Participants

Twenty-three male experienced football players (age: 22.5±2.6 years, height: 181.3±6.3 cm, body mass: 79.9±8.6 kg) recruited from the second to sixth division of the Norwegian national league and participated in football training for a minimum of two times a week, participated in the study. All the participants were right-foot dominant, defined as preferred foot for kicking the ball. Each subject was informed to avoid heavy training and consuming alcohol with a minimum of 24 hours prior to testing and further procedures and risks of participation were explained to each participant. Before the tests, a signed written informed consent was provided by the participants, which was approved by the Norwegian Centre for Research Data project nr: 42440 and performed according to the declaration of Helsinki and of the journal [38].

Experimental design

A within-subject design was employed, where all participants performed a series of different COD tests, strength tests and plyometric tests in one test occasion. This was conducted to investigate the association of performance measures in the strength and plyometric tests with performance in force- and velocity-dominant CODs. Muscle activation data were collected from the tests to investigate the similarities in activity levels between the different exercises.

Procedures

Prior to testing, two days of familiarization was required, practising the strength, plyometric and COD tests, to avoid results at the day of testing being affected from a learning effect. A minimum of four days was set between each day of familiarization, to promote recovery and to establish one repetition maximum (1RM) in the different strength exercises. Participants were encouraged to provide maximal effort in all exercises on the familiarization days to avoid unnecessary fatiguing attempts to establish 1RM and to control whether it was their maximal effort at the day of testing. The protocol was identical for the days of familiarization and testing, starting with the COD test before the strength and plyometric tests were performed in a randomized order. First, on the day of testing, height and body mass were measured. Afterwards, electromyography (EMG) was placed at ten muscles of the dominant lower limb before dressing up in a full body motion capture suit. After all the equipment was attached, the warm-up protocol for the COD test was initiated.

The warm-up consisted of five minutes jogging at a self-selected pace, before performing three runs with increasing intensity up to 60%, 70% and 80% of self-perceived maximum intensity. Afterwards, participants performed a specific warm-up of running 15 metres, then turning either right or left with a 65° or 110° at 80% of self-perceived maximum intensity. In total, four runs (right or left: 2 x 2: 65° or 110°) were performed in a randomized order with 60 s rest between each run. After the warm-up, the COD test was performed at maximal intensity with a three-to-five-minute rest between each attempt.

The COD test consisted of a 45° and 180° turn with a 4 m and 20 m sprint approaching the COD, making four runs in total (Fig 1). The COD track was designed to investigate the extremities of force- vs velocity-dominant turns, while accounting for sprint distance approaching the turn. All CODs consisted of left turns, making the dominant foot (right) perform the COD step. The COD test was performed on an indoor court surface (Taraflex Sport Evolution M 7.0 mm, Unisport, Finland). Participants began from a standing-still start, with the front foot placed 20 cm behind the starting line placed either 4 m or 20 m from the COD area (Fig 1). The participant started running on their own accord with maximal effort. The time started when crossing the first timing gates placed along a two-metre-long starting line running towards the COD area, where the subject performed a left turn (45° or 180°) and re-accelerated 4 m to finish off the test. For an attempt to be approved for the 45° COD, the subject had to turn from right to left around a cone placed in the midpoint of the COD area. The cone was removed when performing the 180° COD, then both feet were required to be inside of the COD area (Fig 1) for an attempt to be approved. In case of slipping or violations of the requirements due to any other unpredicted events, a reattempt was required. Also, to ensure maximum effort, a test attempt resulting in a performance decrease of 0.1 s or more from the second familiarization day resulted in a reattempt.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. Change of direction test set up with approach of 4 or 20 m with timing gates on 4 and 20 m with COD of 45 and 180° followed by 4 m sprint.

https://doi.org/10.1371/journal.pone.0238580.g001

After the COD test, the strength and plyometric exercises were performed in a randomized order (Figs 2 and 3). As the participant was already warm, the only warm-up prior to the strength and plyometric exercises consisted of performing the exercise at submaximal intensities, leading up to a maximal effort attempt. For each of the strength and plyometric tests, the participant had three attempts. However, one attempt for establishing 1RM in the strength exercises was mostly enough, due to the estimates from the second day of familiarization. Proper performance technique (mainly depth) in the different exercises was controlled for by using 3D modelling. An extra attempt was necessary if requirements of the tests were violated. The test-attempt with highest performance, within the given test requirements, was used for further analysis.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 2. Illustration of the strength exercises.

(A) Unilateral quarter squat. (B) Bilateral parallel squat. (C) Lateral squat.

https://doi.org/10.1371/journal.pone.0238580.g002

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 3. Illustration of the different plyometric exercises.

