The effects of repeated snatch lifts on lumbar spine mechanics in high-intensity functional trained athletes

Abstract

This study aimed to determine if lumbar spine joint mechanics altered during a power snatch protocol. Eighteen high-intensity functional training (HIFT) athletes performed two countermovement jumps (CMJ) before and after a 30-repetition power snatch protocol at 70% of one repetition maximum. Three-dimensional kinematics and kinetics were collected simultaneously during the power snatch protocol. Average data from five events of the second and third, and twenty-ninth and thirtieth power snatches were compared. At the end of the power snatch protocol, there was a significant decrease in CMJ height (5.4 cm). The power snatch protocol showed a significant decrease in T12-L1 intervertebral joint extension angle (5.3°, p = 0.04; g = 0.65) at the recovery event. Velocity significantly (p = 0.03–0.05), moderately (g = 0.52–0.75) increased for the T12-L1 left rotation (23 °.s⁻¹) and L5-S1 right rotation (23.9 °.s⁻¹) at the end of the second pull event, whereas L5-S1 right lateral flexion velocity (17.8 °.s⁻¹) decreased at the catch event. Finally, the L5-S1 flexor moment (−4.06 Nm) significantly changed to an extensor moment (0.46 Nm) at the transition event (p = 0.018, g = 0.75). These findings suggest that substantial neuromuscular impairment induced by high-intensity repeated power snatches does not necessarily result in a decline of lumbar spine lifting mechanics, rather athletes may elicit a protective ‘stiffening’ of the spinal region.

Introduction

High-intensity functional training (HIFT) is a popular concurrent training paradigm known for executing multiarticular movements, such as the weightlifting snatch, to develop broad physiological and metabolic adaptations.^1,2 The snatch lift is typically performed by ballistically lifting a loaded barbell from the floor to an overhead position in one motion. The lift requires coordination, speed, and explosive strength,^3,4 yielding higher power outputs compared to non-Olympic-style lifts. ⁵ In contrast to traditional weightlifting programs, HIFT programs often prescribe weightlifting derivatives at high volumes of light-to-moderate loads. For example, the benchmark workout ‘Isabel’ is 30 repeated snatches, performed in as short of time as possible, lifting a standardised absolute load of 61 kg aimed to improve anaerobic threshold and strength.^2,6 While evidence highlights the multi-system physiological and psychosocial benefits of the HIFT method,^7,8 performing high volume, complex movements in exhaustive states is speculated to increase neuromuscular impairment, deteriorate movement patterns,^9–11 and increase the risk of injury. ¹⁰

The snatch is identified as one of the core lifts contributing to lumbar spine injury in HIFT athletes.^11–13 The incidence of injuries to the lumbar region in HIFT athletes is reported to be between 19% to 36%,^1,11,14 with L5-S1 disc herniation being more common in competitive HIFT male athletes >19 years old, compared to female athletes (7.5% and 1.7% respectively). ¹³ Lumbar injuries in HIFT athletes are suggested to result from: (i) poor technique under substantial or repeated load, (ii) repetitive spinal hyper-extension and deviating from neutral spine/zone, (iii) weak trunk/core stability, or (iv) excessive loading of the posterior spinal region.^11,13 In-vitro research demonstrates segmental deviations from the neutral zone increase (non-linear) stress load on passive structures, raising the risk of injury if motion exceeds the elastic zone. ¹⁵ However, the biomechanical mechanisms influencing lumbar spine injury regarding the snatch technique and movement efficacy in HIFT athletes are not fully understood.^8,16 During exhaustive repeated lifting tasks such as the snatch, increased neuromuscular impairment may lead to insufficient motor control adjustments and altered joint kinematics, ultimately affecting joint moments. ¹⁷

Several non-Olympic-style weightlifting studies report significant changes in lifting strategies in the sagittal plane at the end of repetitive lifting protocols among healthy male adults and athletes.^18–20 In some cases, the kinematic changes were detrimental to performance due to increased lumbar flexion, a greater reliance on passive structures, and reduced postural stability. These factors are thought to elevate the risk of injury. ²⁰ The lumbosacral (L5-S1) region sustains peak compressive force, and the thoracolumbar (T12-L1) region sustains peak shear forces and bending moments during the catch and recovery of a single snatch. ²¹ Therefore, changes in lifting strategy or altered multiplanar lumbar spine kinematics over repeated snatch lifts may lead to increased joint moments and increased risk of injury to these spinal regions.

