The Effect of a Coordinative Training in Young Swimmers' Performance

is study investigated the effects of a coordinative in-water training. Total 26 young swimmers (16 boys) were divided in a training group (that performed two sets of 6 25-m front crawl, with manipulated speed and stroke frequency, two/week for eight weeks) and a control group. At the beginning and end of the training period, swimmers performed 50-m front crawl sprints recorded by seven land and six underwater Qualisys cameras. A linear mixed model regression was applied to investigate the training effects adjusted for sex. Differences between sex were registered in terms of speed, stroke length, and stroke index, highlighting that an adjustment for sex should be made in the subsequent analysis. Between moments, differences were noticed in coordinative variables (higher time spent in anti-phase and push, and lower out-of-phase and recovery for training group) and differences between sex were noticed in

adjusted for sex. Differences between sex were registered in terms of speed, stroke length, and stroke index, highlighting that an adjustment for sex should be made in the subsequent analysis. Between moments, differences were noticed in coordinative variables (higher time spent in anti-phase and push, and lower out-of-phase and recovery for training group) and differences between sex were noticed in performance (stroke length and stroke index). Interactions (group * time) were found for the continuous relative phase, speed, stroke length, and stroke index. The sessions exerted a greater (indirect) in uence on performance than on coordinative variables, thus, more sessions may be needed for a better understanding of coordinative changes since our swimmers, although not experts, are no longer in the early learning stages. Keywords:coordination; youth; motor adaptability; ecological dynamics; biomechanics 1. Introduction Performance of locomotor tasks, such as walking, running, or swimming, requires the coordination of spatiotemporal patterns of upper and lower limbs muscle activity [1]. Following the dynamical system approach, coordination will be the result of the interac- tion among organismic (physical, psychological, morphological, and physiological), task (speci c to the task to perform and related to the goal and rules that governing the task), and environment (those that are external to the movement system as light, temperature, or altitude) constraints [2]. Therefore, to achieve a certain goal or performance, a constant management of those constraints must be carried out, as they limit the performers' action. Therefore, and within the ecological dynamics' framework, there is no ideal motor coordi- nation solution towards which all learners should aspire, but rather functional coordination patterns that arise from a self-organization [2–5]. During sport, if more functional movement patterns emerge as a result of movement variability due to the constant interacting constraints, a great movement exploration could be provided, allowing the performer to search for more varied and effective movement solutions to t task dynamics [6]. It has been argued that presenting the relevant constraints during the different skill development phases is the key for learners acquiring functional Int. J. Environ. Res. Public Health2022,19, 7020.

a great movement exploration could be provided, allowing the performer to search for more varied and effective movement solutions to t task dynamics [6]. It has been argued that presenting the relevant constraints during the different skill development phases is the key for learners acquiring functional Int. J. Environ. Res. Public Health2022,19, 7020.

Int. J. Environ. Res. Public Health2022,19, 7020 2 of 14 movement behaviour [6,7]. In fact, through the analysis of different constraint-led ap- proach in swimming, it was observed that changes in behaviour raised during action e.g., [8–10], but uctuations in the movement patterns themselves may or may not be functional [6], increasing the constant importance of manipulating constraints during the learning process [5]. In swimming, especially in the front crawl technique, different constraints have been analysed to understand how they in uence coordination changes. Regarding environ- mental constraints, studies have been analysing different speeds e.g. [11–14] and when using some equipment as paddles, swim suits, and parachutes (e.g., [15–17]), meaning that swimmers have to face different drag magnitudes. Using task constraints, the in uence of breathing action was studied [18–20]. However, it is with organismic constraints that more studies have been done, using characteristics or variables related (directly or indirectly) to physical (stroke frequency—SF—and stroke length—SL; e.g., [21,22]), physiological (fatigue and energy cost—e.g., [23,24]) and morphological (sex—e.g., [9,25]) aspects. In the above-mentioned studies, only one conducted a longitudinal analysis, but with adult swimmers [20], there is an existing lack of studies in young swimmers, who are still in the learning process. In fact, it was found that age changes the physical capacities and, therefore, the athletes' performance considerably [26], and, after the puberty period, slight coordination increases occur [27]. Nevertheless, it is still dif cult to determine the best age for motor learning, but the predispositions seem best up to early adulthood [27]. There is no coordination related interventional studies in young front crawl swimmers and only few exist that characterise young swimmers' patterns analysing characteristics or variables related (directly or indirectly) to physical (SF and SL) and morphological (anthropometry, maturation, and sex) aspects [10,28–31]. In a continuous movement conducted only with ngers, it was observed that when movement frequency increases (and, consequently, speed), the coordination mode becomes unstable, only to be replaced by another stable mode (e.g., [32]). In fact, this in uence was also observed in adult studies when swimming front crawl, suggesting that speed and SF were

