Author(s): ZHU, X., BOISBLUCHE, S., PHILIPPE, K., DEMY, M., SUN, P., PRIOUX, J., Institution: ENS RENNES (ÉCOLE NORMALE SUPÉRIEURE DE RENNES), Country: FRANCE, Abstract-ID: 1924

INTRODUCTION:
In professional rugby union (RU), the transition from pre-season (focusing on capacity building) to early in-season (emphasizing tactical specificity and recovery) represents a fundamental shift in training stressors. Understanding this transition is essential to evaluate monitoring model sensitivity across varying modalities. This study quantified the evolution of external (ETL) and internal (ITL) training loads during this critical period.
METHODS:
14 professional players (7 Forwards, 7 Backs) from a French Pro D2 RU championship were monitored. The study followed a longitudinal design divided into 2 periods based on seasonal structures: (i) the pre-season (6 weeks) and (ii) the early in-season (first 6 weeks of competition). ETL was quantified using 16 GPS metrics, from which the Acute/Chronic Workload Ratio (ACWR) was calculated via a 7:28 day rolling window[1]. To account for "hidden loads" not captured by GPS, resistance training intensity was quantified via coach-planned Gym Rating of Perceived Exertion (RPE) for each session[2]. ITL was assessed using over 30 Heart Rate Variability (HRV) parameters and one wellness questionnaire, with rolling Z-score normalization applied to account for individual longitudinal fluctuations and seasonal effects[3]. To identify temporal relationships between subjective fatigue and physiological markers, Pearson’s correlation and Cross-Correlation Function (CCF) analysis (±3-day window) were employed to distinguish between leading and lagging effects[4]. Dimensionality reduction was performed using Least Absolute Shrinkage and Selection Operator (LASSO) regression (to select 10 core GPS predictors) and Principal Component Analysis (PCA) to extract a "Global HRV Index"[5]. Finally, linear regression models with standardized residuals were utilized to quantify "unexplained fatigue" induced by the integration of locomotor and non-locomotor metrics[6].
RESULTS:
Pre-season showed a distinct loading profile, with Forwards’ contact-intensive load peaking at 1.24 ACWR before transitioning into a stabilization period during the early in-season, facilitating autonomic recovery.CCF revealed that for Forwards, physiological markers predicted subjective fatigue 24h in advance ("Golden Window"), while Backs showed same-day synchronicity.Locomotor metrics (GPS) explained 90.2% of variance for Backs but only 33.8% for Forwards.Integrating Gym RPE yielded a 31.06% relative R2 improvement for Forwards (p=0.049), whereas its impact on Backs was marginal(5.48% improvement, p=0.418).
CONCLUSION:
GPS alone underestimates pre-season physiological strain, especially in Forwards.This study highlights the necessity of multidimensional monitoring (locomotor+resistance+physiological) in pro RU.Practitioners should use advanced HRV for central fatigue and position-specific recovery to optimize internal load status.
REFERENCES: 1.Griffin,SportsMed,2020; 2.Inoue,SpMedOp,2022; 3.Sauliere,Biostat,2019; 4.Brandl,JMCR,2024; 5.Kolodziej,SJMSS,2023; 6.Altman,NatMeth,2016.