Force-Velocity Profile of Competitive Kayakers: Evaluation of a Novel Single Kayak Stroke Test

Flatwater kayaking is an Olympic discipline where athletes are paddling with maximal effort to complete a given distance in the shortest possible time. The ability to efficiently perform the complex paddling movement depends on a number of various kayakers’ physiological and biomechanical factors (Van Someren and Howatson, 2008). It is important to note that upper body muscle mechanical capacities (force, velocity and power) have a great impact on the final results of the competition, especially during shorter distances (i.e., 200 m canoe sprint discipline) (McKean and Burkett, 2014). Different testing protocols are conducted to identify athletes' strengths and weaknesses with respect to the specific demands of a kayak sport discipline and, consequently, to prescribe training according to the individual needs (McGuigan et al., 2013). However, many physical tests do not provide valuable information to guide athletes' training programs (McKean and Burkett, 2010). Therefore, sport scientists are in a continuous quest for tests which could provide meaningful information for enhancing athletes' sport performance.

Besides regular on-water specific drills, developing and achieving optimal kayak performance inevitably include dry-land resistance training of upper body muscles (McKean and Burkett, 2014). One of the reasons for including dry-land exercises is the inability of kayakers to perform their outdoor on-water training sessions during undesirable weather conditions (i.e., low temperatures, rain, etc.). To determine strength of kayakers, basic tests such as the overhead medicine ball throw (López-Plaza et al., 2018), one-repetition maximum (1RM) tests of compound movements including bench press and prone bench pull exercises (McGuigan et al., 2013; McKean and Burkett, 2010, 2014; Ualí et al., 2012), isokinetic tests of elbow flexors (Clarkson et al., 1982), shoulder extension (Tesch, 1983), isometric tests including the hand grip (Hamano et al., 2015), and the bilateral prone bench pull and one arm cable row (Ualí et al., 2012) are used. Several attempts to develop more specific tests consisted of using simulation of the kayak stroke (i.e., to exert a pulling force in the horizontal direction involving trunk rotation, shoulder extension, and elbow flexion in one motion) with an isokinetic device (Van Someren and Howatson, 2008) and tasks under isometric conditions (Steeves et al., 2018). To date, the most specific dry-land test consists of recording maximal power during cycling movements performed on a modified kayak ergometer (kayak Wingate test) (Van Someren and Howatson, 2008). One problem is that the mentioned basic exercises bear little resemblance to the kayak stroke, while the Wingate test reveals peak power, but not the maximal force and velocity capacities. Therefore, it could be important to develop a specific test to evaluate the maximal mechanical capacities of the muscles involved in single kayak strokes.

A testing procedure which has gained popularity over the last years consists of the evaluation of the force-velocity (F-V) relationship. There is compelling evidence that the F-V relationship of multi-joint exercises follows a linear shape (Jaric, 2015). Tests described above (i.e., bench press, bench pull, isokinetic tests, etc.) are limited since they are performed under a single pre-defined movement condition, resulting in a single outcome (e.g., power output, paddling frequency, or exerted force). Consequently, distinguishing among different muscle capacities (Jaric, 2015, 2016) could not be done using the above-mentioned tests. A potential solution could be based on the approximately linear and exceptionally strong F-V relationship typically observed in various functional tasks, such as vertical jumps, the bench press, the bench pull, etc. (García-Ramos et al., 2017; Giroux et al., 2017; Zivkovic et al., 2017). Therefore, a linear regression model [F(V) =F₀ – aV] can be used to distinguish between force (F₀), velocity (V₀) and maximal power (P_max) producing capacities of the tested muscle groups (Jaric, 2015). In addition to these variables, the F-V slope (i.e., the ratio between F₀ and V₀) also provides meaningful information to design individualized training programs (Morin and Samozino, 2016). It has been shown that the F-V relationship recorded for the bench press exercise (García-Ramos et al., 2016; Sreckovic et al., 2015), vertical jumps (García-Ramos et al., 2017), as well as the cycling and arm cranking ergometer (Driss and Vandewalle, 1998) enables reliable and valid assessment of mechanical capacities of muscles acting during multi-joint tasks. However, no protocol has been developed to assess the F-V relationship of muscles involved in tasks simulating the kayak stroke.