Temporary exercise-induced changes or declines in motor control are associated with neuromuscular impairment and reductions in maximal voluntary force, power and velocity.^9,22 Due to the lower-limb biomechanical similarities present during the propulsive phase of the snatch and countermovement jump (CMJ), ²³ average reductions in CMJ height can serve as a sensitive indicator of indirectly measuring reductions in velocity and lower limb neuromuscular impairment^22,24 during weightlifting movements.^25,26 While previous HIFT weightlifting studies have identified an increase in neuromuscular impairment via reductions in countermovement jump height, the associated biomechanical changes were not reported. ²⁷ Furthermore, while kinematics of a single snatch in weightlifting have been extensively reported,^3,28,29 biomechanical research regarding the effects of multiple lifts on multiplanar lumbar spine mechanics as performed in HIFT are currently limited. Therefore, the aims of this study were to determine if 30 power snatches executed for time alters (i) lower limb neuromuscular function, and/or (ii) lumbar spine T12-L1 and L5-S1 joint mechanics from the early to the late stage of the protocol. We hypothesised that repeated power snatch lifts at high intensity would increase neuromuscular impairment and significantly modify lumbar spine mechanics, increasing injury risk at the T12-L1 and L5-S1 intervertebral joints.

Methods

Participants and study design

A priori statistical power analysis indicated 18 participants were required to yield a statistical power of 80% (α=0.05; effect size = 0.25; r = 0.75) (G*Power version 3.1.7, Institute of Experimental Psychology, Heinrich Heine University, Dusseldorf, Germany). Eighteen healthy trained male HIFT athletes between the age of 18 to 30 years (Table 1) were recruited. In this study HIFT was defined as a style of training comprising multimodal functional movement executed at high intensity, developed to improve performance and general physical fitness parameters. ¹ All participants engaged in a HIFT training program, which included monostructural training, bodyweight movements, and weightlifting derivatives such as the snatch, clean, deadlift, and shoulder press. ¹ Pre-testing screening included the Adult Pre-Exercise Screening System [(APSS); Exercise and Sport Science Australia, Queensland, Australia], a physical activity recall log, and an injury history questionnaire. Prior to participation, participants were informed of the study procedures and provided written informed consent. Inclusion criteria required participants to be trained HIFT male athletes with a minimum of twelve months of HIFT experience, currently training three to five days per week, with a current self-reported one-repetition maximum (1RM) power snatch >60 kg (Table 1). This absolute minimum load was established as 61 kg as this is a common load used for repeated Olympic-style lifting in HIFT competition, ³⁰ and ensured participants were sufficiently trained and capable of completing the protocol safely and proficiently. Participants were required to be actively competing in HIFT affiliate competitions or online sanctioned events with no history of an acute musculoskeletal injury within the past six months. Ethical approval was obtained by the Charles Sturt University Human Research Ethics Committee (H20210).

Table 1.

Participant characteristics.

Subjects (n = 18)	Mean ± SD
Participant age (years)	25.16 ± 4.41
HIFT training experience (years)	3.75 ± 2.20
HIFT weekly training minutes (min)	376.72 ± 108.86
Height (cm)	180.74 ± 8.82
Mass (kg)	83.02 ± 8.13
Reported 1RM (kg)	69.44 ± 9.34
Power Snatch load 70% 1 RM (kg)	47.97 ± 7.19

Procedures

Participants had their body mass, height and body segment depths (chest, xiphoid and pelvis depth) measured prior to fifty-two passive-reflective markers adhered to specific bony landmarks on the skin and shoe surface, defining the foot, shank, thigh, pelvis, and trunk segments.^31–33 Participants wore minimal clothing (shorts and lifting shoes) to avoid obstructing or occluding the passive-reflective markers.

After completing a standardised warm-up comprising three minutes of self-paced low-intensity cycling (WattBike, Nottingham, England) and a ten-minute mobility routine, ³⁴ participants were familiarised with the testing protocol, including the CMJ and the power snatch lift. Familiarisation involved three CMJs and five sets of two power snatches starting with the 20 kg Trojan empty barbell, (TROB1500, Joondulap, Australia) and self-selecting weight increases up to 70% of the participant's self-reported 1RM power snatch weight. The testing protocol was then outlined to the subject.

