing from statistical analysis support.

Sixth, the practical implications remain somewhat stronger than the evidence supports. The study did not directly test real-time decision-making, training-load prescription, injury prevention, overreaching prevention, or performance optimization. Therefore, statements suggesting that HRV can guide periodization, prevent maladaptation, or optimize recovery should be framed as potential applications rather than demonstrated outcomes of the present study.

Seventh, the manuscript should address participant identifiability. The sample consists of only 13 elite male wrestlers from a national team, and the manuscript reports that one participant was a Paris 2024 Olympic bronze medalist. This information may make individual participants identifiable. The authors should either remove potentially identifying competitive details or explicitly confirm that participants provided consent for publication of potentially identifiable information.

Finally, the English has improved, but additional editing is still needed. Several sentences remain awkward or grammatically imprecise, and some hyphenation and punctuation issues remain. The manuscript would benefit from one more careful scientific language revision, especially in the Abstract, Methods, Discussion, figure legends, and Practical Applications section.

Comments on the Quality of English Language

The English has improved compared with the previous version, but additional scientific editing is still needed. Several phrases remain awkward or imprecise, including expressions related to “cycling based ET”, “within subject repeated-measures”, “during training HRV”, and the interpretation of the proprietary lnRMSSD-derived HRV Score. Some sentences in the Discussion and Practical Applications still overstate the findings or are phrased in a way that may imply causal or applied conclusions not directly tested in the study. A final professional language revision is recommended to improve clarity, grammar, hyphenation, and scientific precision.

Author Response

 We thank the reviewer for the careful and constructive second-round evaluation, and for recognizing the improvements made in the previous revision. Below we address each point in turn; italicized text reproduces the reviewer’s comment, followed by our response and the corresponding manuscript location.

Comment 1: First, the study design must be framed even more cautiously throughout the manuscript. The protocols were not randomized or counterbalanced, and ET was always performed on Day 7 whereas HIIT was always performed on Day 22. … They should be interpreted as a comparison between two specific field-based conditioning sessions used in an elite wrestling camp.

Response: We agree, and have extended this cautious framing to every section reporting or discussing the protocol comparison, not only the Methods. Specifically: Abstract (Methods): “…completed two training protocols — differing in intensity, duration, modality, and work-rest structure — in a fixed, non-randomized order (ET on Day 7, HIIT on Day 22)…” Abstract (Conclusion): “Because protocol order was fixed rather than randomized, these findings should be interpreted as a comparison between two specific field-based conditioning sessions rather than a controlled test of exercise intensity alone.” Section 2.1: “Potential order effects — including cumulative fatigue, adaptation across the training camp, and other time-dependent changes — cannot be completely excluded.” Section 4 (Discussion, opening): “As protocol order was fixed rather than randomized, the following interpretations should be read as pertaining to these two specific field-based conditioning sessions rather than as an isolated test of training intensity.” Conclusions: “…because the protocols differed in duration, modality, and structure … it should be interpreted with particular caution” and “The fixed order of the protocols and the small sample size further warrant cautious interpretation.” Limitations (4th point): explicitly notes that cumulative fatigue and training adaptation across the camp cannot be fully excluded, and that protocol order was set by the national team coaching staff rather than the research team.

Comment 2: Second, the validity and interpretation of HRV during exercise require further methodological clarification. … The authors should report signal-quality criteria, artefact correction thresholds, the proportion of corrected beats, whether any recordings or segments were excluded, and whether results change when only standardized segments are analyzed. The LF/HF ratio during exercise should be treated as descriptive only, not as a direct measure of sympathovagal balance.