To address the problems discussed above, we conducted a study to explore the feasibility of a novel test for assessing the muscle mechanical capacities involved in the kayak stroke. Specifically, the aims of this study were to evaluate the shape (i.e., linear or polynomial) and strength (i.e., magnitude of the r coefficient) of the F-V relationship as well as to examine the reliability and external validity of the F-V relationship parameters obtained during the single kayak stroke test (SKST). We hypothesized that (1) the F-V relationship would be highly linear, (2) the F-V relationship parameters would present acceptable reliability, and (3) the F-V relationship parameters obtained during the SKST would present high external validity with respect to the same variables obtained during traditional tests (short Wingate kayak test, bench press and prone bench pull).

Linear and 2^nd order polynomial F-V relationship models obtained from the single kayak stroke test are depicted in Figure 2. Note that all regression models revealed exceptionally high correlation coefficients (r > 0.990) and that none of the coefficients of the linear regression models were below the 95% CI of the correlation coefficients obtained from the corresponding polynomial regression models.

Figure 2

The within-session and between-session reliability of the F-V relationship parameters was high with the only exception of the between-session reliability observed for the F-V slope (CV = 12.24-13.12%; ICC = 0.28-0.40). In addition, results of the depended t test showed that there were no between- nor within-session differences (all p ≥ 0.10) (Table 1).

The F-V relationship parameters were generally highly correlated with the same variables obtained from traditional tests. Nearly perfect correlations were observed between the maximal power reached during the bench press, prone bench pull, modified Wingate test and P_max assessed from the F-V relationships during the single kayak strokes (r range = 0.825-0.975). Correlations between F₀ variables were very large (r range = 0.751-0.831), as well as V₀ for the prone bench pull (r = 0.770-0.829). Non-significant correlations were observed for the V₀ obtained during the bench press exercise (r = 0.355-0.471) (Figure 3).

Figure 3

This study aimed to explore the feasibility of a novel test for evaluating the F-V relationship of muscles involved in the single kayak stroke. The results with respect to the SKST revealed (1) high linearity of the F-V relationship, (2) acceptable reliability of the obtained F-V relationship parameters, and (3) generally high external validity of the F-V relationship parameters with respect to the same variables obtained from traditional tests (short Wingate kayak test, bench press and prone bench pull). These findings suggest that the SKST can be utilised to evaluate the muscle mechanical capacities of upper body muscles in kayakers.

To our knowledge, this is the first study which has explored the F-V profile in competitive kayakers not only during the SKST, but also during the bench press and prone bench pull exercises. The high linearity of the F-V relationship observed in the present study during the SKST is in line with our first hypothesis and it brings additional support with respect to the high linearity of the F-V relationship obtained during multi-joint exercises (Giroux et al., 2017; Jaric, 2016; Sreckovic et al., 2015; Zivkovic et al., 2017). An important point to consider is that we used a very wide range of loads (from 1 kg until the 1RM load) covering almost the entire spectrum of the F-V relationship (Figure 2). It should be noted that the F-V relationship of the single kayak stroke remains linear even when the experimental points are located very close to the force and velocity intercepts. Similar results were obtained by Riviere et al. (2017) and Cuevas-Aburto et al. (2018) during the squat and bench press exercises, respectively. Therefore, within the range of loads which can be tested during multi-joint exercises (e.g., a squat, a bench press, or a SKST), it seems that the F-V relationship follows a linear relationship.

