Training and Detraining Effects of Eccentric Exercise and Static Stretching on Hamstring Aponeurosis Dimensions

ABSTRACT

This study examined the effects of 10 weeks of eccentric knee‐flexor training or matched‐duration static stretching on hamstring aponeurosis dimensions and muscle size, and explored the changes after 4 weeks of detraining. Eleven healthy untrained men completed a within‐participant contralateral intervention (3 sessions·week⁻¹, 10 weeks) followed by 4 weeks of detraining. One leg performed eccentric knee‐flexion exercise on an isokinetic dynamometer (3 × 10 reps at 5°·s⁻¹; 50% maximal eccentric torque), while the contralateral leg performed static stretching (3 holds matched to eccentric time under tension). Hamstring outer aponeurosis volumes, inner aponeurosis volumes and widths, and muscle volumes were quantified using magnetic resonance imaging. Outer aponeurosis volume increased after both training protocols across all muscles (p < 0.05) and was maintained during the 4 week detraining period, with comparable changes observed between stretching (~19.0%–31.8%, d = 0.91–1.69) and eccentric training (~15.8%–22.8%, d = 0.85–1.13). Inner‐aponeurosis volume and width of biceps femoris long head and semimembranosus also increased after both training protocols (~7%–23%; p < 0.05; d = 0.33–0.86) and were preserved during detraining. Muscle hypertrophy occurred only after eccentric training (~22%–42%; p < 0.05; d = 0.98–1.65) in biarticular hamstrings and remained above baseline in the detraining period. Stretching exercise produced minimal changes in muscle volume (p > 0.05). Changes in muscle volume were not significantly associated with changes in aponeurosis dimensions in almost all cases (p > 0.05). These findings suggest that hamstring aponeurosis dimensions can increase after both eccentric loading and static stretching, whereas muscle hypertrophy is largely eccentric‐specific and more sensitive to short‐term detraining.

1. Introduction

Hamstring strain injuries are commonly linked to modifiable factors such as reduced eccentric strength or flexibility [1]. However, these factors do not fully explain why the biceps femoris long head (BFlh) is the most frequently injured hamstring muscle, or why a substantial proportion of injuries involve the myotendinous junction (MTJ) [2]. This discrepancy suggests that other factors beyond muscle strength, particularly local morphological features, may contribute to injury risk [3, 4]. One such feature is aponeurosis size, which shows large interindividual variability in hamstrings and is not consistently related to muscle size [3, 5, 6]. Importantly, it remains unknown whether BFlh aponeurosis remodels less with chronic loading and unloading than the other hamstrings, which could help explain its disproportionate strain‐injury incidence.

From a physiological perspective, the aponeurosis is not merely a passive connective tissue sheet but a mechanically active structure [7, 8]. Outer aponeuroses provide broad surface attachments for muscle fibers and contribute to force transmission toward the free tendon and surrounding fascial tissues [7, 8, 9, 10]. Inner aponeuroses, or intramuscular tendinous sheets, provide deeper fiber‐insertion surfaces and may influence internal force transmission and local strain distribution within the muscle belly [7, 8, 9, 10]. Although the distinct mechanical roles of inner and outer hamstring aponeuroses remain unclear, both structures are likely to contribute to how force and strain are distributed across the muscle–tendon unit [7, 8, 9, 10]. Therefore, concurrent adaptations in outer and inner aponeuroses may be relevant as they could influence both the surface‐level force transmission toward the free tendon and the deeper internal strain distribution within the muscle belly. Additionally, during contraction, muscle fibers shorten under near‐constant volume, causing the muscle belly to bulge and imposing substantial transverse strain on the aponeurosis [8, 11, 12]. This deformation influences muscle mechanics, including the fiber rotation, shear force transfer, and the local stress environment near the MTJ [8, 11, 13]. Modeling studies showed that a wider aponeurosis reduces peak fiber strain, whereas a narrow proximal aponeurosis (such as BFlh) produces steep strain gradients, thereby increasing the injury risk [4, 14]. In contrast, semitendinosus (ST) and semimembranosus (SM) muscles possess broader aponeuroses and more uniform fiber‐insertion geometry [15], leading to more even strain distribution and possibly contributing to their lower injury incidence.

Despite its mechanical importance, the capacity of the aponeurosis to adapt to different forms of chronic loading remains poorly understood [16, 17, 18]. Eccentric training represents a high‐force, active lengthening stimulus that increases muscle size and strength and has been shown, in some cases, to induce BFlh aponeurosis remodeling [19]. In contrast, static stretching imposes prolonged, low‐load tensile deformation at long muscle lengths and is widely used in both athletic and clinical practice. While such loading can stimulate connective tissue remodeling [20], it remains unknown whether static stretching alters hamstring aponeurosis dimensions. Furthermore, although hamstrings hypertrophy after eccentric training and its partial reversal during detraining have been documented [21, 22], the response of the hamstring aponeurosis to training cessation has not been examined. While studies in other muscles have shown that training‐induced changes in muscle size are accompanied by proportional adaptations of the aponeurosis [23], evidence in hamstrings is currently lacking.

Understanding how the aponeurosis adapts, and whether such changes scale with muscle hypertrophy, could offer significant insight into the development of more evidence‐based injury‐prevention strategies. Therefore, the purpose of this study was to determine the effects of a 10 week eccentric training and a matched‐duration static stretching intervention on the aponeurosis dimensions and muscle size in hamstring muscles. A second aim was to examine the time course of these adaptations during a subsequent detraining period (4 weeks), to establish whether muscle and aponeurosis exhibit similar or distinct retention patterns after training cessation. A third aim was to assess whether training‐induced changes in muscle size are associated with proportional changes in aponeurosis dimensions. Based on previous research [4, 14, 19, 23], we hypothesized that: (1) eccentric training would elicit greater increases in both muscle and aponeurosis size compared to static stretching; (2) both muscle and aponeurosis adaptations would decrease during a subsequent detraining period and (3) changes in muscle size would be accompanied by proportionate changes in aponeurosis dimensions.