Response: We have added the requested methodological detail to the extent the available data allow, and we are transparent about one point that could not be fully resolved. Signal-quality criteria and exclusions (Section 2.5): “Signal quality was categorized automatically by the application as Good, Okay, or Poor; all retained recordings, at rest and during the TRAINING phase, were rated Good … No exercise segment was excluded unless signal loss prevented reliable RR interval acquisition; in the present dataset, no recordings or segments met this exclusion criterion, and complete TRAINING-phase data were available for all 13 participants under both protocols.” Artefact-correction transparency: we explicitly state that “the manufacturer does not disclose the specific artefact-detection and correction algorithm used,” and we now cite the only available independent validation of this algorithm (Perrotta et al., 2017; Himariotis et al., 2022), noting in the Limitations that this validation exists only for resting conditions, not high-intensity continuous exercise. Proportion of corrected beats: we are not able to report an exact numeric artefact-correction rate or threshold, because the Elite HRV consumer application does not expose this figure to the user — only the categorical Good/Okay/Poor rating is available. We state this constraint explicitly (Section 2.5 and Limitations) rather than omit it. Standardized-segment sensitivity analysis: we agree this would strengthen the manuscript, but it is not feasible with the available data: TRAINING-phase HRV was exported by the application only as a whole-session average, and the underlying time-stamped RR-interval series needed to extract equal-duration sub-segments was not retained. This is now stated explicitly in the Limitations: “A sensitivity analysis restricted to standardized, equal-duration segments of each session was not feasible with the available data and is recommended for future studies directly comparing protocols of different duration.” LF/HF as descriptive only: this qualification now appears at every mention of the LF/HF ratio — Results 3.3, Discussion, and Limitations — e.g., “LF/HF values reported during exercise should be interpreted as descriptive indicators of frequency-domain change rather than direct measures of autonomic balance.”

Comment 3: Third, the lnRMSSD-derived HRV Score requires more cautious handling. … Raw RMSSD and/or lnRMSSD should be treated as the primary HRV outcomes, whereas the proprietary HRV Score should be secondary or exploratory. Claims based on the proprietary score should be softened, especially where raw RMSSD does not show the same pattern.

Response: We have restructured the manuscript’s outcome hierarchy accordingly. Section 2.5: “raw RMSSD (together with SDNN) was treated as the primary time-domain HRV outcome in the present study; the lnRMSSD-derived HRV Score was treated as a secondary, exploratory outcome.” This hierarchy is carried through Results 3.4: “a pattern consistent with a true difference that the present sample (n = 13) was underpowered to confirm at the corrected significance threshold”; Discussion: “Because raw RMSSD — the primary time-domain outcome — showed no significant Timepoint × Training Type interaction, the following pattern observed in the secondary, proprietary HRV Score should be interpreted as exploratory”; and Conclusions: “a pattern not observed in the primary raw RMSSD outcome … because this pattern rests on a single exploratory, non-transparent metric, it should be interpreted with particular caution.”

Comment 4: Fourth, the statistical analysis section should be made more transparent. … The authors should also avoid stating that the study had “adequate statistical power” for all relevant inferences, because the sample is very small and the study is underpowered for many protocol-specific and interaction effects.

Response: Section 2.6 has been substantially expanded to specify the exact model (“HRV parameter ~ Training Type × Timepoint + (1 | Participant ID)”, REML), the degrees-of-freedom method (Kenward-Roger approximation), the confidence-interval method (Wald), and the assumption checks performed (Shapiro-Wilk and Kolmogorov-Smirnov residual normality, extreme-residual identification by cluster), with full diagnostic output provided in Supplementary Material S1. Results 3.0 now reports the residual normality outcome for every parameter, including the one exception (heart rate), linked explicitly to a physiological explanation (TRAINING-phase variance compression) developed further in the Limitations. The Bonferroni correction is now explicitly justified and its power cost acknowledged: “we acknowledge that this conservative correction, combined with the small sample size, reduces power to detect true between-protocol differences, several of which showed large uncorrected effect sizes.” Regarding statistical power, we removed the blanket “adequate statistical power” framing. The power analysis (Section 2.2) now states only that the sample met the a priori target “for the primary interaction effect,” and we independently re-verified this calculation (Cohen’s f = 0.30, α = .05, power = .80, assumed correlation r = 0.5, ε = 1) rather than asserting adequate power for all inferences.

Comment 5: Fifth, the manuscript should clarify the role of artificial intelligence tools. … The authors should provide reproducible analysis code or supplementary statistical output if they wish to claim independent verification. The AI-use statement should comply with the journal’s policy and should clearly distinguish language editing from statistical analysis support.

