Abstract
pplied muscular strain and hamstring strain capacity have a joint interaction on hamstring strain injury (HSI) with modifiable risk factors frequently assessed. However, to date there is limited observations on the interaction between these factors. The purpose of the present study was to observe if spatiotemporal characteristics, running kinematics and muscle activation were related to modifiable risk factors of HSI. Twenty-two competitive team sport athletes (24.7±4.3 years, 1.82±0.07 m, 84.9±8.5 kg) participated whereby the Bicep femoris long head (BFLH) fascicle length assessed via ultrasound and isokinetic eccentric hamstring strength was assessed. With running assessment performed at 18 km/h, capturing running kinematics and muscle activation. Multiple linear regressions were used to examine the relationship of running kinematics and muscle activation on the modifiable risk factors of HSI on. The overall model (F2,19) was statistically significant for both relative eccentric hamstring strength (F = 23.58,p< 0.001) and BFLHfascicle length (F = 18.87, p< 0.001) highlighting spatiotemporal characteristics, running kinematics and hamstring activation were found to be significantly related to the modifiable risk factors. There is a complex interrelationship between running mechanics and hamstring muscle properties, with the potential of either cause or consequence association. Keywords:eccentric strength; fascicle length;
relative eccentric hamstring strength (F = 23.58,p< 0.001) and BFLHfascicle length (F = 18.87, p< 0.001) highlighting spatiotemporal characteristics, running kinematics and hamstring activation were found to be significantly related to the modifiable risk factors. There is a complex interrelationship between running mechanics and hamstring muscle properties, with the potential of either cause or consequence association. Keywords:eccentric strength; fascicle length; bicep femoris; spatiotemporal; hamstring strain injuries 1. Introduction The incidence of hamstring strain injuries (HSIs) within sport frequently result from performing one of two high risk actions, kicking or high-speed locomotion [1], with inciting events occurring in phases of acceleration and deceleration [2]. The elevated risk of a HSI occurrence during these high velocity actions is due to the requirement for the hamstrings to produce high forces (i.e., up to 10.5 N/kg for the bicep femoris long head [BFLH] during the terminal swing phase) [3], to resist rapid knee extension [1,4–6]. An eccentric muscle action is proposed to occur during the terminal swing phase of sprinting based on three-dimensional(3D) modelling [7–13], although this is contested within the literature with Van Hooren and Bosch [14] postulating that an isometric muscle action occurs during the terminal swing phase based on animal models. Recent research has added further ambiguity to this area; Yoon et al. [15] observed that the BFLHfascicle actively shortens during the late swing phase; however, they also observed asynchronous behaviour of the Muscles2025,4, 44 https://doi.org/10.3390/muscles4040044
Muscles2025,4, 44 2 of 16 muscle-tendon unit (MTU). The asynchronous action observed with the MTU could suggest that length changes may occur primarily in the tendon rather than the muscle at increasing sprinting velocities [15]. This could support the notion from Van Hooren and Bosch [14] whereby high forces generated during an eccentric muscle action are the resultant cause of a HSI event [14,16,17]. However, it is important to acknowledge that these findings may partially reflect methodological limitations associated with ultrasound imaging and modelling of dynamic movements at high velocities. Despite the ambiguity on what muscle action is occurring during the terminal swing phase, eccentric hamstring strength is strongest modifiable predictor of HSI risk and has been shown to reduce the risk of HSIs [18–20]. The results of hamstring strength training interventions have indicated that the hamstring muscles adapt rapidly to the stimulus applied, specifically with the inclusion of an eccentric training stimulus, there is a rapid increase in both eccentric hamstring strength and BFLHfascicle length [21]. Both of these adaptive responses to eccentric hamstring focused strength training have not only demonstrated reductions in HSI occurrence [20,22,23], but also subsequent increases in performance of athletic tasks such as sprinting and jumping [24–27]. Recently, absolute measures of eccentric hamstring strength, assessed using an isokinetic dynamometer, and BFLHfascicle length were shown to have no meaningful association, contrastingly when measures were taken in relative to body mass (relative eccentric hamstring strength) and a measure of relative BFLHfascicle length (i.e., relative to the anatomical distance between attachments sites [ischial tuberosity and lateral tibial epicondyle]), this markedly increased with moderate to nearly perfect relationships observed [28]. Individuals with impaired hamstring muscle function, through either a history of HSI occurrence or acute fatigue, have demonstrated alterations in running kine- matics, kicking mechanics, muscle activation patterns, and lengthening muscle tissue mechanics [ . Researchers demonstrated that a previous HSI significantly reduced horizontal force production during high speed running (~80% maximum velocity) [33], potentially resulting in a decrease in peak hip flexion and the peak knee extensor moment that occurred during the late swing phase of running, although