2. Materials and Methods

2.1. Sample Size Calculation

An a priori power analysis was performed using a repeated‐measures analysis of variance (ANOVA) (G*Power, v3.1; F tests, ANOVA: repeated measures, within factors). Each participant contributed 24 repeated measurements per magnetic resonance imaging (MRI)–derived outcome (4 muscles × 2 contralateral conditions × 3 time points). The calculation was based on one group, 24 measurements, a significance level of α = 0.05, a conservative correlation among repeated measures of 0.50, and a nonsphericity correction of ε = 0.75. The effect size was derived from the observed before‐to‐after change in aponeurosis maximum width (before: 3.42 ± 0.76; after: 3.68 ± 0.72) [19]. Assuming a conservative before–after correlation of r = 0.50, this corresponded to Cohen's f = 0.248 for a within‐subject effect. Under these assumptions, the required sample size to achieve > 80% power was n = 10. This calculation was intended to estimate power for the overall within‐participant repeated‐measures effect, primarily the change across time and the comparison between eccentric training and stretching, rather than for higher order muscle‐specific interactions.

2.2. Participants

Eleven healthy young men (mean ± SD, 23.3 ± 3.7 years; 170.7 ± 4.9 cm; 66.8 ± 11.7 kg) were recruited to allow for potential attrition during the intervention. The participants had not engaged in structured resistance training for at least 1 year and reported no musculoskeletal injury to either lower limb within the past 2 years. After receiving written and verbal explanations of all study procedures, risks, and requirements, each participant provided written informed consent. The study was approved by the institutional Ethics Committee (No. 23057‐1) and conducted in accordance with the Declaration of Helsinki.

2.3. Experimental Design

We used a within‐participant contralateral design, meaning each participant's two legs received different treatments (interventions) and therefore each person served as their own control, similar to previous studies [24, 25]. This design minimizes between‐participant differences and improves the ability to detect true intervention effects [25]. Therefore, the contralateral limb was used as a relative comparator between two loading stimuli, rather than as an absolute no‐intervention control. This is important given the substantial individual variability in muscle and aponeurosis volume adaptations reported after resistance training [18, 19, 26].

The experiment included baseline testing, a 10 week intervention period, and a subsequent 4 week detraining phase (Figure 1). During each testing session, aponeurosis dimensions and muscle volumes of all hamstring muscles were assessed bilaterally. Muscle volume data from the same cohort have also been reported elsewhere, but in the present study they were analyzed to determine whether changes in aponeurosis dimensions occurred independently of, or in parallel with, changes in muscle size. One week later, participants began the intervention protocol, which involved three supervised sessions per week for 10 weeks. For each participant, one leg was allocated to eccentric knee‐flexor training and the contralateral leg to static stretching. Allocation was performed using block randomization with a block size of two to counterbalance leg dominance, so that the eccentric intervention was not systematically assigned to either the dominant or nondominant limb. Because the final sample included 11 participants, dominance was approximately but not perfectly balanced. Participants were instructed to maintain their usual physical activity, diet, and sleep patterns throughout the study and to refrain from performing additional resistance training or stretching. All outcome variables were reassessed 4–7 days following the final intervention session. All assessments were performed at the same time of day for each participant (between 10:00 and 20:00) to minimize diurnal variation on the measurements.

FIGURE 1.

Study design and MRI‐based assessment of hamstring aponeuroses. (A) Within‐participant contralateral design in 11 untrained young men: One leg completed eccentric knee‐flexor training and the contralateral leg completed static stretching (3 sessions·week⁻¹ for 10 weeks), followed by 4 weeks of detraining. (B) Three‐dimensional reconstructions of manually segmented hamstring aponeuroses, with inner aponeuroses shown for BFlh (light purple) and SM (light blue), and outer aponeuroses shown for BFlh (purple), BFsh (yellow), ST (light green), and SM (gray). Outer aponeuroses refer to the superficial/peripheral connective tissue sheets located along the muscle belly surface, whereas inner aponeuroses refer to intramuscular tendinous sheets visible within the muscle belly. Representative axial MRIs at different thigh levels are shown for anatomical reference. The distal panel is ‐enlarged for visualization purposes. A coordinate system [anterior (A)–posterior (P) and medial (M)–lateral (L)] is provided to indicate image orientation. (C) Example axial images illustrating segmentation of the inner aponeuroses of BFlh and SM; the inset provides a magnified view used to quantify inner‐aponeurosis width as the maximum mediolateral distance on slices where the inner aponeurosis was visible. BFlh, biceps femoris long head; BFsh, biceps femoris short head; SM, semimembranosus; ST, semitendinosus.

A subsequent 4 week detraining period followed the intervention. Participants were instructed to maintain their usual daily routines while avoiding resistance training, stretching, high‐intensity exercise, and any additional structured lower limb training. Normal low‐intensity activities, such as short cycling commutes and household tasks, were permitted if they were part of the participant's habitual routine. Compliance was monitored through weekly verbal check‐ins, during which participants were asked whether they had performed any prohibited or additional structured exercise. All participants confirmed adherence to the detraining instructions.

2.4. Aponeurosis Dimensions and Muscle Volume Assessments

Aponeurosis dimensions and muscle volumes were evaluated using a 3‐T MRI scanner (MAGNETOM Skyra, Siemens Healthineers, Germany). Participants were positioned supine with the hips and knees fully extended [26]. Cushions supported the trunk, pelvis, and lower legs to prevent contact between the hamstring region and the scanner bed, and two custom foot pads were used to minimize hip rotation in both limbs. The ankles and distal shanks were secured with a nonelastic strap to maintain consistent alignment. To minimize posture‐related fluid shifts that could influence morphological measurements, participants rested in the supine position for at least 20 min before imaging and were instructed to keep their lower limbs relaxed and still throughout the session. Obtained scans were meticulously inspected for motion artifacts, and scans were repeated whenever any displacement or blurring was detected. Before each MRI session, participants were instructed to avoid vigorous exercise for 24‐h and verbally confirmed before scanning that no intense exercise had been performed during the preceding day.

Axial MR images of the thigh were obtained using a 32‐channel spine coil combined with two 18‐channel body array coils. Aponeurosis dimensions and muscle volume were quantified from a T1‐weighted VIBE two‐point Dixon sequence (echo times: 2.46/3.69 ms; repetition time: 5.98 ms; slice thickness: 5 mm; field of view: 315 × 420 mm; reconstructed matrix: 384 × 512; bandwidth: 610 Hz; flip angle: 12°; acquisition time: 19 s per scan). Three axial imaging blocks were collected, beginning just above the ischial tuberosities and extending distally to ensure that the full length of each thigh was captured and therefore the entire aponeurosis and muscle volume. Across sessions, the field of view was reproduced using the same anatomical landmarks, with the proximal coverage set slightly above the ischial tuberosities and distal coverage extending beyond the distal hamstring muscle–tendon region. The coronal and sagittal localizer images were used to confirm comparable proximal–distal coverage before acquisition.