Response: We have both clarified the AI-disclosure language and made the underlying claim verifiable rather than aspirational. Section 2.6 now reads: “The authors additionally implemented an independent replication of the core model results in Python (statsmodels) … Claude (Anthropic) was used to assist in generating and troubleshooting this Python code under direct author supervision, and all code, outputs, and statistical interpretations were reviewed, run, and verified by the authors themselves. This replication reproduced jamovi’s fixed-effects F-tests (matching its Kenward-Roger-based results to within rounding error across all main and interaction effects for all five parameters) as well as its random-effects variance estimates.” The Python script itself is provided as Supplementary Material S2, together with its summary output (Table S2: variance components, ICC, Shapiro-Wilk statistics, and fixed-effects F-tests), so the replication is independently reproducible by the editor or reviewers rather than merely asserted. Acknowledgments now distinguishes the two tools’ roles explicitly: language editing (ChatGPT-4o) versus code-generation assistance under direct author supervision (Claude), with an explicit statement that all scientific content, data interpretation, and conclusions were conceived, written, and critically revised by the research team.

Comment 6: Sixth, the practical implications remain somewhat stronger than the evidence supports. … Therefore, statements suggesting that HRV can guide periodization, prevent maladaptation, or optimize recovery should be framed as potential applications rather than demonstrated outcomes of the present study.

Response: We have systematically reworded the Discussion and Practical Applications sections to distinguish observed findings from untested applications. Representative examples: “Whether this pattern supports HIIT’s inclusion as a time-efficient alternative in high-performance training programs is a potential application that was not directly tested in the present study and would require further research examining training outcomes directly”; “the present study did not directly test their use for guiding training decisions, and this application remains to be demonstrated in future research”; “…although the present study did not directly evaluate their use for real-time decision-making or load prescription” (Practical Applications); and “whether this observation translates into a demonstrated usefulness of HRV monitoring in high-performance settings was not directly tested here and remains a potential application” (Conclusions).

Comment 7: Seventh, the manuscript should address participant identifiability. … The authors should either remove potentially identifying competitive details or explicitly confirm that participants provided consent for publication of potentially identifiable information.

Response: We have removed the specific, potentially re-identifying detail (Olympic medal, competition, and year) and replaced it with a general statement of the cohort’s competitive caliber: Section 2.2: “Several participants had extensive international competition experience, including medal-winning performances at major continental and global championships.” We judged that even a year-redacted mention of an Olympic medal remained re-identifying, given the small and sparsely populated pool of Danish Greco-Roman Olympic medalists, and therefore opted for generalization rather than a consent-based disclosure of identifying detail.

Comment 8 / Quality of English Language: … Several phrases remain awkward or imprecise, including expressions related to “cycling based ET”, “within subject repeated-measures”, “during training HRV”, and the interpretation of the proprietary lnRMSSD-derived HRV Score. … A final professional language revision is recommended to improve clarity, grammar, hyphenation, and scientific precision.

Response: The manuscript has undergone a further complete language revision targeting precisely these examples and their recurrences throughout the text, in addition to a broader hyphenation, punctuation, and grammar pass. “cycling based ET”: consistently hyphenated and set off with em dashes at every occurrence (Discussion, Conclusions): “two different training modalities — cycling-based ET and HIIT — …” “within subject repeated-measures”: standardized to “within-subject, repeated-measures design” throughout (Abstract, Results). “during training HRV”: reworded to “HRV values obtained during the training sessions were averaged to represent the overall within-session HRV response” (Section 2.1). HRV Score interpretive phrasing: consistently qualified as “proprietary, non-linear, database-calibrated” and framed as a secondary, exploratory outcome at every mention (see response to Comment 3). We additionally corrected inconsistent em-dash spacing, incomplete hyphenation, a missing Oxford comma, and two instances of an unnecessary comma before “and” in compound predicates sharing a single subject, identified during a full line-by-line re-read of the revised manuscript.

This manuscript is a resubmission of an earlier submission. The following is a list of the peer review reports and author responses from that submission.

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

This manuscript addresses an important and practically relevant question: how HIIT and endurance training (ET) acutely affect HRV in elite Greco-Roman wrestlers, using repeated measures across baseline, pre-exercise, during exercise, and 24 h post-exercise. The topic is timely, the sample is elite and sport-specific, and the within-subject crossover approach is a clear strength. The statistical framework is also broadly appropriate for the study design.