In preparation for the CMJs, participants were instructed to stand erect and motionless with their hands crossed on opposite shoulders to avoid occluding the right greater trochanter marker. The CMJ consisted of flexing the hip and knees and dorsiflexing the ankles to reach an approximate 90° knee flexion squat, then rapidly jump extending the hips, knees and ankles to achieve maximal vertical height before landing. For each CMJ participants were instructed to jump as high as they could. Participants performed two pre-testing CMJs, followed by a three-minute rest, 30 repeated power snatches, immediately followed by another two post-testing CMJs.

For the repeated snatch protocol, 30 power snatches were completed for time, at 70% of each subject's self-reported 1RM. Participants were familiar with the protocol and instructed to perform the test at a self-paced, high-intensity level while prioritising safety. To replicate the HIFT environment, the researcher provided verbal encouragement, and a clock was visible to the participants throughout the protocol. Participants utilised a hook grip on the bar, double shoulder width apart, in the power snatch start position (Figure 1). Each repetition was deemed complete if the loaded barbell was lifted and vertically displaced overhead in one continuous motion overhead with hip, knee, and elbow reaching full extension at the end of the lift. ³⁵ Participants were permitted to drop the barbell onto the lifting platform from hip height after completing each repetition (Figure 1). The first repetition commenced when the digital timer began continuously counting up from 0. Participants performed repeated power snatches ballistically until completing 30 repetitions. The clock was stopped as soon the bar hit the floor after repetition number 30.

Figure 1.

Power snatch protocol set-up.

The testing protocol (CMJ and power snatches) was performed with one foot on each of the two Kistler multicomponent piezoelectric force platforms (2500 Hz; Kistler, 9281CA, Winterthur, Switzerland) with built-in amplifiers, each connected to control units, (Kistler, 5233A, Winterthur, Switzerland). Participants were instructed to keep their feet as close to the centre of each force platform as possible (Figure 1). Trunk and lower limb kinematic data were also recorded using eight Qualisys Oqus 700+ time-synchronised three-dimensional motion capture cameras (250 Hz; Qualisys, Oqus 700+, Gothenburg, Sweden) and two Qualisys Miqus M3 high-speed video cameras (50 Hz; Qualisys, Miqus M3, Gothenburg, Sweden) positioned on a ceiling-mounted frame. Prior to data collection, a static trial was recorded to define the participant's segments for biomechanical modelling purposes.

Data analysis

The CMJ height was calculated based on the vertical displacement of the right greater trochanter marker. Throughout the repeated power snatch protocol, thoracolumbar (T12-L1) and lumbosacral vertebral joint space (L5-S1) angles, velocities, and moments in all three anatomical planes, were calculated. Each body segment was defined in the Cardan x-y-z sequence, where the x-axis defined T12-L1 and L5-S1 flexion-extension, the y-axis T12-L1 and L5-S1 lateral flexion, and the z-axis T12-L1 and L5-S1 rotation. The second and third power snatches and the twenty-ninth and thirtieth power snatches were analysed to compare each subject's early and late protocol kinematics and kinetics. The first power snatch was excluded from the data analysis due to possible motor learning and stability effects.^36,37

All raw kinematic coordinates, GRFs, free moments and centre of pressure data were filtered through fourth-order zero-phase Butterworth low-pass digital filters before calculating individual joint kinematics and net internal joint moments (f_c = 18 Hz). All data were collected continuously using Qualisys Track Manager software Version 2020.3 (Qualisys, Gothenburg, Sweden), then imported and analysed using Visual 3D software Version 4 (C-Motion, Germantown, MD). Biomechanical models were constructed with the thigh, shank and foot defined as a frusta of right cones, and the trunk, lumbar, thorax and pelvis defined as elliptical cylinders. ³⁸ Segment masses and inertial properties for the trunk, lumbar, thorax and pelvis were defined by Pearsall, Reid, ³⁹ and the thigh, shank and foot segment masses were defined by methods established by Zatsiorsky, Seluyanov. ⁴⁰

The five events of the power snatch were defined by the knee joint angle in the sagittal plane. ³ The five events (Figure 2) were: (i) start of the first pull (first maximum knee flexion); (ii) start of the transition (first maximum knee extension); (iii) end of the second pull (second maximum knee extension); (iv) end of the catch (third maximum knee flexion), and (v) end of the recovery (third maximum knee extension). Kinetic and kinematic data collected at the five events from the second and third power snatch and the twenty-ninth and thirtieth power snatch were averaged for each subject, respectively.