2.5. Aponeurosis Dimensions and Muscle Volume Analysis

MRI datasets were anonymized before analysis and supplied in NIfTI (.nii) format which were analyzed in 3D Slicer (v5.8.1). All segmentations were performed manually by a single investigator (CS), who had access only to the anonymized image files and was unaware of participant identity, the intervention assigned to each limb (eccentric training or static stretching), and the assessment time point (Before, After, or Detraining), using axial images as the primary working plane, with sagittal and coronal views used to verify anatomical continuity. For segmentation, the outer aponeurosis was defined as the hypointense connective tissue sheet located along the superficial or peripheral surface of each muscle belly, whereas the inner aponeurosis was defined as the intramuscular hypointense tendinous sheet visible within the muscle belly. For each participant and time point (Before, After, and Detraining), the outer aponeuroses of the BFlh, BF short head (BFsh), SM, and ST, the inner aponeuroses of the BFlh and SM, and the full muscle volumes of all four hamstring muscles were segmented slice‐by‐slice from the most proximal slice in which the structure was visible to the most distal slice where visualization remained clear. Segmentation relied on careful delineation of fascial and connective tissue borders using the Slicer “Segment Editor” tools, with boundaries cross‐checked in orthogonal planes as needed. For both inner and outer aponeuroses, segmentation followed a standardized anatomical rule based on the continuity between the connective tissue sheet and the muscle belly. On each axial slice, the aponeurosis was included only when the hypointense structure remained in direct contact with the adjacent muscle tissue. Slices in which the same structure appeared without muscular contact, potentially corresponding to proximal or distal free tendon, were excluded. This criterion ensured that volumetric estimates reflected only the aponeurotic portion integrated within the muscle belly and not the free tendon segments (Figure 1B). Cross‐sectional area (CSA) was extracted for each slice, and volumes for muscles and aponeuroses were calculated as the sum of slice‐wise CSAs multiplied by slice thickness (Figure 1B). For BFlh and SM, the maximum inner‐aponeurosis width was additionally measured on every slice containing inner aponeurosis, with the largest mediolateral distance recorded as the representative value for that scan (Figure 1C). This segmentation protocol was applied to both the eccentric and stretching limbs across all time points [4, 18]. Inner aponeurosis measurements were restricted to BFlh and SM because these muscles had clearly identifiable intramuscular aponeurotic sheets that could be consistently followed across axial MRI slices. In contrast, inner tendinous structures in BFsh and ST were not consistently distinguishable from adjacent fascial or tendinous tissues and were not quantified.

2.6. Between‐Session Reliability

Between‐session reliability of aponeurosis dimensions and muscle volume was assessed in six participants from the main intervention cohort (age: 25.2 ± 5.0 years; height: 171.1 ± 4.0 cm; body mass: 69.5 ± 15.3 kg). Reliability was evaluated using two MRI sessions performed 1 week apart before the start of the intervention. The first scan was conducted 1 week before baseline, whereas the second scan corresponded to the baseline assessment used in the main longitudinal analysis. To standardize the evaluation and avoid potential asymmetries influencing the findings, all reliability measurements were conducted exclusively on the left leg. Each participant underwent the same MRI protocol on two separate occasions, scheduled 1 week apart. On both days, aponeurosis dimensions and the hamstring muscle volumes were acquired following identical positioning, scanning, and preprocessing procedures to isolate day‐to‐day variability in the measurement process.

2.7. Intervention Sessions

Each intervention session was conducted with participants seated on an isokinetic dynamometer (Biodex System4, Biodex Medical Systems, USA), positioned with their hips flexed to 120°. The pelvis and trunk were stabilized using straps to limit hip rotation and pelvic displacement, ensuring that movement occurred primarily at the knee joint. Across the 30 intervention sessions, the order in which the eccentric and stretching conditions were performed was randomized and approximately counterbalanced for each participant, so that neither intervention was consistently performed first.

2.8. Eccentric Training

For eccentric training, the exercise range of motion (ROM) was set from 50% to 100% of each participant's maximal seated knee‐joint ROM, which was reassessed every 2 weeks and updated for the subsequent training block. To account for passive torque and gravitational influences, a passive knee‐extension trial at 5°/s was first conducted across the same ROM. The training intensity was standardized to 50% of each participant's most recently measured maximal voluntary eccentric torque [27]. This load was selected based on preliminary work from the broader project showing that eccentric knee‐flexor contractions at 50% of maximal eccentric torque acutely reduced hamstring shear modulus, whereas contractions at 25% or 75% did not [27]. More details can be found in Appendix S1.

Participants completed a warm‐up consisting of five eccentric contractions. They were instructed to generate force rapidly as soon as the lever arm began moving and to maintain the prescribed torque throughout the ROM. The main protocol consisted of three sets of 10 repetitions performed at 5°/s, with real‐time visual feedback of the target torque provided on a computer monitor. A 2‐min rest interval was given between sets. Mean torque across all repetitions was used to calculate the relative torque expressed as a percentage of the maximal eccentric torque measured during the most recent testing week (Appendix S1: Table S1).

2.9. Static Stretching

For static stretching, the knee joint angle was set to 100% of each participant's maximal knee joint ROM, again determined from the latest weekly measurement. A passive knee‐extension test was performed to match the ROM conditions of the contralateral limb. The knee joint was then held at the maximal ROM (onset of pain) for the same duration as the eccentric intervention warm‐up period. Following this, participants completed three static stretching repetitions, each held at 100% ROM for a duration equivalent to the time required to complete 10 eccentric repetitions, corresponding to ~60–74 s per repetition (mean 68.26 ± 5.43 s) depending on the participant's biweekly ROM (Appendix S1, Table S2). A 2‐min rest period was provided between repetitions. Throughout the stretching intervention, participants were instructed to remain completely relaxed and to avoid any voluntary muscle activation. More details can be found in Appendix S1.