in running kine- matics, kicking mechanics, muscle activation patterns, and lengthening muscle tissue mechanics [ . Researchers demonstrated that a previous HSI significantly reduced horizontal force production during high speed running (~80% maximum velocity) [33], potentially resulting in a decrease in peak hip flexion and the peak knee extensor moment that occurred during the late swing phase of running, although these differences were not assessed [33]. Under fatigued conditions, where it would be expected that the ability for the hamstrings to produce force would be impaired, healthy soccer players displayed significant differences in swing phase kinematics in comparison to non-fatigued con- ditions [34]. Contrastingly, Silder et al. [32] demonstrated no significant differences in mechanics or muscle activation between previously and non-previously injured limbs when running at 60-, 80-, 90-, and 100% of maximum sprinting speed. However, the impact of an HSI event on hamstring and athletic performance have been made retrospec- tively [5,29–33], where the changes in kinetics, kinematics, and muscle activation could be a result of motor adaptation following injury in order to optimise and/or to protect the system from further injury [35]. In a recent review on the current evidence for ham- string kinematics [36], the authors highlighted that as the primary mechanism of HSI is strain and that several kinematic parameters directly influence strain running kinematics could represent a modifiable risk factor for future injury [36]. However, as there is not a single driver to kinematic changes in muscular strain suggesting it is potentially an interaction between several kinematic features, such as lumbo-pelvic control, anterior pelvic tilt, forward trunk lean, trunk lateral flexion, maximal hip flexion, and likely the under pinning kinetic features, such as force production, which undoubtedly links to sprinting kinematic characteristics [36–38]. Currently many of these findings are based on retrospective studies and therefore further investigation is warranted. Applied muscular strain (i.e., running kinematics) and the hamstring strain capacity, specifically the BFLHfascicle lengths and eccentric hamstring strength, have a joint interac-
of these findings are based on retrospective studies and therefore further investigation is warranted. Applied muscular strain (i.e., running kinematics) and the hamstring strain capacity, specifically the BFLHfascicle lengths and eccentric hamstring strength, have a joint interac-
Muscles2025,4, 44 3 of 16 tion on HSI [36]. Bramah et al. [36] highlighted that lumbo-pelvic control as a parameter has the strongest level of evidence to support the risk of HSI; however, it was noted that no single biomechanical parameter has been identified as a driver for applied hamstring strain or HSI risk. Therefore, a combination of parameters including force production characteristics or strain capacity should be considered and monitored, highlighting the need for a combined observation of key features associated with applied strain and strain capacity. Quantitative evaluation of these factors requires methods capable of captur- ing both external motion and internal tissue behaviour. Three-dimensional (3D) motion capture is considered the gold standard of kinematic assessments, allowing for the quan- tification of joint angles, angular velocities, and segment coordination during high-speed running [36,39]. In contrast, hamstring strain capacity can be evaluated through a combi- nation of architectural assessments via ultrasound imaging to determine fascicle length and pennation angle [40–42], and force-based measurements, with isokinetic dynamometry providing a reliable method of assessing eccentric strength [43]. However, to date there are limited observations on the interaction between these factors. Recently, meaningful relationships have been identified between eccentric hamstring strength and late swing phase mechanics at the knee [44], with the potential that running kinematics or the task specific mechanical demands may result in physical and architectural adaptations [45,46]. Moreover, on completion of a four-week Nordic hamstring exercise training programme, improvements in the late swing phase knee mechanics were also observed [47]. The aim of the present study was to observe if spatiotemporal characteristics, running kinematics and muscle activation was related to relative eccentric hamstring strength and BFLHfascicle length. It was hypothesised that running kinematics would have a relationship with the modifiable risk factors of hamstring strain injuries. 2. Materials and Methods 2.1. Experimental Setup An observational experimental design was performed to determine the effect of run- ning kinematics and muscle activation on the modifiable risk factors of HSIs (i.e., BFLH fascicle length and eccentric hamstring strength). A priori sample size estimation based on the differences observed by Sever et al.