Figure 2.

Power snatch events left knee angle (x-axis) identifying the events of the power snatch (i) start of first pull: first maximum knee flexion, (ii) start transition: first maximum knee extension, (iii) end of the second pull: second maximum knee extension, (iv) catch: third maximum knee flexion, (v) recovery: third maximum knee extension. Note: This is a single representative example.

Statistical analyses

Prior to statistical analysis, outliers were assessed, with measures greater than two standard deviations away from the mean removed and excluded from the analysis, ⁴¹ with no imputations made. Statistical analysis was conducted using the Statistical Procedures for the Social Sciences (version 24, SPSS Inc., Chicago, IL). Data analyses were conducted using each respective snatch event (i.e., start of the first pull; start of the transition end of the second pull; end of the catch; and end of the recovery) using linear mixed-effects models to analyse the difference between the main (variable, early protocol) and interaction (variable, late protocol) effects. Participants were set as a random intercept, and the model was fitted via maximum likelihood estimation using a 1 × 1 covariance matrix structure. The model analysed the difference between CMJ height and T12-L1 and L5-S1 joint mechanics (i.e., angle, velocity and moment) between the early of the protocol (group 1) and end of the protocol (group 2) of each respective dependent variable at each respective power snatch event. The fixed effects on the dependent variable resulted in one combination, with eighteen participants in each combination. Statistical significance was set at α < 0.05. Following the linear mixed model, main and interaction effects were evaluated using the Fisher's Least Significant Difference (LSD) method pairwise comparison. Where significant differences were observed between interventions, practical differences in means were interpreted using Hedges g effect size (ES) statistic, where g < 0.2 = “trivial”, 0.2 ≤ g < 0.5 = “small”, 0.50 ≤ g < 0.8 = “medium” and g ≥ 0.8 = “large”. ⁴²

Results

Countermovement jump

Countermovement jump height significantly decreased post-protocol (42.30 cm ± 8.13 cm to 36.95 cm ± 6.46 cm; p = 0.036, g = 0.66).

Repeated power snatch protocol

Participants’ average time to complete the thirty repeated power snatch protocol was 234.72 ± 70.83 s. Significant joint angle differences between the early stage of the power snatch protocol and the late stage of the power snatch protocol were identified only for the T12-L1 extension angle, which significantly reduced at the recovery event (27.60° ± 7.37° to 22.34° ± 7.41°, p = 0.043, g = 0.65). There were no significant differences identified in other joint angles (online Supplementary Resource).

Lumbar joint velocities at the late stage of the protocol displayed significant differences when compared to the early stage of the protocol. The T12-L1 right rotation velocity at the end of the second pull event increased in the left direction at the late stage of the power snatch protocol (7.26°.s⁻¹ ± 24.21°.s⁻¹ to −14.72°.s⁻¹ ± 30.50°.s⁻¹, p = 0.042, g = 0.65; online Supplementary Resource). At the same event, the L5-S1 left rotation velocity increased in the right direction at the late stage of the protocol (−5.08°.s⁻¹ ± 29.22°.s⁻¹ to 18.77°.s⁻¹ ± 33.26°.s⁻¹, p = 0.031, g = 0.69; online Supplementary Resource). The L5-S1 left lateral flexion velocity at the catch event, decreased in the right direction at the late stage of the protocol (9.35°.s⁻¹ ± 29.27°.s⁻¹ to −8.46°.s⁻¹ ± 20.92°.s⁻¹, p = 0.05, g = 0.52). There were no significant differences in other joint velocities (online Supplementary Resource).

Significant differences were identified within joint moments with the L5-S1 flexion moment changing to an L5-S1 extensor moment increasing at the transition event at the late stage of the protocol compared to the early stage of the protocol (0.46 Nm ± 5.63 Nm to −4.06 Nm ± 5.12 Nm, p = 0.018, g = 0.75). There were no other significant differences in other joint moments (online Supplementary Resource).