(A) Drop jump. (B) Countermovement jump. (C) Skate jump.

https://doi.org/10.1371/journal.pone.0238580.g003

The chosen exercises conducted in the vertical direction performed bilaterally and unilaterally are often used in training interventions [10, 34] and the strength and plyometric tests were matched in movement direction. Additionally, the skate jump and lateral squat were chosen since they could simulate the lateral direction of the COD step [39]. The parallel squat was performed bilaterally. The depth required was that a visualized line between the patella and trochanter major should be parallel to the ground. The unilateral squat was performed with the dominant foot in a Smith machine and bending the knee until reaching 40 degrees. The lateral squat was performed by moving the dominant foot to the side, with an extension of the hip initiating the movement. The dominant foot’s heel should not be placed any further forward than the toes of the non-dominant foot. The depth of the dominant foot needed to be 90° while the non-dominant foot was required to be stretched out.

In the unilateral vertical countermovement jump, the performance variable was jump height. The test was performed with the dominant foot, with hands placed akimbo and a straight back, limiting usage of the back and arms to contribute to jumping performance. The non-dominant foot was required to remain passive throughout the jump. The drop jump was performed from an individually customized drop height ranging from 0.15 to 0.6 metres, found at the days of familiarization. Optimal drop height was the height from which the subject obtained the highest reactive strength index (RSI) value, which was the performance variable of the test (flight-time/contact time). The RSI value was obtained by a contact grid. When performing the drop jump, hands were akimbo. Skate jump was performed laterally, jumping with the dominant foot for maximal distance. For an attempt to be approved, the subject needed to stand still in the landing. Distance was measured manually with measuring tape to the closest 0.01 m.

Measurements

Total time to complete the COD test was measured with electronic timing gates (Brower Timing Systems, Salt Lake City, Utah, USA). RSI and jump height were found using infrared optical contact grid (Ergotest innovation, Porsgrunn, Norway) with a resolution of < 0.02 s. The contact grid consists of two units, irSOURCE and irMirror, which reflects an infrared carpet a few millimetres above the floor. Contact time was recorded when disrupting the infrared carpet. Jump height was calculated by flight time, using the equation:

Muscle activity was measured using EMG (Ergostest Innovation, Porsgrunn Norway) with a sampling rate of 1 kHz. The EMG sensors were attached to electrodes (Zynex Neurodiagnostics, CO, USA) on ten different muscles of the subjects’ dominant foot. The skin had to be shaved and washed with alcohol before placing the electrodes (11 mm contact diameter and 2 cm centre-to-centre distance), along the direction of the presumed muscle fibres on the lateral and medial vastii, rectus femoris, adductor longus, biceps femoris, semitendinosus, soleus, gastrocnemius, gluteus medius and maximus muscles according to the recommendations of Hermens, Freriks [40].

The EMG raw signal was amplified and filtered using a preamplifier located as close as possible to the pickup point to minimize noise induced from external sources through the signal cables. The preamplifier had a common-mode rejection ratio of 106 dB and the input impedance between each electrode pair was > 10¹² Ω. The EMG raw signal was then bandpass-filtered (fourth-order Butterworth filter) with cut-off frequencies of 20 Hz and 500 Hz. The resulting EMG signals were converted to root mean square (RMS) signals using a hardware circuit network (frequency response 450 kHz, averaging constant 12 ms, total error ± 0.5%). Peak EMG signal in the COD step and acceleration step where the dominant foot was in contact with the ground was used for further analysis, together with the peak EMG in the strength and plyometric exercises. EMG was collected in Musclelab 10.5.69 (Ergotest innovation A.S, Porsgrunn, Norway) and synchronized with a contact grid and a 3D motion-capture system: Xsens motion capture, MVN link (Xsens Technologies B.V. Enschede, Netherlands). Xsens is a full-body motion-capture system, based on 17 inertial sensors on different anatomical points with a sample rate of 240 Hz.

Statistical analyses

Correlational analysis was performed by Pearson’s r. Level of significance was set at p < 0.05. Strength performance for correlation was expressed by Wilks points, a valid way of normalizing strength performances by accounting for body mass [41]. Jump height/length and the reactive strength index were performance measures used for the plyometric exercises. A one-way ANOVA with repeated measures was used to compare muscle activation in the different strength and plyometric tests with the COD step, and first subsequent step of the dominant foot (acceleration step). When significant differences occurred, the Holm−Bonferroni post hoc test was conducted. Violations of the assumption of sphericity was adjusted for by the Greenhouse−Geisser correction.

Correlations

The high correlations between plyometric performance and COD were expected and in accordance with previous research [42–45]. However, the low correlation between strength with COD performance was somewhat surprising, as strength qualities are factors influencing COD performance [15, 43], although strength is also context-specific. As the direction of forces produced in exercises should mimic the CODs [34], movements in the plyometric and strength tests were matched on bilateral vertical, unilateral vertical and lateral unilateral exercise to induce similar direction of forces produced and the muscles range of motion.