modifiable risk factors of hamstring strain injuries. 2. Materials and Methods 2.1. Experimental Setup An observational experimental design was performed to determine the effect of run- ning kinematics and muscle activation on the modifiable risk factors of HSIs (i.e., BFLH fascicle length and eccentric hamstring strength). A priori sample size estimation based on the differences observed by Sever et al. [48] where the smallest correlation between hip strength and running kinematics (r = 0.68, f 2 = 0.86), a priori alpha level of 0.05, minimum statistical power of 80% resulted in a required sample size of 15. Participants were observed on two separate occasions within a one-week period, each interspersed by≥48 h at the same time of day. Participants’ modifiable risk factors were assessed during the first testing occasion whereby BFLHfascicle length and isokinetic eccen- tric strength measurements for hamstrings [28]. On the second testing bout, participants performed a submaximal treadmill assessment running at 18 km/h with both 3D and EMG measurements taken [49]. 2.2. Participants Twenty-two male competitive team sport athletes (Tier 1–3 according to McKay et al. [50]) who incorporated high-speed running (24.7±4.3 years, 1.82±0.07 m, 84.9±8.5 kg ) participated within the present study. None of the participants reported that they had previously performed any structured sprint training, with exposure to technical elements being described during sport-based warm-ups. The study was approved by the institutional ethics committee (HSR1718–040) providing written consent. The study also conformed to the principles of the Declaration of Helsinki (2013).
Muscles2025,4, 44 4 of 16 2.3. Procedures 2.3.1. Muscle Architecture BFLHmuscle architecture images were collected in a prone position with the hip in neutral and the knee fully extended. All images were collected at the halfway point between the ischial tuberosity and the knee joint fold along the longitudinal axis of the muscle belly utilising a two-dimensional, B-mode ultrasound (MyLab 70 xVision, Esaote, Genoa, Italy) with a 7.5 MHz, 10 cm linear array probe with a depth resolution of 67 mm. A layer of conductive gel was placed across the probe; the probe was then placed on the skin over the scanning site perpendicular to the skin. During collection of the ultrasound images, minimal pressure was applied to the skin. The assessor manipulated the orientation of the probe slightly if the superficial and intermediate aponeuroses were not parallel. These methods are consistent to those used previously by the same authors [40]. Imaging of both limbs took between 8 and 12 min per participant. Sonograms were analysed offline with Image J version 1.52 software (National Insti- tute of Health, Bethesda, MD, USA). Images were calibrated to the known field of view, then a fascicle of interest was identified. Muscle thickness, pennation angle, observed fascicle length, and distance between fascicle end point and super fascial aponeurosis were measured 3 times within each image, to enable complete fascicle length estimation using a previously established reliable linear equation [40]. Bicep femoris estimation is given by the following equation: fascicle length = L + (h÷sin(β)). Where L is the observable fascicle length, h is the perpendicular distance between the superficial aponeurosis and the fascicles visible end point, andβis the angle between the fascicle and the superficial aponeurosis. 2.3.2. Isokinetic Eccentric Strength Participants performed a standardised warm-up following the collection of ultrasound images, including 5 min of submaximal cycling, followed by 2 sets of 5 repetitions of body weight squats, forward lunges, and leg swings. Isokinetic strength of the knee flexors was assessed using an isokinetic dynamometer (125 AP, KinCom, TN, USA) sampling at 120 Hz. Participants were seated and secured to avoid secondary joint movement,
warm-up following the collection of ultrasound images, including 5 min of submaximal cycling, followed by 2 sets of 5 repetitions of body weight squats, forward lunges, and leg swings. Isokinetic strength of the knee flexors was assessed using an isokinetic dynamometer (125 AP, KinCom, TN, USA) sampling at 120 Hz. Participants were seated and secured to avoid secondary joint movement, with the hip flexed to 90 ◦ . The range of motion (ROM) of the knee was determined as 0–90 ◦ (i.e., full extension to 90 ◦ of flexion), and the limb length and limb weight for each subject were recorded, with limb weight being measured at rest at 0 ◦ , for gravitational correction during data analysis [43]. Eccentric knee flexion was at a single standardised angular velocity (60 ◦ · s −1 ), as this angular velocity to detect future HSI risk [19]. Subjects performed 5 submaximal incremental repetitions, which were used for familiarisation purposes. Three maximal eccentric knee flexion efforts were performed, and a 60 s rest period was provided between each repetition. Participants were instructed to pull the dynamometer head as “hard and fast as possible”, with strong verbal encouragement provided during the task. Raw torque/angle data was analysed using a custom designed Excel spreadsheet (Microsoft, Redmond, WA, USA). Phases of acceleration and deceleration were initially deleted from the analysis using a tolerance of±1 ◦ ·s −1 , with the included isokinetic range gravity corrected. 2.3.3. Three-Dimension and Task Electromyography Following a standardised warm up, which is crucial for the collection of EMG data [51]. All 3D motion data was collected over a 15 s duration using infrared cameras (250 Hz) operating through Qualisys Track Manager (QTM) (Oqus 7+, Qualisys AB, Partille, Sweden) on a single treadmill (T9450 HRT Vision Fitness, Cottage Grove, WI, USA) for all running