maturation, and sex) aspects [10,28–31]. In a continuous movement conducted only with ngers, it was observed that when movement frequency increases (and, consequently, speed), the coordination mode becomes unstable, only to be replaced by another stable mode (e.g., [32]). In fact, this in uence was also observed in adult studies when swimming front crawl, suggesting that speed and SF were the major in uencing factors on changes in front crawl swimming coordination. The aim of this study was to develop an eight-weeks coordinative training in young swimmers to investigate if more stable modes emerged at the fastest front crawl race (50-m). It was hypothesized that swimmers included in the coordinative training expressed a great coupled between upper limbs after the intervention, comparing to control group. 2. Materials and Methods 2.1. Participants Total 26 young swimmers (16 boys), free from injury and training, at least, six times/week, participated in the current study. Participants were divided into two groups. Considering that they belonged from two different teams, one team was considered the control (CG) and the other team, the in-water coordinative training group (TG). Neverthe- less, the entire sample compete at the same level and their training characteristics were similar. To be included in this study, swimmers should have participated in at least 70% of complementary training sessions (11 sessions) and in the two measuring moments. Follow- ing this, three (all boys) and four swimmers (three boys and one girl) were excluded from the CG and TG, respectively. Table characteristics of each group separated by sex. The local ethics committee approved the procedures and all the swimmers' parents signed a consent form in which the protocol was explained. In addition, a maturation evaluation was accomplished, concluding that all swimmers were in the post-pubertal maturational stage (stage 4 or higher; [33]).

Int. J. Environ. Res. Public Health2022,19, 7020 3 of 14 Table 1. Age, height, body mass, and training background characteristics of the control and coordina- tive training groups. Control Group (n = 11) Training Group (n = 8) Age (years) 14.8 0.9 14.8 0.7 Height (cm) 166.6 0.1 170.3 0.1 Arm span (cm) 167.1 0.2 173.0 0.1 Body mass (kg) 58.3 10.1 58.1 10.2 Swimming practice (years) 5.5 1.0 5.5 0.9 2.2. Training Sessions To obtain a larger group of swimmers, the different training sessions were performed with two different teams, but their coaches followed a similar swimming training plan (regarding frequency, volume, and intensity). The CG only performed the normal training sessions in the swimming pool (without any complementary sessions) and the TG per- formed an additional in-water coordinative training session two/week for eight weeks (total 16 sessions). Those coordinative training were supervised by the rst the rst author. 2.3. In-Water Coordinative Training Sessions Considering that speed and SF were the main in uencing factors in front crawl swim- ming technique [12,22], it was used a task manipulation focusing on these two variables. However, as the aim of the current study was also to understand the impact of coordination on performance, the maximal speed was also the target of the in-water coordinative train- ing. Therefore, two sets of 6 25-m maximal speed front crawl [34] were implemented in each training session and SF was manipulated in each repetition as follow: (i) preferred SF; (ii) slightly lower SF; (iii) greatly lower SF comparing to the preferred one; (iv) preferred SF; (v) slightly higher SF; and (vi) greatly higher SF comparing to the preferred one. In each repetition the individual SF and the respective speed value was registered and feedback was given to swimmers. The maximal speed was measured in the pre-test as well as their preferred SF, using a stopwatch. During the training sessions, the time at 25-m was always measured (individually) and each swimmer was asked the number of upper limb cycles performed. The interval between each repetition was established at 1 min and between sets was 3 min.