Our second hypothesis was also confirmed since the F-V relationship parameters obtained during the kayak strokes generally revealed high reliability. Particularly, the within session reliability of all F-V relationships was high. Absolute between session reliability ranged from low (for the F-V slope) to high (for F₀ and P_max). Acceptable reliability of the F-V relationship parameters has also been reported for other multi-joint exercises such as vertical jumps (Cuk et al., 2014; García-Ramos et al., 2017), bench press (García-Ramos et al., 2016; Sreckovic et al., 2015) and the leg cycle ergometer exercise (García-Ramos et al., 2018). It is known that for obtaining the F-V relationship parameters with high reliability, the experimental points should be as close as possible to the axis intercepts (García-Ramos and Jaric, 2018). The addition of one experimental point very close to the velocity-intercept during the SKST could be responsible for the higher reliability of V₀ variable observed in the present study compared to the reliability of the same variable in exercises such as the vertical jump and the bench press (Cuevas-Aburto et al., 2018; García-Ramos et al., 2016). Regarding the somewhat lower between-session reliability for the F-V slope and V₀, there are still no data to elucidate whether the observed phenomenon originates from the tested task, from the participants’ skill, or from other factors. Nevertheless, the discussed findings generally suggest that obtained variables could be reliable indices of mechanical capacities of the muscles engaged in a single kayak stroke (Van Someren and Howatson, 2008).

Confirming our last hypothesis, the results of the external validity showed high agreement of F₀ with corresponding variables obtained from traditional exercises commonly used for testing kayakers (bench press and prone bench pull) and P_max with the maximal power reached during the modified Wingate test on a kayak ergometer (García-Pallarés et al., 2010; Papandreou et al., 2018; Van Someren and Howatson, 2008). Finally, while the external validity of V₀ obtained during the SKST was high with respect to V₀ obtained during the prone bench pull, the agreement of V₀ with the same variable from the bench press ranged from low to moderate. Since this is the first study evaluating the F-V relationship obtained from single kayak strokes, there are no data regarding agreement of F₀, V₀ and P_max with directly measured force, power and velocity obtained from other upper body performance tests. The lower correlations observed for V₀ in the bench press (pushing exercise) compared to the bench pull (pulling exercise) could be explained by different movement patterns with respect to the SKST (pulling exercise) and also by the lower reliability of V₀ obtained during the bench press exercise when light loads were not used for F-V modelling (Cuevas-Aburto et al., 2018). The high external validity of the F-V relationship parameters suggests that the SKST could be used for routine dry-land testing of kayakers as an alternative to presumably less specific tests, such as the bench pull and bench press exercise.

As a limitation of this study, it is possible that the reliability of the SKST would be reduced with less experienced kayakers and, therefore, the use of this test might be limited to kayakers with a high level of expertise. Although a custom-made machine was built for the present study, coaches interested in assessing the F-V relationship could use the commercially available “Catch Force” machine. However, the “Catch Force” machine has an inclination of 11^⍛ with respect to the horizontal plane which can negatively impact the reliability of V₀ due to the higher distance between the experimental points and the velocity-intercept (García-Ramos and Jaric, 2018). It should also be noted that the sample size was small (6 women and 8 men). This was due to the restrictive inclusion criteria regarding the level of kayakers and also to the difficulties of elite athletes to attend the laboratory on multiple occasions. Therefore, the sample size was very difficult to increase. Although multicenter reliability studies have been proposed to solve the limitation of the small sample size (García-Ramos and Janicijevic, 2020), this was not possible in the present study due to the custom-made device used to evaluate the novel single kayak stroke test. Finally, this study did not compare the results of the SKST with on-water kayak performance. Future studies should elucidate whether the SKST is able to discriminate between athletes of different competitive levels, disciplines (sprinters vs. endurance) and sexes.

In conclusion the single kayak stroke test is a feasible procedure for testing maximal upper-body mechanical capacities of kayakers. The feasibility of the single kayak stroke test is supported by the high linearity of the F-V relationship as well as by the acceptable reliability and high external validity of the F-V relationship parameters. Due to the higher face validity of the SKST, its implementation can be proposed for obtaining more meaningful information about the maximal mechanical capacities (i.e., F₀, V₀, F-V slope, and P_max) of upper-body muscles in comparison to the modified kayak Wingate test (which only provides information on peak power) or traditional resistance training exercises such as the bench press and the bench pull. However, future studies should explore the association between the F-V parameters obtained during the SKST and on-water kayaking performance.