Twenty-one elite female rowers volunteered to participate in this study (age 26 ± 4 years; height 1.81 ± 0.06 m; weight 74.2 ± 7.3 kg). All rowers were professional athletes of the national team, had 5.9 ± 2.2 years of rowing experience, and trained for at least 12 sessions each week. Participants included 19 heavyweight and two lightweight rowers. All rowers competed in international rowing regattas, 19 participated in Rowing World Championships, and 11 participated in the Olympic Games. Five participants won Olympic medals. Participants completed a 2,000-m ergometer test in 414.2 ± 10.8 s (range: 399.3–437.8 s), placing them between the 95th and 99th percentiles of the international ranking for indoor rowing in 2019 (Concept II) (Concept2 Logbook, 2019) | 2,000m Indoor Rower Rankings). Prior to participation, participants were informed about the experimental procedures of the study and provided written informed consent. The study was conducted according to the principles of the Declaration of Helsinki and was approved by the departmental ethics committee of the Vrije Universiteit, Amsterdam, the Netherlands (VCWE-2020-122).

This study was an observational non-randomized training study, in which all rowers were monitored during their general preparation phase and a subgroup of seven rowers were monitored during their competitive preparation phase (see Table 1). Testing was performed and muscle morphology was assessed the week before (pre) and in the week after general preparation (post-1) and in the week after competitive preparation in a subgroup of seven female rowers (post-2). Both training regimens were completed during preseason, when athletes prepared for the upcoming competitive season. For reasons unrelated to the study, two participants did not complete the post-1 measurement and were excluded from analysis.

During the general preparation, participants performed 8 weeks of concurrent training with emphasis on HLRT. They performed two HLRT sessions per week, with the aim to increase muscle force production by increasing the PCSA of their muscles. Participants executed resistance exercises with high load, aiming for 3–4 sets of 5–10 repetitions per set with an intensity of ~70–85% of the 1-repetition maximum. HLRT sessions included exercises such as deadlifts, squats, bench press, and unilateral leg exercises, which were all typical exercises performed by the participants in their previous resistance training program. During their competitive preparation, a subgroup of seven female rowers completed a subsequent 16 weeks of concurrent endurance and HLRT with APL, aiming to increase muscle power and longitudinal growth of their muscle fibers. In addition to a session of regular resistance training, rowers executed one session of APL per week, which consists of exercises such as counter movement jumps, aqua bag squat jumps, dumbbell broad jumps, and stretch exercises. Participants were instructed to execute these exercises in circuit, using a large range of motion and with a high velocity. Each circuit consisted of 4–6 exercises and exercises were executed with 3–5 sets of 2–10 repetitions. Note that participants performed both HLRT and APL sessions in addition to their other training, which consisted of core stability (two times a week) and row-specific endurance training sessions (8–12 times a week, see Table 1). For weekly details of the resistance training sessions, see Supplementary Table 1. Training sessions were performed on-site at the facilities of the Royal Rowing Federation and were overseen by the resistance and conditioning coach of the female rowing team.

Data are presented as individual values or as mean ± SD. Paired-sampled t-tests or, when data were not normally distributed, non-parametric Wilcoxon signed-rank tests were used to detect differences in morphological and functional adaptations after general preparation (pre–post-1) and after competitive preparation in a subgroup of rowers (post-1–post-2). Magnitude of the effects was assessed by effect size (ES) that were obtained as the mean difference divided by the pooled standard deviation. ESs were reported in standardized Cohen's d units with their 95% confidence interval, and absolute ESs were considered to be negligible (<0.20), small (0.20–0.50), medium (0.50–0.80), or large (>0.80) (Cohen, 1992). Individual training adaptations were reported, and relationships between individual adaptations in PCSA and fascicle length (expressed relative to baseline) were assessed using Pearson or Spearman correlations. Stepwise multiple regression analyses were performed to explain changes in rowing ergometer power output based on baseline values and changes in muscle morphology. Differences were considered statistically significant if p < 0.05. Tendencies were reported for p < 0.10.