2.10. Statistical Analysis

2.10.1. Reliability Analysis

Between‐session reliability of all MRI‐derived measurements was evaluated in six participants from the main intervention cohort. Reliability was evaluated using a two‐way mixed‐effects, single‐measure intraclass correlation coefficient for consistency [ICC_(3,1)]. This model was selected because the same measurement protocol and rater were used across repeated scans, and the aim was to determine the consistency between test and retest values. Reliability was assessed by comparing the initial MRI acquisition with a second scan performed 1 week apart under identical positioning and scanning conditions. ICC values were interpreted following the criteria [28]: values < 0.50 were classified as poor, 0.50–0.75 as moderate, 0.75–0.90 as good, and > 0.90 as excellent reliability. The standard error of measurement was calculated as SEM = SD_difference × √1 − ICC, where SD_difference represents the standard deviation of the participant‐level differences between test and retest measurements. To further determine the magnitude of change required to exceed measurement error, the minimal detectable change (MDC) at the 95% confidence level was computed using MDC₉₅ = 1.96 × √2 × SEM. This value represents the smallest detectable difference that can be interpreted as a true change in aponeurosis or muscle volume rather than a product of measurement variability.

2.10.2. Changes After Training and Detraining

Statistical analyses were conducted in R (v 4.3.0, Vienna, Austria) using linear mixed‐effects models (lme4, lmerTest). Data normality was assessed using quantile–quantile plots together with the Kolmogorov–Smirnov test. For each MRI‐derived variable, Muscle, Exercise condition (eccentric vs. stretching), and Time (Before, After, and Detraining) were entered as fixed effects. Participant was included as a random factor to account for the repeated‐measures structure of the data and the nonindependence of observations obtained from multiple muscles, conditions, and time points within the same individual. Sum‐to‐zero contrasts were used, and denominator degrees of freedom were estimated using the Kenward–Roger method. Model significance was evaluated with Type III ANOVA. When significant main effects or interactions were detected, post hoc comparisons were performed using estimated marginal means with Holm adjustment. To characterize training‐ and detraining‐related adaptations, relative change scores were calculated for all outcomes. Percentage changes were computed for both the Before‐to‐After and After‐to‐Detraining intervals using the formula Δ = (Time₂ − Time₁)/Time₁ × 100.

Statistical significance (p < 0.05) was determined from the mixed‐effects models and model‐based post hoc comparisons, whereas percentage changes, 95% confidence intervals, and Cohen's d values were used to describe the direction and magnitude of the observed changes. Effect sizes were quantified using partial eta squared (${\eta }_p^2$), interpreted as negligible (< 0.01), small (0.01–0.059), medium (0.06–0.139), and large (≥ 0.14) [29]. Cohen's d was also calculated as the mean difference expressed in standard deviation units, with values of 0.2, 0.5, and 0.8 interpreted as small, medium, and large effects, respectively.

2.10.3. Association Between Changes in Aponeurosis Dimensions and Muscle Volume

The relationship between changes in aponeurosis dimensions (volume and inner‐aponeurosis width) and muscle volume was examined using correlation analysis. Pearson's correlation coefficients were calculated for variables that met the assumption of normality, as assessed by the Kolmogorov–Smirnov test, whereas Spearman's rank‐order correlations were used for non‐normally distributed data. The correlation coefficients were reported together with their 95% confidence intervals. Correlation strength was interpreted as follows: 0–0.10 trivial, 0.10–0.39 weak, 0.40–0.69 moderate, and 0.70–0.99 strong [30].

3. Results

3.1. Between‐Session Reliability Analysis

Detailed reliability outcomes for all MRI‐derived aponeurosis measurements are presented in the Appendix S2. Between‐session reliability was generally high (ICC = 0.84–0.99), with low relative SEM (0.2%–3.5%) and MDC (0.5%–9.8%).

Absolute mean ± SD values for all MRI‐derived outcomes across time points and intervention conditions are provided in Appendix S3.

3.2. Outer Aponeurosis

Individual participant data for hamstring outer aponeurosis volume across the Before, After, and Detraining periods in both the eccentric and stretching interventions are illustrated in Figure 2. The analysis showed a significant main effect of Time (F _(2,230) = 70.90, p < 0.001, ${\eta }_p^2$ = 0.38) and Exercise Condition (F _(1,94.41) = 8.49, p = 0.004, ${\eta }_p^2$ = 0.08) for outer aponeurosis. Post hoc tests indicated a significant increase in outer aponeurosis volume following training, while no differences were found between the post and detraining period (p > 0.05). A significant main effect of Muscle (F _(3,230) = 269.15, p < 0.001, ${\eta }_p^2$ = 0.78) was also found with SM presenting greater aponeurosis volume than the other muscles. All interactions were nonsignificant (all p ≥ 0.29, ${\eta }_p^2$ ≤ 0.03).

FIGURE 2.

Upper panels: Individual data points for outer aponeurosis volumes of Biceps Femoris long (BFlh), short head (BFsh), Semitendinosus (ST), and Semimembranosus (SM) at Before, After, and Detraining period (n = 11). Each participant is shown with connected points to illustrate their progression over time. Colors indicate the eccentric‐training (blue), static stretching conditions (magenta). Lower panels: Before‐to‐After (light colors) and After‐to‐Detraining (dark colors) percent change for the same muscles. Each point shows an individual participant's change, and the overlaid markers with error bars show the group mean and its 95% confidence interval. Significant increases following both interventions were found for the outer aponeurosis of all the muscles (p < 0.05). No significant differences were found between the after and detraining period (p > 0.05). *Different from before training (p < 0.05).

Eccentric training increased outer aponeurosis volume in BFlh (+22.8%, 95% CI [15.0%, 30.7%]; d = 1.13), and in BFsh (+22.8%, 95% CI [11.8%, 33.8%]; d = 0.98), with comparable increases observed in SM (+15.8%, 95% CI [8.3%, 23.3%]; d = 0.95), and in ST (+21.1%, 95% CI [13.5%, 28.5%]; d = 0.85). Static stretching produced comparable adaptations with eccentric training, increasing outer aponeurosis volume in BFlh (+30.1%, 95% CI [22.6%, 37.6%]; d = 1.69), BFsh (+20.4%, 95% CI [12.1%, 28.7%]; d = 1.00), SM (+19.0%, 95% CI [10.4%, 27.6%]; d = 0.91) and in ST (+31.8%, 95% CI [25.2%, 38.4%]; d = 1.46).