Discussion

The current study is unique in examining changes in lower limb neuromuscular status and lumbar spine mechanics during repeated power snatch lifts in trained male competitive HIFT athletes. The 30 repeated power snatch lifts executed at high intensity resulted in a 12.65% reduction in average CMJ height post-protocol, indicating temporary lower limb neuromuscular impairment. This finding aligns with previous research reporting a 7.35% reduction in CMJ height post repeated power clean protocol. The authors attributed the decline in performance to intramuscular changes from the high-intensity repetitive exercise protocol. ²⁷ This evidence supports our first hypothesis that repeated power snatch lifts executed at high intensity would induce neuromuscular impairment.

Contrary to the second hypothesis, significant changes in lumbar spine mechanics at the late stage of the repeated power snatch protocol did not result in deleterious changes that would increase the risk of injury. Our data showed in some cases, the mechanical changes were associated with maintaining spinal stability in neuromuscular impaired states. For instance, at the late stage of the protocol during the recovery event, a moderate reduction in T12-L1 joint extension angle resulted in a more upright trunk position, reduced hyperextension and improvement towards the neutral zone (1.89° ± 0.47° flexion/extension based on in-vitro measures, as such, there is likely some individual variation ⁴³ ). Although, T12-L1 intervertebral joint angles at the recovery event exceeded the neutral zone for the T12-L1 segment, the moderate improvement (i.e., closer towards the neutral zone) is thought to be beneficial as it likely attenuated some of the stress concentrations on passive structures,^15,43 potentially lowering the risk of injury compared to the early stage of the protocol. This change is important as the T12-L1 region is subjected to compressive and peak shear forces during the recovery phase, as the barbell system is directed overhead. ²¹ A possible explanation for the observed reduction in T12-L1 extension is the increased co-contraction of the core stabiliser muscles at the recovery event. Previous research indicates that during repeated sagittal lifting tasks, increased flexion angle and enhanced antagonist co-contraction are employed to protect against spinal instability. ⁴⁴ This may be further supported by the small, albeit non-significant, increase in T12-L1 flexor velocity (p = 0.286, g = 0.36; see online Supplementary Resource) in the late stage of the protocol. This motor control strategy is particularly beneficial at low trunk moments in upright trunk postures, contributing to improve spinal joint stiffness, as well as potentially lowering the risk of stability failure and injuries associated with load tolerance. ⁴⁴ Therefore, the reduction in T12-L1 extension angle at the recovery event likely attenuated the stress on passive structures and improved spinal stability in impaired states, potentially reducing the risk of load related injuries.

Protective motor control adaptations were evident as neuromuscular impairment progressed during the repeated snatch protocol, as both T12-L1 and L5-S1 axial rotation velocities moderately increased in the opposite direction at the second pull event. Although lateral flexion velocities did not change substantially (T12-L1 p = 0.529, g = 0.21 and L5-S1 p = 0.654, g = 0.15; see online Supplementary Resource), there is still potential to be meaningful, contributing to the overall protective mechanism. Increased spinal velocity in the transverse and frontal planes during repeated sagittal plane loaded trunk movement is associated with decreased coordination and stability in the sagittal plane, ⁴⁵ research indicates that provided neuromuscular impairment is not severe, increased rotation and lateral flexion velocities may temporarily restore spinal stability. ⁴⁶ Authors suggest stability is maintained by co-activation or mechanical coupling simplifying complex activation patterns into a lower-dimensional state.^46,47 As electromyographic studies identify the erector spinae as a primary effector organ during weightlifting pull derivatives,^48,49 one possible explanation for the observed velocity change may be attributed to increased oblique abdominal muscle activation, recruited to provide multiplanar control due to their orientation, when contractile characteristics of the trunk extensors are altered during repeated high-intensity snatch lifts. ⁴⁶ While adaptations from proprioceptive open-loop feedback delay further instability impairment to improve lumbar joint stiffness and stability, ¹⁷ research reports increased spinal loading on the surrounding passive structures. ⁵⁰ However, our results observed no significant changes in joint moments at the corresponding events. Nevertheless, it is important to note that if the protocol were prolonged or impairment became severe, the likelihood of injury would likely increase. Previous research suggests that if repeated exhaustive lifting continues and neuromuscular impairment worsens, the capacity to compensate for load perturbations diminishes. ⁴⁵