Rowers improved their rowing ergometer power output by 2 ± 2% during the general preparation phase, consisting of concurrent endurance and HLRT (Table 2, Figure 1A). Rowing ergometer power output per kg body mass showed similar improvement (+2 ± 3%, ES = 0.42, P = 0.002), as no changes in body mass were observed (−0.32 ± 0.85 kg, P > 0.05, ES = −0.04). Morphological adaptations included changes in vastus lateralis muscle architecture, with no change in muscle volume and a small reduction in fascicle length (Table 2, Figures 1B–E). Effect sizes of training adaptations have been summarized in Figure 1F. Stepwise multiple regression analysis revealed that changes in rowing ergometer power output were largely explained by the baseline values of muscle volume, pennation angle, and power output (R² = 0.68, P < 0.01). When baseline power output was excluded as a predictor, the change in rowing ergometer power output tended to be explained by the changes in muscle volume and baseline values of pennation angle (R² = 0.32, P = 0.068).

During the competitive preparation phase, a subgroup of participants further improved their rowing ergometer power output by 5 ± 3% following concurrent endurance and HLRT with APL (Table 2, Figure 2A). Similarly, rowing ergometer power output per kg body mass showed a tendency to improve (+6 ± 5%, ES = 0.74, P = 0.063), since no changes in body mass were observed (−0.35 ± 1.63 kg, P > 0.05, ES = −0.03). Morphological adaptations were non-significant in this small sample of seven rowers (Table 2, Figures 2B–E), but average ES was large for fascicle length, medium for pennation angle, small for muscle volume, and negligible for PCSA (Table 2). Effect sizes of training adaptations have been summarized in Figure 2F. Data were insufficient (n = 6) to model changes in rowing ergometer power output using stepwise multiple regression analyses. In brief, results might indicate large but non-significant increases in fascicle length, and no changes in PCSA.

Rowers showed substantial differences in their individual adaptations in the general preparation phase for muscle volume (ranging from −10 to +15%), fascicle length (−21 to +8%), and PCSA (−9 to +35%). Also, the subgroup of rowers revealed different individual adaptations after competitive preparation for muscle volume (−8 to +29%,), fascicle length (−10 to +36%), and PCSA (−32 to +44%). Similarly, confidence intervals of the ESs indicated large individual differences following the general preparation phase for muscle volume (ES from −0.18 to 0.19), fascicle length (−0.86 to −0.08), and PCSA (−0.03 to 0.73) and following the competitive preparation phase for muscle volume (−0.10 to 0.74), fascicle length (−0.54 to 2.40), and PCSA (−0.99 to 0.84). In Figure 3, the individualized training adaptations are shown for fascicle length and PCSA after the general and competitive preparation phases. Results illustrate large adaptations and large individual differences in training adaptations of vastus lateralis muscle volume and architecture in elite athletes.

Figure 4 illustrates that individual adaptations in fascicle length and PCSA were strongly and inversely related, after both the general preparation and competitive preparation (r = −0.83, P < 0.01 and r = −0.95, P < 0.01, respectively). These results indicate that even though individual training adaptations may differ, changes in muscle architecture of the m. vastus lateralis are highly interrelated, both following concurrent training with a focus on HLRT and concurrent training with APL.

Although many studies have investigated the contribution of muscle strength, power, muscle activation, kinematics, or muscle fiber composition to rowing performance (Secher, 1983; Hagerman, 1984; Hofmijster et al., 2007; Lawton et al., 2011; Miarka et al., 2018; Volianitis et al., 2020), the literature on muscle morphology in rowers remains limited. Initially, studies focused on anatomical cross-sectional area (ACSA) of the quadriceps femoris, demonstrating that mid-thigh ACSA in elderly oarsman was larger compared to that of sedentary controls, and significantly related to 2000-m ergometer performance (Yoshiga et al., 2002) and to leg extension power (Asaka et al., 2010). ACSA of the m. vastus lateralis was also larger in young rowers compared to that of controls over the entire muscle length [at 30, 50, and 70%; (Ema et al., 2014)]. These authors also demonstrated that m. vastus lateralis is the largest of the quadriceps femoris muscles, and that the m. vastus lateralis comprises a larger percentage of the total quadriceps femoris volume in rowers compared to controls (Ema et al., 2014). To the best of our knowledge, there is one previous study (van der Zwaard et al., 2018b) that reported muscle morphology in terms of physiological CSA, fascicle length, pennation angle, and also muscle volume of the vastus lateralis in elite rowers. Results showed that m. vastus lateralis volume was strongly related to 2,000-m rowing ergometer performance and that its muscle architecture impacted lower body peak power (van der Zwaard et al., 2018b). In this study, muscle volume at baseline (653 ± 103 mL) was similar to that previously reported in Olympic female rowers (van der Zwaard et al., 2018b). Surprisingly, rowers in the present study displayed smaller PCSA (56.1 ± 8.3 cm²) and pennation angles (16.8 ± 2.2°), and larger fascicle lengths (11.7 ± 1.1 cm) compared to this previous study (van der Zwaard et al., 2018b), possibly due to differences in their training history. To the authors' knowledge, our study is the first presenting training-induced adaptations of whole-muscle morphology of m. vastus lateralis and performance during the general and competitive preparation in elite rowers.