For all muscles and in both training modes, after‐to‐detraining changes were negligible, indicating no significant decline in outer aponeurosis volume occurred during the detraining period (Figure 2). In eccentric training, small positive but nonsignificant after‐to‐detraining changes in aponeurosis volume were observed, with increases of +5.2% for BFlh (95% CI [−0.4%, 10.8%]; d = 0.33), +3.9% for BFsh (95% CI [−4.0%, 11.8%]; d = 0.17), +4.1% for SM (95% CI [−3.5%, 11.6%]; d = 0.19), and +4.6% for ST (95% CI [0.3%, 8.9%]; d = 0.23). The stretching condition demonstrated a similar response, with after‐to‐detraining changes of +0.6% for BFlh (95% CI [−3.0%, 4.3%]; d = 0.14), +2.9% for BFsh (95% CI [−5.6%, 11.3%]; d = 0.20), +4.7% for SM (95% CI [−2.7%, 12.1%]; d = 0.24), and +0.6% for ST (95% CI [−4.1%, 5.3%]; d = 0.01).

3.3. Inner‐Aponeurosis

3.3.1. Volume

Figure 3 presents individual participant data for hamstring inner‐aponeurosis volume across the Before, After, and Detraining periods in both the eccentric and stretching interventions. A significant main effect of Time was identified (F _(2,110) = 6.85, p = 0.002, ${\eta }_p^2$ = 0.11), indicating that inner‐aponeurosis volume increased following training. A significant main effect of Muscle (F _(1,110) = 138.39, p < 0.001, ${\eta }_p^2$ = 0.56) was also present, with the SM showing greater inner‐aponeurosis volume than BFlh. The main effect of Exercise Condition was not significant (F _(1,10.11) = 1.46, p = 0.254, ${\eta }_p^2$ = 0.13), and none of the interactions reached significance (all p ≥ 0.242, ${\eta }_p^2$ ≤ 0.009).

FIGURE 3.

Upper panels: Individual data points for inner aponeurosis volumes (left panel) and width (right panels) of Biceps Femoris long head (BFlh) and Semimembranosus (SM) at Before, After, and Detraining period (n = 11). Each participant is shown with connected points to illustrate their progression over time. Colors indicate the eccentric‐training (blue) and static stretching conditions (magenta). Lower panels: Before‐to‐After (light colors) and After‐to‐Detraining (dark colors) change scores for the same muscles. Each point shows an individual participant's change, and the overlaid markers with error bars show the group mean and its 95% confidence interval. Significant increases following both interventions were found for the inner aponeurosis volume and width (p < 0.05). Similar changes between the two interventions were found (p > 0.05). No significant differences were found between the post and detraining period (p > 0.05). *Different from before training (p < 0.05).

Eccentric training increased inner‐aponeurosis volume in BFlh by +10.8% (95% CI [2.0%, 19.6%]; d = 0.33) and also a similar response was found in SM (+21.8%, 95% CI [8.50%, 35.1%]; d = 0.40). Static stretching also induced significant increases of inner‐aponeurosis volume in BFlh by +23.3% (95% CI [10.7%, 35.9%]; d = 0.63) and in SM by +9.1% (95% CI [1.9%, 16.3%]; d = 0.53).

During detraining, inner‐aponeurosis volume showed no evidence of reduction in either intervention. In the eccentric limb, BFlh volume changed by +0.6% (95% CI [−10.5%, 11.6%]; d = 0.08), and SM changed by +4.2% (95% CI [−5.9%, 14.3%]; d = 0.01). The stretching intervention showed a comparable pattern, with BFlh increasing by +8.8% (95% CI [1.00%, 16.7%]; d = 0.51), and SM increasing by +12.9% (95% CI [1.9%, 23.9%]; d = 0.51).

3.3.2. Width

Figure 3 depicts individual participant data for hamstring inner‐aponeurosis width across the Before, After, and Detraining periods in both the eccentric and stretching interventions. The analysis showed a significant main effect of Time (F _(2,110) = 5.00, p = 0.008, ${\eta }_p^2$ = 0.08) and Muscle (F _(1,110) = 176.92, p < 0.001, ${\eta }_p^2$ = 0.62), with SM presenting greater inner‐aponeurosis width than BFlh. The main effect of Exercise Condition was not significant (F _(1,72.38) = 0.31, p = 0.577, ${\eta }_p^2$ = 0.004), and all interactions were also nonsignificant (all p ≥ 0.296, ${\eta }_p^2$ ≤ 0.012).

Eccentric training increased inner‐aponeurosis width in BFlh by +7.1% (95% CI [−1.1%, 15.1%]; d = 0.55). While in SM by +12.9% (95% CI [−2.2%, 27.9%]; d = 0.55). Static stretching induced similar improvements, with inner‐aponeurosis width of BFlh increased by +19.6% (95% CI [4.6%, 34.5%]; d = 0.86), while for the SM by +8.9% (95% CI [−5.0%, 22.9%]; d = 0.37).

After‐to‐detraining changes in inner‐aponeurosis width were small and variable, with 95% CIs crossing zero in both interventions. In the eccentric condition, BFlh changed by −0.9% (95% CI [−11.2%, 9.4%]; d = −0.06) and SM by +2.3% (95% CI [−5.3%, 9.8%]; d = 0.20). In the stretching condition, BFlh changed by −2.8% (95% CI [−9.2%, 3.7%]; d = −0.29) and SM by +2.3% (95% CI [−3.8%, 8.5%]; d = 0.26).

3.4. Muscle Volume

Individual participant data for hamstring muscle volume across the Before, After, and Detraining periods in both the eccentric and stretching interventions are displayed in Figure 4. Analysis indicated significant main effects of Muscle (F _(3,230) = 376.89, p < 0.001, ${\eta }_p^2$ = 0.83) and Time (F _(2,230) = 35.15, p < 0.001, ${\eta }_p^2$ = 0.23), as well as a significant main effect of Exercise Condition (F _(1,35.77) = 35.71, p < 0.001, ${\eta }_p^2$ = 0.50) for muscle volume. Significant Muscle × Time (F _(6,230) = 2.47, p = 0.025, ${\eta }_p^2$ = 0.06) and Exercise Condition × Time (F _(2,230) = 13.22, p = 0.0006, ${\eta }_p^2$ = 0.10) interactions were also found. Post hoc tests indicated that eccentric training increased significantly (p < 0.05) the muscle volume of all biarticular hamstring muscles (ST, SM, and BFlh) while no changes were found for BFsh (p > 0.05). Static stretching did not induce any significant changes in the muscle volume of the hamstring muscles (p > 0.05). Also, significant reductions were observed in muscle volume in all biarticular hamstring muscles from after‐to‐detraining period after eccentric exercise while no significant differences were identified for the stretching interventions.

FIGURE 4.