Conversely, at the catch event, L5-S1 lateral flexion velocity changed as velocity decreased and changed direction at the late stage of the protocol. Reductions in spinal velocity during repeated iso-inertial trunk extensions are associated with impaired motor control (time delays in the neuromuscular system), reduced muscle contractility, altered ability to compensate for load perturbations and greater instability. ⁴⁵ However, as the reduction in L5-S1 lateral flexion velocity magnitude at the catch event in our study was 0.89°^.s⁻¹, and the subsequent recovery event showed values return to early protocol, we suggest that athletes were able to maintain stability despite impairment. Noteably, all spinal velocity values reported in this study remained lower than the mean maximum trunk rotation velocity of 38.04°^.s⁻¹, or trunk lateral flexion velocity of 35.45°^.s⁻¹ considered the lower threshold probability risk of spinal injury during repetitive occupational tasks. ⁵¹ Therefore, we speculate that as neuromuscular impairment increased during the repeated snatch protocol, changes in muscle recruitment patterns to a lower-dimensional motor control states may explain the changes in T12-L1 and L5-S1 velocity, aiding to maintain stability at the T12-L1 and L5-S1 levels. This may temporarily reduce the risk of stability failure in well-trained HIFT athletes at the late stage of the protocol at the second pull event and catch events. However, we recommend further investigation using electromyography to analyse changes in core activation patterns during repeated weightlifting tasks at high-intensity.

During the transition event, despite lumbar spine lifting kinematics remaining mostly unchanged, the L5-S1 joint moment changed from a very small flexion moment in the early stage of the protocol to a slightly greater extensor moment late in the protocol, thus supporting our hypothesis. While the respective joint angle remained unchanged, L5-S1 extensor velocity showed a non-significant decrease (see online Supplementary Resource). Decreased L5-S1 velocity is associated with reduced force-producing capabilities and contractile speed of the spinal extensors^17,45 and increased inertial effects on the load/barbell, which may attribute to the increased L5-S1 moment.^52,53 As the spinal extensors assist in resisting the barbell's forward torque through the transition phase, ⁵⁴ altered lower limb neuromuscular function could decrease in the impulse force at the start of the pull (increasing the inertial load of the barbell as the load is constant and closer to its linear resting inertia), thereby increasing the total extensor moment.^29,53 Coupled with impaired activation this finding could indicate an increased load-sharing of the passive structures. ⁵⁵ However, as we did not measure the force-producing capacity of the extensor muscle, we cannot confirm extensor impairment or the extent of passive tissue loading through the transition phase of the power snatch. Despite increased loading, the L5-S1 extensor moment was below the maximum mean moment of 23.64 Nm, which is considered a low probability risk of spinal injury. ⁵¹ Therefore, although we observed an increase in the L5-S1 extensor moment, it was unlikely to increase the probability of injury.

In summary, results showed that the repeated power snatch protocol resulted in moderate neuromuscular impairment in well-trained HIFT athletes, evident by the significant reductions in CMJ height. Previous studies have shown as neuromuscular impairment increases, the capacity to compensate for load perturbations diminishes. However, contrary to our hypothesis, neuromuscular impairment did not result in deleterious changes in T12-L1 and L5-S1 joint kinematics and kinetics. Results indicate changes in spine mechanics demonstrated improvements towards the neutral zone, showing reduced hyperextension at the late stage, reducing stress on passive structures, consequently lowering the risk of injury. Implementation of a protective motor control strategy, through small increases in spinal velocities, as seen in this study was not suggested to result in increased spinal loading as joint moments at corresponding time points did not significantly increase. Although there was an increase in the moment during the late stage of the protocol, which could elevate the load on the passive structures around L5-S1, these moments remained lower than those previously reported as indicators of a low threshold for injury risk. Overall, despite the increased neuromuscular impairment, lumbar spine lifting mechanics of well-trained HIFT male athletes were generally unchanged or slightly improved during repeated power snatch lifts for time.

Supplemental Material

sj-docx-1-spo-10.1177_17479541251376811 - Supplemental material for The effects of repeated snatch lifts on lumbar spine mechanics in high-intensity functional trained athletes

Supplemental material, sj-docx-1-spo-10.1177_17479541251376811 for The effects of repeated snatch lifts on lumbar spine mechanics in high-intensity functional trained athletes by Ashlee E Pilgrim, Kerry J Mann, Eric J Drinkwater, Rylee A Dionigi and Gavin Buzza in International Journal of Sports Science & Coaching