Rowing ergometer power output increased by 2% during the general preparation phase, consisting of 8 weeks of concurrent training with a focus on HLRT. In a subgroup of rowers, power output further improved by 5% after competitive preparation, consisting of 16 weeks of concurrent training with a focus on APL. In a recent meta-analysis, Thiele et al. (2020) showed that recreational and subelite level rowers improved their sport-specific rowing performance (e.g., 2,000-m rowing ergometer time) following 4–9 weeks of resistance training with a small effect (ES = 0.32). Similar or even larger effect sizes were observed in elite female rowers of our study for improvements in maximal rowing ergometer power output, which is a critical determinant of 2,000-m rowing ergometer performance (Bourdin et al., 2004), during the general (ES = 0.22) and competitive preparation phases (ES = 0.52). Contrary to our expectations, these results illustrate that compared to resistance training in recreational or subelite rowers, elite rowers demonstrate similar or even larger improvements of their rowing ergometer power output after concurrent resistance and endurance training.

Muscle volume remained unchanged during the general preparation phase with concurrent training focused on HLRT, whereas vastus lateralis fascicle length decreased (−5 ± 8%, ES = −0.47). Surprisingly, PCSA did not change, which was likely due to the large individual differences in muscle hypertrophy (+6 ± 14%, ES = 0.35). HLRT is known to increase muscle size (ACSA, PCSA, and muscle volume) after 8–14 weeks of training, with muscle fiber hypertrophy as primary underlying adaptation (Abe et al., 2000; Aagaard et al., 2001; Folland and Williams, 2007). Muscle fiber hypertrophy is observed in all fiber types following 8 weeks of heavy resistance training, both in men and in women (Staron et al., 1994). Furthermore, men and women show similar relative improvements in hypertrophy following lower body HLRT (Abe et al., 2000; Folland and Williams, 2007). Recently, it was observed that resistance training with high loads induced similar whole-muscle hypertrophy (+8.3% increase, ES = 0.53) compared to training with low loads [+7.0% increase, ES = 0.42; (Schoenfeld et al., 2017)]. Although these adaptations are slightly larger than the ES we observed for PCSA (+6%, ES = 0.35), this was not unexpected, as this meta-analysis predominantly included studies with untrained individuals (Schoenfeld et al., 2017). It remains unknown whether training with high vs. low loads will evoke similar adaptations in elite athletes and whether training regimens may lead to different adaptations at the muscle fiber level, such as type II muscle hypertrophy with HLRT and type I muscle fiber hypertrophy in training with low loads (Grgic and Schoenfeld, 2018). In addition, our results show that vastus lateralis fascicle length decreased during 8 weeks of concurrent training with HLRT (−5 ± 8%, ES = −0.47), contrasting previous observations with 10 weeks of concentric HLRT in recreationally active participants [+4–7%; (Blazevich et al., 2007; Franchi et al., 2014; Trezise and Blazevich, 2019)]. Although this explains why muscle volume was not increased in elite rowers of our study, further investigation should elucidate whether this adaptation in fascicle length is distinct for elite athletes. Potentially, the number of sarcomeres in series is reduced to accommodate a shift in optimum knee angle toward extension, such that sarcomere optimum length on the active length-tension curve occurs at optimum joint angle (Herring et al., 1984; Bobbert et al., 2020). In brief, here we show that 8 weeks of general preparation (concurrent training focused on HLRT) did not change muscle volume but reduced fascicle length and resulted in a 2% improvement in rowing ergometer power output.