1. Please resolve the inconsistencies in participant characteristics.
   The manuscript reports different sample characteristics in different sections. In the Subjects section, the athletes are described as having a mean age of 21.26 ± 1.34 years, body mass of 74.82 ± 8.01 kg, and height of 177.07 ± 6.56 cm, whereas in the Results section the same sample is reported as 23.0 ± 3.47 years, 76.61 ± 9.08 kg, and 1.76 ± 0.07 m. These values must be reconciled and made consistent throughout the manuscript.
2. Please clarify the crossover and order-control procedure.
   The manuscript states that the study used a longitudinal crossover design with a 15-day washout period, but the training protocol section indicates that HRV measurements were conducted on Day 7 (ET) and Day 22 (HIIT) of the camp. This presentation does not make it clear whether the order was randomized, counterbalanced, or fixed. Because order effects could influence HRV responses, this point needs explicit clarification.
3. Please define the “general HRV” index.
   The term “general HRV” is used in the methods, results, discussion, and conclusions, and it appears to be one of the key outcomes. However, the manuscript does not define how this index was calculated. Please provide the formula, software output, or reference used to derive it.
4. Please align the interpretation of 24 h recovery with the reported statistics.
   The Results section reports that SDNN and RMSSD showed no significant between-protocol differences at individual time points, whereas the Discussion and Conclusions state that SDNN and RMSSD remained lower after HIIT at 24 h and suggest slower autonomic recovery. These statements appear stronger than the reported statistics support. Please either provide the corresponding post hoc results for the 24 h comparisons or moderate the interpretation.
5. Please temper conclusions regarding long-term adaptations.
   The manuscript concludes that HIIT may elicit “comparable, if not greater, physiological adaptations” relative to ET. However, the study evaluates acute HRV responses and 24 h recovery only; it does not assess chronic adaptation or long-term performance change. The wording should be revised so that the conclusions remain within the scope of the collected data.
6. Please acknowledge that the two protocols are not fully matched.
   The ET session consisted of a 2-hour cycling bout at 70–85% HR followed by exhaustion, whereas HIIT involved alternating cycling and running intervals with different work-rest structures. Because modality, duration, and structure differ, the manuscript should avoid implying that the comparison isolates intensity alone. A more cautious interpretation would strengthen the paper.
7. Please consider whether some figure captions overinterpret the findings.
   Several captions make causal or comparative statements that are stronger than the statistical results reported in the main text. Aligning the captions more closely with the results would improve consistency.
8. Minor editorial issues should be corrected.
   The manuscript contains small formatting and terminology inconsistencies, including the use of both RMSSD and rMSSD, and the affiliation line includes a formatting error (“2Dept.”). These should be standardized.

Overall, the manuscript has clear strengths: an elite athlete sample, a repeated-measures design, multiple HRV time points, and an appropriate mixed-model analysis. At the same time, the internal inconsistencies and interpretive overreach need to be corrected before the work can be considered ready for publication.

Reviewer 2 Report

Comments and Suggestions for Authors

The manuscript addresses a relevant topic and includes elite-level wrestlers, which is a notable strength. However, in its current form, the study has substantial methodological and conceptual limitations that considerably reduce its scientific impact.

A major concern is the insufficient literature review regarding HRV specifically in combat sports and wrestling populations. The Introduction relies predominantly on general HRV and endurance-based literature, while recent evidence related to autonomic monitoring in combat sports is only minimally addressed. As a result, the sport-specific rationale and scientific gap remain insufficiently developed.

Another important limitation is the ecological validity of the study design. The exercise protocols are largely based on endurance-oriented conditioning methods (prolonged cycling and generic HIIT formats) rather than wrestling-specific training demands. This is problematic given that the study aims to draw conclusions relevant to elite wrestlers. Moreover, the rationale for selecting these protocols is not adequately explained or justified.

The primary methodological concern remains the poor standardization between the ET and HIIT protocols. The interventions differed substantially in duration, structure, and likely overall physiological load, making it difficult to attribute HRV responses specifically to training modality. Additionally, the interpretation of HRV during exercise is insufficiently nuanced, particularly regarding LF/HF ratio and autonomic balance during high-intensity exercise, where important methodological limitations exist.

The novelty of the study is also limited, as the main finding—that HIIT induces greater acute HRV suppression than endurance exercise—is largely expected and does not substantially extend current knowledge. Furthermore, recovery-related conclusions are overinterpreted given that recovery was assessed only at a single 24 h time point.