Upper panels: Individual data points for muscle volumes of Biceps Femoris long (BFlh), short head (BFsh), Semitendinosus (ST), and Semimembranosus (SM) at Before, After, and Detraining period (n = 11). Each participant is shown with connected points to illustrate their progression over time. Colors indicate the eccentric‐training (blue), static stretching (magenta) conditions. Lower panels: Before‐to‐After (light colors) and After‐to‐Detraining (dark colors) change scores for the same muscles. Each point shows an individual participant's change, and the overlaid markers with error bars show the group mean and its 95% confidence interval. Significant increases following both eccentric training was found for the muscle volume in all biarticular (ST, SM and BFlh) hamstrings muscles (p < 0.05). Significant decrease in muscle volume was found following detraining (p < 0.05). *Different from before training (p < 0.05); ^#Different from the static stretching training (p < 0.05); ^Different from after training (p < 0.05).

The corresponding descriptive changes showed that eccentric training increased muscle volume in BFlh (+29.6%, 95% CI [25.1%, 34.2%]; d = 1.49), SM (+22.0%, 95% CI [16.2%, 27.8%]; d = 0.98), ST (+42.4%, 95% CI [38.2%, 46.6%]; d = 1.65), and BFsh (+24.7%, 95% CI [19.1%, 30.4%]; d = 1.06). Static stretching produced smaller positive increases of muscle volume of BFlh (+3.9%, 95% CI [−0.1%, 8.0%]; d = 0.18) and BFsh (+2.4%, 95% CI [−2.0%, 6.8%]; d = 0.12). Further, nonsignificant increases were also found for SM (+7.3%, 95% CI [2.9%, 11.7%]; d = 0.37) and ST (+9.4%, 95% CI [5.1%, 13.7%]; d = 0.33). During the detraining period following eccentric training, model‐based post hoc comparisons showed significant reductions in muscle volume in BFlh (−17.9%, 95% CI [−24.9%, −11.1%]; d = −1.74), SM (−15.3%, 95% CI [−21.5%, −9.1%]; d = −1.20), and ST (−13.6%, 95% CI [−19.4%, −7.8%]; d = −1.10). BFsh also showed a relatively large descriptive reduction (−15.4%, 95% CI [−18.2%, −12.7%]; d = −1.50), but the corresponding model‐based post hoc comparison did not reach statistical significance. This discrepancy should be interpreted cautiously given the modest sample size and variability within the repeated‐measures analysis. After detraining, muscle volume in the biarticular hamstrings remained above baseline following eccentric training (all p < 0.05).

In the stretching condition, after‐to‐detraining descriptive changes were observed in all muscles, with reductions of −11.1% in BFlh (95% CI [−16.1%, −6.1%]; d = −0.85), −10.5% in BFsh (95% CI [−13.5%, −7.5%]; d = −0.80), −11.3% in SM (95% CI [−18.2%, −4.3%]; d = −0.70), and −5.3% in ST (95% CI [−9.8%, −0.9%]; d = −0.40). However, only the SM reduction reached significance in the model‐based post hoc comparisons.

3.4.1. Before‐To‐After Training Changes Association

Across all muscles, the eccentric intervention produced trivial to moderate and nonsignificant correlations between changes in muscle volume and aponeurosis measures (−0.28 to 0.47, p ≥ 0.15) (Figures 5 and 6 and Appendix S4). The static stretching intervention showed a similar pattern, with nonsignificant trivial to weak associations (−0.30 to 0.38, p ≥ 0.25), with two exceptions. In the stretching‐trained leg, a strong significant negative correlation was identified between change in BFlh muscle volume with change in inner‐aponeurosis width (r = −0.70, 95% CI [−0.92, −0.17], p = 0.017), whereas a significant strong positive correlation between change in muscle volume and change in inner‐aponeurosis width was observed in SM (r = 0.77, 95% CI [0.32, 0.94], p = 0.006) (Figure 6).

FIGURE 5.

Association between training‐induced changes in hamstring muscle volume and outer aponeurosis volume. Scatterplots show individual participant percent change from Before‐ to after‐intervention (10 weeks) for eccentric knee‐flexor training (top row, blue) and time‐matched static stretching (bottom row, magenta) in biceps femoris long head (LH), biceps femoris short head (SH), semitendinosus and semimembranosus. Each data point corresponds to one participant (n = 11). No significant associations were found between changes in muscle volume and outer aponeurosis volume (p > 0.05).

FIGURE 6.

Association between training‐induced changes in hamstring muscle volume and inner aponeurosis volume and width. Scatterplots show individual participant percent change from Before‐ to After‐intervention (10 weeks) for eccentric knee‐flexor training (top row, blue) and time‐matched static stretching (bottom row, magenta) in biceps femoris long head (LH), and semimembranosus. Least‐squares regression lines are displayed only when the correlation was significant (p < 0.05). Each data point corresponds to one participant (n = 11). Significant associations were only observed between changes in hamstring muscle volume and inner aponeurosis width in the stretching‐trained leg.

3.4.2. After–to‐Detraining Changes Association

During detraining, correlations between changes in muscle volume and aponeurosis measures were trivial to moderate and nonsignificant across all muscles and both exercise conditions (r = −0.45 to 0.21, p ≥ 0.16) (Appendix S4).

4. Discussion

Aponeurosis dimensions increased in both conditions, with broadly comparable responses between eccentric training and stretching. Eccentric training increased biarticular hamstring (BFlh, SM, and ST) muscle volume, whereas static stretching did not. During detraining, aponeurosis volumes and widths were maintained in both legs, while muscle volume in the eccentrically trained leg decreased but remained above baseline. In the stretching leg, muscle volume was largely unchanged, except for a reduction in SM. The near absence of significant associations between changes in muscle volume and aponeurosis dimensions suggests that these tissues adapt largely independently.

4.1. Differential Tissue Adaptations to Eccentric Training and Static Stretching

Partially in line with our primary hypothesis, eccentric training led to hypertrophy of the biarticular hamstring muscles while no differences were found after the static stretching (Figure 4). Surprisingly, despite these differences in muscle size adaptations, both interventions increased aponeurosis dimensions to a similar extent with static stretching inducing slightly greater changes in the outer aponeurosis volume (Figures 2 and 3). These findings align with previous work showing aponeurosis width adaptations following strength training (3 times per week) [19, 23, 31], but differ from recent evidence reporting no aponeurosis adaptations (2‐times per week) [18]. These discrepancies may partly relate to differences in training frequency, which could lead to greater cumulative loading, although this interpretation remains speculative given the limited number of available studies. Also, the muscle × time interaction was not significant for both inner and outer aponeurosis size adaptations, suggesting a similar remodeling across hamstrings aponeurosis. This does not support that the reduced aponeurosis hypertrophic capacity (size) in BFlh as an explanation for its higher strain‐injury incidence [3], pointing instead to other mechanisms not captured by aponeurosis size (such as regional strain concentrations), which warrant direct verification in future studies.