given to swimmers. The maximal speed was measured in the pre-test as well as their preferred SF, using a stopwatch. During the training sessions, the time at 25-m was always measured (individually) and each swimmer was asked the number of upper limb cycles performed. The interval between each repetition was established at 1 min and between sets was 3 min. 2.4. Test Procedures One month before starting the intervention program, swimmers performed the same pre and post-tests evaluations, trying to understand if the training characteristics were similar between groups (since they belong to two different teams). A standardized 1000 m warm-up at low to moderate swimming intensity was conducted in a 25-m indoor pool before experiments. Afterwards, each swimmer performed a self-paced 50-m front crawl at maximal speed, starting in-water (without diving), with a non-breathing pattern in the centre of the pool to avoid start, turn, and breathe effects on coordination. After each trial, participants were informed of their performance and if their time was not within 2.5% of their 50-m race time, he/she repeated the trial (to ensure that swimmers performed at their best). 2.5. Apparatus While performing the 50-m front crawl test, swimmers used 10 anatomical re ective landmarks in each body side (iliac crest, acromion, lateral humerus epicondyle, and radius- and ulnar-styloid processes), enabling a 3D dual media working volume creation, where the orthogonal axes were de ned as x, y, and z for horizontal, medio-lateral, and vertical (z = 0 de nes the water surface) movements, respectively. A 13 camera setup (MoCap) was used, with seven land plus six underwater cameras (Oqus 3+ and Oqus Underwater, Qualisys AB, Gothenburg, Sweden) operating at 100 Hz. The calibrated volume was de ned using

Int. J. Environ. Res. Public Health2022,19, 7020 4 of 14 underwater, above water, and twin system to merge the rst and the latter calibrations (according to the manufacturer's guidelines). 2.6. Biomechanical Variables Swimming speed was assessed through the ratio of the hip displacement in an upper limbs cycle (distance travelled between two consecutive entries of the same hand) to its total duration. SL was determined by the horizontal distance travelled by the hip during an upper limbs cycle and SF was determined as the number of cycles performed per minute. Stroke index (SI) was computed by the product of speed and SL, and intra-cyclic velocity variation (IVV) was calculated through the ratio between speed standard deviation to mean speed. 2.7. Upper-Limbs Coordination Analysis Coordination between right and left upper limbs was assessed through the continuous relative phase (CRP) [35,36]. CRP assessment between upper limbs (arm–shoulder–trunk angle) was performed for two upper limbs cycles, recorded in the central part of the pool, with cycle duration expressed in percentage allowing its comparison. The CRP was calculated through the subtraction of the phase angle of the two oscillators at each point in time over the entire cycle (i.e., the left shoulder phase angles were subtracted from the right one). CRP values can range from 0–360 , but a variation of 30 was accepted for the determination of a coordination pattern [37–39]. Therefore, three different modes could be found: in-phase (when 330 < CRP < 30 ), anti-phase (when 150 < CRP < 210 ) and out-of-phase (when 30 < CRP < 150 and 210 < CRP < 330 ). From that analysis, different variables were extracted to examine the coordination between upper limbs: (i) the mean CRP and its variability through the Standard Deviation of CRP (SD of CRP) over a cycle; and (ii) the relative time spent in in-phase, out-of-phase, and in anti-phase (all expressed in %) to inform about the coupling between upper limbs coordination. The relative time between two propulsive upper limbs actions was also calculated, namely the index of coordination (IdC; [12]), characterized as the time between the be-

Deviation of CRP (SD of CRP) over a cycle; and (ii) the relative time spent in in-phase, out-of-phase, and in anti-phase (all expressed in %) to inform about the coupling between upper limbs coordination. The relative time between two propulsive upper limbs actions was also calculated, namely the index of coordination (IdC; [12]), characterized as the time between the be- ginning of propulsion of the rst right and the end of propulsion of the rst left upper limbs actions, and between the beginning of propulsion of the second left and the end of propulsion of the rst right upper-limbs actions. IdC was calculated based on the division of the upper limbs actions in four phases: (i) entry and catch, corresponding to the time since the entry of the hand in-water until it starts to make the backward movement; (ii) pull, since the end of the previous action until achieve the vertical alignment of the shoulder ( rst propulsive phase); (iii) push, since the end of the previous action to the exit the hand of the water (second propulsive phase); and (iv) recovery, covering the time from the exit of the hand until its new entry. The IdC and each cycle phase were expressed as the percentage of the duration of a complete upper limbs cycle and the sum of pull and push phases, and of catch and recovery phases, indicate the duration of propulsive and non-propulsive phases, respectively [12]. Three different synchronisation modes are possible to identify in front crawl: (i) opposition (IdC = 0%), when one upper limb begins the propulsive phase and the other is nishing it, providing continuous motor action; (ii) catch-up (IdC < 0%), existing a lag time between propulsive phases of the two upper limbs; and (iii) superposition (IdC > 0%), describing an overlap in the propulsive phases of both upper limbs. 2.8. Statistical Analysis To understand if experimental groups had similar training characteristics, an analysis was conducted for the pre-test and the tests performed one month before. All the statistical analysis were conducted with linear mixed models adjusted for sex (a widely used method