A subgroup of rowers showed small but non-significant increases in muscle volume after competitive preparation consisting of concurrent training focused on APL, together with large but non-significant increases in fascicle length (+13 ± 19%, ES = 0.93) and no changes in PCSA. The large average ES for fascicle length likely relates to stretching of fascicles that occurs during the rapid eccentric contractions in the stretch-shortening cycles involved in the APL (Grgic et al., 2020; Monti et al., 2020). Yet, morphological adaptations did not reach statistical significance, likely due to the small sample (n = 7). Previous literature showed that fascicle length increased rapidly after 5 weeks of eccentric training in humans (+3%), with no further increases at 10 weeks (Blazevich et al., 2007). In young men, such increases in fascicle length have still been observed after 10 weeks of eccentric training [+12%; (Franchi et al., 2014)]. Few studies have examined muscle architecture following plyometric training. However, recently, time course changes in fascicle length with plyometric training were studied in healthy men (Monti et al., 2020). It was shown that fascicle length changes could be observed already after 2 weeks including six sessions of plyometric training (+2.2%), with further increases after 4 (+4%) and 6 weeks [+4.4%; (Monti et al., 2020)]. Our results show relatively large but nonsignificant increases in fascicle length (+13%) after 16 weeks of concurrent training with APL, similar to the changes observed with 10 weeks of accentuated eccentric loading in resistance-trained men [+14%; (Walker et al., 2020)]. It remains to be determined whether the time course of fascicle length adaptations in elite athletes is similar to that of healthy men. A recent meta-analysis indicated that plyometric and resistance training interventions may produce similar effects on whole-muscle hypertrophy in untrained individuals (Grgic et al., 2020). However, in this study, PCSA did not change with concurrent training focused on APL (ES = −0.07), illustrating that radial muscle hypertrophy was negligible in elite athletes. Similarly, it was previously observed that pennation angle increased in untrained individuals with plyometric [+6%; (Monti et al., 2020)] and eccentric training +5–21%; (Blazevich et al., 2007; Franchi et al., 2014)], however, pennation angle decreased in the rowers in our study after concurrent training with APL (−2°, −9 ± 15%, ES = −0.71), similar to observations in elite athletes after sprint and jump training [−3°; (Blazevich et al., 2003)]. Therefore, adaptations in muscle morphology are likely different in elite athletes compared to untrained individuals. Our findings show that 16 weeks of competitive preparation (concurrent training with APL) may have induced longitudinal fascicle growth instead of radial hypertrophy. Also, rowing ergometer power output was improved by 5 ± 3%.

Adaptations in fascicle length could have mechanical and metabolic implications. Changes in fascicle length underpin changes in torque–angle relationships (Blazevich et al., 2007) and attribute to a shift in optimum knee angle toward flexion in simulated isokinetic sprint cycling (Bobbert et al., 2020). Here, longer fascicles also shift optimal power production toward higher pedaling rates and interestingly lead to higher peak power production (Bobbert et al., 2020). This seems to be explained by a better alignment with the peak power vs. pedaling rate curves of the other leg muscles (Bobbert et al., 2020). The same may apply to rowers, who aim to produce high power outputs while attaining high stroke frequencies. Importantly, long fascicles are also advantageous for athletes that need to combine a well-developed sprint and endurance performance (van der Zwaard et al., 2018a,b), such as rowers. Likely, this is because longer fascicles, in contrast to a larger PCSA, do not negatively affect the diffusion distance of the muscle fibers, hampering the balance between oxygen supply and demand in the muscle (van der Zwaard et al., 2018a). In the late 2000s, it has been suggested that adaptations in fascicle length are most likely influenced by training range of motion rather than contraction velocity or contraction mode (Blazevich et al., 2007). Recently, it was confirmed that longitudinal fascicle growth seems to be independent of lengthening velocity (Marzilger et al., 2020). For contraction mode, however, eccentric and concentric training show distinct differences in adaptations of fascicle length when training load is matched for their concentric or eccentric 1-repetition maximum [+12 vs. +5%; (Franchi et al., 2014)]. Mechanical stretch is likely an essential stimulus for longitudinal fascicle growth, but may be less prominent in plyometric training compared to the stretch observed with eccentric training (Monti et al., 2020). We suggest future studies to investigate the time course and critical stimuli for training-induced longitudinal and radial muscle growth, also in elite athletes.