Eccentric loading provides a distinct mechanical stimulus to the aponeurosis through two complementary mechanisms. First, high‐force lengthening (i.e., eccentric contractions) generates steep local strain gradients at the fiber‐aponeurosis junction, particularly in anatomically narrow regions [4, 32, 33]. Repeated exposure to these localized strain concentrations may stimulate aponeurosis widening, with a wider aponeurosis is expected to reduce peak stress and improve load distribution along the muscle–tendon interface [4, 14, 34]. This may be advantageous because distributing strain over a broader aponeurotic insertion area could reduce local strain concentration near the MTJ, although this interpretation is based on modeling evidence and was not directly tested in the present study. Second, eccentric contractions impose force‐dependent biaxial loading on the aponeurosis. Actively contracting muscle fibers behave near‐isovolumetrically and bulge under high tension, generating transverse stresses within the aponeurotic sheets [11, 12]. Repeated exposure to this combined longitudinal and transverse loading may further promote structural expansion of the aponeurosis.

In contrast, as no studies have directly examined how static stretching affects aponeurosis morphology, the underlying mechanisms remain uncertain and only tentative explanations can be provided. Aponeuroses and tendons respond to long‐duration, low‐load strain, which stimulates fibroblast activity, collagen I synthesis, and extracellular matrix turnover even when forces are modest [20, 35]. In this study, even though the total duration between the two interventions was matched, the cumulative time spent at long muscle lengths during stretching exceeded that of the eccentric training, providing a potent duration‐dependent stimulus for matrix remodeling. Stretching also exposes the aponeurosis to uniaxial mechanical loading in contrast to active contraction that promotes biaxial loading [11]. As muscle fibers lengthen, they generate shear and transverse stresses across the connective tissue sheets potentially promoting collagen realignment and structural expansion [11, 16]. These stresses are expected to propagate into deeper tissue layers (such as inner aponeurosis or intramuscular tendons), which is consistent with the observed adaptations in both outer (Figure 2) and inner aponeurosis dimensions (Figure 3). In addition, prolonged stretching may increase intramuscular hydrostatic pressure and induce creep‐like behavior in collagen networks [36, 37]. Sustained tension, together with transient reductions in perfusion and fluid redistribution, could promote microstructural remodeling through disruption and subsequent reformation of collagen cross‐links [36, 37, 38]. Taken together, these findings indicate that aponeurosis hypertrophy can be induced by both high‐force, short‐duration active lengthening and low‐force, long‐duration passive lengthening, whereas muscle hypertrophy appears to depend primarily on the former.

4.2. Detraining Responses of Aponeurosis vs. Muscle

After the detraining period, aponeurosis volumes and width remained unchanged in either intervention (Figures 2 and 3), which is partly in line with our second hypothesis. In contrast, hamstring muscle volume in the eccentrically trained leg decreased by ~14%–18%, although the values remained above baseline, whereas in the stretching leg only the SM showed a significant reduction in muscle volume (Figure 4). These patterns are consistent with previous reports showing partial loss of muscle volume after eccentric loading cessation (3–4 weeks), even though BFlh and SM volume remained above pretraining levels [21, 22].

To date, it remains unclear how the aponeurosis responds to detraining after an intervention [16, 17]. The present findings suggest that its behavior differs from muscle, which is expected given the distinct turnover rates of contractile and connective tissues [39, 40]. Muscle protein synthesis and breakdown adapt quickly to changes in loading, leading to significant declines in CSA and volume within weeks of detraining [41, 42]. In contrast, collagen turnover in tendon and aponeurosis is considerably slower, and any loss of connective tissue structure tends to occur over a longer period [39, 43]. Our results are consistent with this interpretation, suggesting that once aponeurosis hypertrophy was induced after training, its dimensions were maintained during the 4 week detraining period. However, given the variability in the change scores, small changes cannot be excluded. Thus, these findings should be interpreted as evidence of short‐term maintenance only, and further studies are needed to determine whether aponeurosis dimensions regress after longer periods without training.

4.3. Limited Association Between Muscle and Aponeurosis Remodeling

Contrary to our third hypothesis, changes in muscle volume were not significantly associated with changes in aponeurosis dimensions in most of the cases, indicating that these two tissues did not remodel in parallel (Figures 5 and 6, Appendix S4). This differs from the findings of Wakahara et al. [23], who reported a moderate association between vastus lateralis CSA and aponeurosis width following resistance training. However, recent studies show different tissue adaptation patterns in the hamstrings. Lazarczuk et al. [18] observed significant muscle hypertrophy without accompanying aponeurosis remodeling, whereas Maeo et al. [19] reported increases in both muscle and aponeurosis size but did not examine whether these changes were associated. Together, these studies highlight that aponeurosis remodeling can occur with or without substantial muscle hypertrophy, and vice versa.

Several factors likely explain these differential adaptations. First, the aponeurosis appears to respond mainly to cumulative mechanical deformation, whereas muscle hypertrophy depends primarily on high active tension. These stimuli are not necessarily aligned, particularly when comparing passive static stretching with active lengthening eccentric contractions. Consistent with this notion, some studies report increases in muscle size after static stretching [44], whereas others show minimal or no changes [45] which is in line with our findings. Second, aponeurosis strain is influenced by its own geometry and regional stiffness [46, 47], which determine local strain amplification largely independent of whole‐muscle size [4, 14]. Modeling studies demonstrate that local strain concentrations within the aponeurosis are influenced more by tissue morphology than by global muscle hypertrophy [4, 14]. Third, geometric models of pennate muscles indicate that muscle volume reflects both aponeurosis dimensions and fascicle architecture (e.g., fascicle length and pennation angle). Thus, architectural remodeling could weaken or obscure any direct coupling between changes in muscle volume and aponeurosis size [17, 23]. Fourth, this dissociation may reflect tissue‐specific remodeling kinetics, particularly during detraining, because contractile tissue can respond more rapidly to changes in loading than the extracellular matrix‐rich aponeurosis [41, 42]. Finally, hamstring aponeuroses exhibit substantial interindividual variability, and these parameters are only weakly related to muscle size [3, 5, 6]. This variability further reduces the likelihood of a strong, uniform coupling between muscle hypertrophy and aponeurosis remodeling.