and (iii) superposition (IdC > 0%), describing an overlap in the propulsive phases of both upper limbs. 2.8. Statistical Analysis To understand if experimental groups had similar training characteristics, an analysis was conducted for the pre-test and the tests performed one month before. All the statistical analysis were conducted with linear mixed models adjusted for sex (a widely used method for longitudinal continuous data that considers correlation between repeated measures and the maximum likelihood estimators are easily obtained using standard software [40]). Changes in groups over time (group * time interaction) in coordinative (CRP, standard deviation of CRP, in-phase, anti-phase, out-of-phase, IdC, the four upper limbs phases,

Int. J. Environ. Res. Public Health2022,19, 7020 5 of 14 propulsive, and non-propulsive phases) and performance variables (speed, SF, SL, SI, and IVV) were modelled using a linear mixed model regression with random-effects statements on intercept of each participant. Adjustments for sex were conducted in all variables analysed. The covariance type used for the random effects was the unstructured option (completely general covariance matrix). Normality of residuals was visually veri ed and data were expressed as mean SD. Values ofp< 0.05 were considered signi cant and tests were two-sided, with statistical analysis performed using IBM SPSS software version 24.0 (SPSS, Chicago, IL, USA). 3. Results Comparison between pre-test and the tests performed one month before, did not report signi cant differences in any variable included in the current study, neither between groups, nor time or their interaction, thus the differences registered in the post-test could be related to the implemented trainings. Nevertheless, that analysis provided some information about the sex effect, since in some performance variables (speed, SL, and SI) a signi cant effect was found. 3.1. The Effect of Training on Coordinative Variables The implemented coordinative training showed no in uence on standard deviation of CRP, time percentage spent in in-phase, IdC, time percentage spent in entry and catch, propulsive, and non-propulsive phases. However, as presented in Tables, in the time spent in anti-phase, out-of-phase, and push phase differences were noticed from the pre- to the post-test (models 4, 5, and 8, respectively). Furthermore, a signi cant group*time interaction for CRP (Table) was observed. Table 2. Linear mixed model regression for continuous relative phase (CRP), standard deviation of continuous relative phase (SD of CRP), in-phase, anti-phase, out-of-phase, and index of coordination (IdC), with the unadjusted model and adjusted for sex (models 1–6, respectively). Slope (SE); Statistical Inference CRP Unadjusted model Group 12.78 (9.95); p= 0.21 Time 1.14 (2.99); p= 0.71 Group * Time 12.65 (4.61);p= 0.01 Model 1 Group 13.33 (9.84); p= 0.19 Time 0.54 (3.89); p= 0.89 Group * Time 12.74 (4.60);p= 0.01 Sex 6.94 (9.84);p= 0.49 SD of CRP Unadjusted model Group 19.96 (12.02); p= 0.11 Time

sex (models 1–6, respectively). Slope (SE); Statistical Inference CRP Unadjusted model Group 12.78 (9.95); p= 0.21 Time 1.14 (2.99); p= 0.71 Group * Time 12.65 (4.61);p= 0.01 Model 1 Group 13.33 (9.84); p= 0.19 Time 0.54 (3.89); p= 0.89 Group * Time 12.74 (4.60);p= 0.01 Sex 6.94 (9.84);p= 0.49 SD of CRP Unadjusted model Group 19.96 (12.02); p= 0.11 Time 1.26 (4.58); p= 0.79 Group * Time 11.41 (7.06);p= 0.12 Model 2 Group 20.23 (12.03); p= 0.11 Time 0.47 (5.99); p= 0.94 Group * Time 11.52 (7.08);p= 0.12 Sex 3.44 (12.03);p= 0.78 In-phase Unadjusted model Group 0.88 (2.57); p= 0.28 Time 0.17 (0.96); p= 0.86 Group * Time 0.26 (1.48); p= 0.86 Model 3 Group 0.73 (2.54); p= 0.78 Time 1.14 (1.20); p= 0.36 Group * Time 0.40 (1.42); p= 0.78 Sex 1.90 (2.54); p= 0.46

The study examines the impact of training on swimming performance in youth.