Muscles2025,4, 44 5 of 16 trials which was situated within the 3D motion capture area. Retro-reflective markers and cluster sets of four markers were placed onto the body landmarks to define the pelvis, thigh, shank, and foot segments. Surface EMG data captured at 1500 Hz through QTM of BFLHand medial ham- strings (MH)(semitendinosus [ST] and semimembranosus) was measured for all trials. Prior to electrode placement, the participants’ skin was prepared using a standardised process of shaving (with a disposable safe razor), rubbing with a preparation gel and cleaning with an alcohol-based solution. Skin preparation was performed to minimise resistance (i.e., to reduce inter-electrode resistance to values below 5 kΩ[52]. Ag-AgCl electrodes with wireless EMG sensors and a reference pad (Noraxon U.S.A. Inc., Scotts- dale, AZ, USA) were placed on to the surface of the skin of both limbs attached in orientation with the muscle fibres. Electrodes were placed at the mid-point of the BFLH and the medial hamstrings. Correct electrode placement was confirmed prior to com- mencing data collection with manual muscle testing (i.e., by asking the participants to voluntarily contract the hamstrings against manual resistance) and minimal cross-talk will be visually and physically checked via internal and external rotation of the leg with a 90 ◦ knee angle, as per Timmins et al. [ A maximal treadmill sprint assessment was performed. This method was chosen to allow for comparisons of EMG attained at the sub-maximal running speed to be expressed as a percentage of maximal velocity. Following the sub-maximal treadmill trial, a 5 min rest period was provided. Participants performed a maximal treadmill sprint assessment to normalise task EMG data. This method was chosen to allow for comparisons of EMG attained at the sub-maximal running speed to be expressed as a percentage of maximal velocity. Following the sub-maximal treadmill trial, a 5 min rest period was provided. Participants ran at increasing velocities where they were required to maintain a set running velocity for 10 s with 180 s recovery between each trial. Commencing at 18 km·h −1 , with subsequent increases in velocity of 1.5 km·h −1 for each
as a percentage of maximal velocity. Following the sub-maximal treadmill trial, a 5 min rest period was provided. Participants ran at increasing velocities where they were required to maintain a set running velocity for 10 s with 180 s recovery between each trial. Commencing at 18 km·h −1 , with subsequent increases in velocity of 1.5 km·h −1 for each running interval, consistent with Numella et al. [54]. They continued this until they could not maintain the pace for the given duration or a rating of perceived exertion (RPE) of >9 was given, when using a scale of 1–10. This was performed on a high-speed treadmill (Woodway Ergo ELG55, Weil am Rhein, Germany), which was in a fixed position outside the 3D motion capture area. A lower extremity six degrees of freedom kinematic model was created for each par- ticipant including the pelvis, thighs, shanks, and feet using Visual 3D (V3D) (C-motion, version 3.90.21, Gothenburg, Sweden). The model utilised a CODA pelvis orientation to define the location of the hip joint centre [55]. The knee and ankle joint centres were defined as the mid-point of the line between lateral and medial markers. Hip and knee angles during the gait and sagittal plane knee and hip joint angles were deter- mined based on the 3D coordinates of one segment to another. Prior to exporting the first derivative kinematic angular data, an 8 Hz low pass filter was applied to the data to attenuate noise [ To identify three full strides for the left and right limb, gait characteristics for the left and right foot, take off (TO) and touch down (TD) events were identified. TO was identified as the moment the fifth metatarsal ascended (Z) to a height greater than this minimum threshold (0.22 m) for a minimum of 25 frames (0.1 s). Alternatively, TD was identified as the moment the fifth metatarsal reached 0.22 m for a minimum of eight frames (0.032 s). Following the identification of gait events, data was stride normalised from TD to subsequent TD, with contact time being defined as TD to TO. A description of
minimum threshold (0.22 m) for a minimum of 25 frames (0.1 s). Alternatively, TD was identified as the moment the fifth metatarsal reached 0.22 m for a minimum of eight frames (0.032 s). Following the identification of gait events, data was stride normalised from TD to subsequent TD, with contact time being defined as TD to TO. A description of kinematic running metrics identified from 3D motion data is presented in Table.
Description
The study examines the relationship between hamstring strength, architecture, and running biomechanics.