4.4. Methodological Considerations

Despite the advantages of a within‐participant contralateral design over traditional between‐group approaches, this methodological approach also raises the possibility of cross‐education effects. Existing evidence indicates that cross‐education is primarily neural and is mainly reflected in strength outcomes [25, 48], whereas morphological cross‐transfer is minimal or absent [25, 48]. Consistent with this, the present results showed no muscle volume adaptations in the stretching leg (Figure 4). However, because this design did not include a true no‐intervention control leg, it cannot dissociate local mechanical effects from systemic or metabolic influences, nor can it fully exclude subtle bilateral adaptations. Therefore, the findings should be interpreted as differences between two loading stimuli applied within the same systemic environment, rather than as evidence of exclusively local adaptations. Nevertheless, the different responses between tissues and interventions, with muscle hypertrophy occurring mainly after eccentric training whereas aponeurosis adaptations occurred after both eccentric training and stretching, suggest that the findings are unlikely to be explained solely by a generalized systemic effect.

Although the 10‐week intervention was sufficient to detect changes in aponeurosis dimensions, it remains unclear whether these adaptations had reached a plateau or would continue to remodel with longer exposure. Similarly, because the detraining phase lasted only 4 weeks, the present findings should be interpreted as evidence of short‐term maintenance rather than long‐term retention of aponeurosis hypertrophy. Longer follow‐up periods are therefore needed to determine whether aponeurosis dimensions remain stable, regress, or continue to remodel after training cessation.

A further consideration is that the two interventions were matched by time under tension [27] rather than total training volume. Volume is more strongly linked to hypertrophy [49], but our goal was not to determine which intervention would produce greater growth; instead, we aimed to examine how these tissues respond to different types of stimuli. It is possible that using different loading stimuli or equating the interventions by volume instead of time under tension could lead to different adaptations [49].

Although the MRI‐based measurements showed high between‐session reliability, the interpretation of change should also consider the MDC values reported in Appendix S2. The main training‐induced changes in aponeurosis volume were larger than the corresponding MDC values, supporting that these adaptations exceeded expected measurement error. In contrast, smaller changes, particularly for inner‐aponeurosis width and detraining‐related responses, should be interpreted more cautiously. Also, the spatial resolution of the MRI sequences does not allow a clear separation between the aponeurosis and adjacent deep fascial layers. These structures form a continuous connective tissue matrix [9, 50], and their boundaries are not always distinctly identifiable on T1‐weighted images. Consequently, the reported aponeurosis volumes likely represent the combined thickness of the aponeurosis and neighboring fascial tissues rather than the aponeurosis alone. However, the absence of consistent associations between changes in muscle volume and aponeurosis dimensions suggests that these measurements did not simply reflect generalized muscle enlargement. Also, given the reconstructed in‐plane resolution (~0.82 × 0.82 mm) and 5‐mm slice thickness, partial‐volume effects may have influenced aponeurosis volume estimates, particularly at the borders between aponeurosis, fascia, and adjacent muscle. In addition, inner‐aponeurosis volume and width were quantified only for BFlh and SM because these structures were consistently identifiable on MRI. Therefore, the inner‐aponeurosis findings should not be extrapolated to the entire hamstring group.

Although the sample size was supported by the a priori power analysis for the overall repeated‐measures effect, it was modest relative to the complexity of the design. Therefore, higher order interactions, particularly muscle‐specific Time × Exercise Condition effects, should be interpreted cautiously.

Finally, the sample consisted only of healthy young untrained men without recent lower limb musculoskeletal injury, which limits the generalizability of the findings to women, older adults, trained athletes, and clinical populations. Sex differences have been reported in hamstring morphology [51] and strain‐injury incidence [52], raising the possibility that training and detraining responses may also differ between men and women. However, previous studies comparing hamstring aponeurosis thickness between sexes found no consistent differences [53]. Whether the aponeurosis adapts similarly in women therefore remains unknown.

4.5. Practical Applications

Because a wider aponeurosis may reduce peak fascicle strain and strain gradients near the MTJ, increases in aponeurosis dimensions could represent a potentially beneficial morphological adaptation, particularly for reducing mechanical vulnerability during high‐speed running [4, 34]. This interpretation assumes that the observed increases in aponeurosis volume and width improve the distribution of tissue strain along the muscle–tendon interface. From a practical standpoint, both training modes increased hamstring aponeurosis volume, which may help distribute strain more evenly along the MTJ. Static stretching may be preferred when high‐force eccentric loading is not feasible or advisable (such as during early rehabilitation). However, eccentric knee‐flexor training should remain central to hamstring injury prevention because it increases strength and muscle volume [19, 21]. Over the 4‐week detraining period, aponeurosis dimensions showed no clear decline despite partial regression of muscle hypertrophy after eccentric training. Nevertheless, functional outcomes such as strain distribution, mechanical vulnerability, and injury incidence were not directly measured; therefore, these practical implications remain inferential. Future studies should determine how long these aponeurotic adaptations persist and whether they translate to measurable reductions in tissue strain and injury incidence.

5. Conclusion

Eccentric knee‐flexor and static stretching training increased hamstring aponeurosis dimensions, with no significant decline observed after 4 weeks of detraining. In contrast, muscle hypertrophy was mainly induced by eccentric training and partially regressed with detraining. The absence of consistent correlations between muscle and aponeurosis changes indicates largely independent remodeling. Aponeurosis hypertrophy can be induced by either eccentric loading or stretching and maintained over the 4‐week detraining period, whereas increases in muscle size are largely eccentric‐specific.

Funding

The authors have nothing to report.

Ethics Statement

This study was conducted in accordance with the ethical standards of the University ethics committee and received approval (No. 23057‐1). All procedures involving human participants were performed following ethical guidelines, ensuring confidentiality, informed consent, and compliance with relevant regulations.

Consent

Informed consent was obtained from all individual participants included in the study. Participants were provided with comprehensive information about the study objectives, procedures, potential risks, and benefits before obtaining their consent.

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Appendix S1: Supporting Information 1.

Appendix S2: Supporting Information 2.

Appendix S3: Supporting Information 3.

Appendix S4: Supporting Information 4.

Data Availability Statement

The data supporting the findings of this study are available upon reasonable request from the corresponding author.