As specified per exercise, the skate jump correlated best with all CODs, which is explainable as it is the only exercise performed laterally and horizontal power productions are important for accelerating and decelerating in CODs [33, 39]. As COD movements occur unilaterally in the vertical/horizontal direction they require mostly medio-lateral force production [10]. Therefore, the unilateral countermovement jump also correlated with most CODs, but less due to the more vertical application of the jump. That is possibly also the reason why it was not related to the 20 m 180° COD, because this COD requires a greater amount of breaking and propulsive forces [36, 46] produced in both the anterior-posterior and medio-lateral direction [10]. The drop jump performance only correlated significantly with the 20 m 45° COD performance, which might be due to this COD performance being largely influenced by the straight-line sprint prior to the COD step [46–48]. This is because the reactive strength index is dictated by contact time, which also is also associated with sprint performance [49]. Time to exert a high amount of force in such fast dynamic movements is limited. Therefore, the relative amount of forces produced in the sprint prior to the COD step in the 20 m 45° COD is more vertically directed due to an increased velocity [50–52].

None of the strength tests induced a significant correlation with COD performance and this is perhaps related to the specificity of contraction velocity, which is suggested to be specific for both strength and sprinting [53, 54]. Probably the contractions are too slow to have a direct transfer to the different CODs. Furthermore, the pre-stretch of the muscles that may enhance the concentric contraction due to neural potentiation allows the recruitment of a greater number of motor units [55], with larger effect at increasing velocities [56, 57]. The faster pre-stretch in COD and plyometric exercises, thus attaining a higher level of cross-bridges may explain the higher observed muscle activation. However, an explanation of the muscle-specific tasks in the different conditions must not be neglected.

Muscle activation

Firstly, both hamstring muscles revealed lower muscle activation in the strength exercises compared to the CODs (Fig 4). Knee flexor strength is suggested to help maintain trunk stability when decelerating in a COD [42] and producing horizontal force [58], which are important aspects for promoting overall COD performance [10, 59, 60]. As such, despite the relatively large emphasis on gluteus maximus and knee flexors in the different squat variations, the hamstrings are seemingly neglected. In comparison, the drop jump might induce a similar eccentric requirement of the hamstring muscles compared to the COD, as the hamstrings decelerate knee joint moments in the COD [61–63] prior to the concentric propulsive force [10]. The skate jump also revealed similar muscle activation of the hamstrings, which might be due to a faster movement increasing the eccentric load, utilizing a faster stretch-shortening cycle, in comparison to maximum lifts where the velocity is low.

Secondly, the strength exercises were found to induce lower muscle activation of the adductor longus and gluteus medius. Muscle activation was only similar for the gluteus medius when performing the unilateral squat, compared to the CODs. This might be due to the unilateral squat being performed on one foot, without a foot stabilizing, requiring more stabilization of the hip. The differences in the adductor longus muscles were only found for the force-dominant CODs. In the force-dominant CODs, the participants perform a total change in momentum [64] and start turning prior to the COD step [65] where the adductor functions as stabilizers of the hip [66, 67].

Thirdly, muscle activation in the gastrocnemius was found to be lower in the strength exercises when compared to the CODs, except for the lateral squat. This is because the lateral squat is the only strength exercise requiring plantar flexion of the ankle. On the other hand, all of the plyometric exercises require plantar flexion of the ankle, which explains the similar muscle activation to that in the CODs, as the ankle plantar flexes for the decelerating moment in the COD step [64, 67] by absorbing impact forces [59].

Based on the observed positive associations between the plyometric exercises and especially the skate jump with the CODs and observed peak muscle activation it seems that the plyometric exercises should be recommended to develop both force and velocity dominant CODs, while the measured strength exercises should not be recommended based upon the present findings. However, the present study did not investigate if implementing these exercises over a period of time helps the COD performance. Earlier training studies have shown that implementing some of these plyometric exercises in training increased COD performance [34]. The skate jump had the highest associations with the different COD, while only similar exercises performed laterally are supplemented in earlier training interventions [16, 68, 69], it would be interesting to investigate the isolated effect of this exercise in future training studies and how it effects COD performance. Furthermore, the plyometric exercises displayed a greater similarity in muscle activation in the hamstring muscles, adductor longus, gastrocnemius, gluteus medius and a greater association with COD performance compared to the strength exercises. It is possible that strength exercises did not show the same level of association due to the lack of activation in these muscles. As such, it is possible that selecting different strength exercises or modifying the current strength exercises with the goal of stimulating muscle activation more like the COD can better improve COD performance.

Limitations

Although the plyometric exercises are seemingly more specific towards both force- and velocity-dominant CODs, some limitations must be addressed. Firstly, the strength exercises were measured at maximal external load lifted, possibly inhibiting the specificity towards the CODs. A lighter load with the intention of maximizing power output might reveal more correlation in performance measures [19]. Secondly, the number of tested players and wide variety of player level (second to sixth national level) could be a limitation to generalise the findings. But the variation in test performances between the players was low and thereby these performances are comparable for players of different levels. However, more studies with players from same playing level should be performed to investigate if the associations between CODs with strength and plyometric tests are comparable with the ones found in the present study. Thirdly, EMG measurements are always at risk of crosstalk between muscles, especially the fast-dynamic movements like the plyometrics and CODs. Future research should be conducted with a force plate in the COD step and for the different strength and plyometric tests, to investigate the forces produced. These force measurements could provide greater insight into the specificity of the strength and plyometric exercises towards CODs, especially the exercises performed laterally.