In elite female rowers, we demonstrated considerable individual differences in training adaptations of vastus lateralis muscle volume and architecture during general and competitive preparation. To the authors' knowledge, only one study exists presenting individual differences in fascicle length adaptations in resistance-trained men (Walker et al., 2020). These authors observed that changes in fascicle length differed considerably between participants after traditional resistance training (~−21 to +16%) and accentuated eccentric loading [~−7 to +31%; (Walker et al., 2020)], reporting similar ranges compared to those of the elite rowers in our study after concurrent training with HLRT (−21 to +8%) and concurrent training with APL (−10 to +36%), respectively. In addition, this study also demonstrates that training-induced adaptations in muscle morphology can largely differ between individual female rowers, following general and also competitive preparation. Overall, the ESs for adaptations in muscle morphology ranged from largely negative to largely positive (−0.99 to 2.40). Interestingly, we also discovered strong negative relationships between percentual changes in PCSA and fascicle length, which were observed based on individual adaptations both after general (r = −0.83) and competitive preparation (r = −0.95). These results contribute to our understanding of individualized training adaptations in elite female athletes and highlight that similar training stimuli may evoke remarkable differences in training-induced morphological adaptations, even in elite athletes. Therefore, researchers and coaches are encouraged to finetune and monitor training progression of the individual athletes and investigate the specificity of individual training adaptations in more detail.

Resistance and endurance training research is conducted by many exercise scientists and has a profound impact on the field of sports science (Kraemer et al., 2017; van der Zwaard et al., 2020). In this study, elite rowers were monitored during their general preparation phase consisting of concurrent training focused on HLRT and during the competitive preparation phase consisting of concurrent training with a focus on APL. Although scientists are encouraged to study the effects of (resistance) training in different populations (Kraemer et al., 2017), research on the effects of high-load resistance or plyometric training in elite athletes remains limited (Schoenfeld et al., 2017; Grgic et al., 2020). In rowing, a recent meta-analysis (Thiele et al., 2020) revealed only two studies that investigated resistance training in elite rowers (Lawton et al., 2012, 2013). However, training-induced changes in muscle morphology of elite rowers have not yet been reported. Our study shows that during their (competitive) preparation, elite athletes are able to largely alter their muscle morphology, while improving their rowing ergometer power output. The magnitudes of training-induced changes in PCSA after general preparation and in fascicle length after competitive preparation in the female athletes in our study were shown to be remarkably similar to those previously observed after high-load resistance and plyometric training in untrained and trained men (Franchi et al., 2014; Schoenfeld et al., 2017; Walker et al., 2020), which is in line with previous suggestions (Folland and Williams, 2007). However, unlike untrained individuals, these changes came at the cost of a lower fascicle length and a decreased pennation angle, respectively, suggesting that adaptations in muscle morphology may be more tightly regulated in elite athletes.

It is commonly acknowledged that insight into training adaptations of elite athletes is largely built upon empiric studies, which is an inherent limitation of performing experimental training studies with elite athletes (Volianitis et al., 2020). This study is no exception in this respect, with athletes performing concurrent training regimens during their general and competitive preparation. While our study was performed under highly ecologically valid conditions, the experimental design of the study did not allow for inclusion of a control group. Therefore, we cannot rule out that there may be other factors beyond our control that influenced the changes in muscle morphology and rowing ergometer power output in this study.

Another limitation of studying elite athletes is that there are only a few of them. Recruiting large samples of participants in training studies poses inherent difficulties already in untrained populations (Grgic et al., 2020), let alone in elite athletes. Our sample size was similar to the 10–22 athletes reported in previous training studies with elite rowers (Thiele et al., 2020). However, results of the subgroup of seven rowers who performed concurrent training with APL during the competitive preparation did not reach statistical significance, presumably because the statistical power was too low to draw inferences. This was confirmed by sample size calculations based on the current data, illustrating that sufficient statistical power was obtained for the general preparation with 19 participants (detecting morphological adaptations with ESs ≥0.35) and for the competitive preparation with 22 participants (detecting morphological adaptations with ESs ≥ 0.32). Therefore, findings support only preliminary conclusions and warrant more studies to confirm the estimation of training effects.