Muscle strains are a common injury among athletes, with the hamstrings being particularly susceptible to injury in sports that involve high-speed running (16,29,42,51). For example, musculotendon strains accounted for nearly half of all injuries for a National Football League team during preseason practice, with hamstring strains being the most common and requiring the most time (average of 8.3 d) away from sport (14). Similarly in the Australian Football League, hamstring strains are the single most common injury, with approximately six injuries per club per season and 33% of those being recurrent injuries (25). The susceptibility of the hamstrings to injury during high-speed running is likely linked to the biomechanical demands placed on the muscle, although debate continues as to whether injury is more likely during the stance or the swing phase of a running gait cycle (23,38,41,52). Resolving this issue is relevant for designing the type of resistance training that may be most effective for preventing initial or recurrent injuries. Specifically, injury prevention programs would ideally incorporate aspects (e.g., lower extremity postures, muscle lengths, contraction type) that are most similar to the conditions associated with injury, such that the athlete can optimize the gains in functional strength (24) and minimize the risk of future injury.
Animal models have been used to establish relationships between musculotendon mechanics and injury risk. When stretching a maximally activated muscle, the degree of muscle injury is related to the magnitude of fiber strain and negative work done by the muscle (17,30,31). However, when active stretch-shortening cycles are imposed on a musculotendon unit, the degree of strain taken up by the fibers can change with time, making the muscle potentially more susceptible to injury after repeated loading cycles (9). These insights are potentially relevant for understanding hamstring injury mechanisms. In particular, the biarticular hamstrings are known to undergo an active stretch-shortening contraction during the late swing phase of running gait (21,47), with the amount of negative work increasing significantly with running speed (10). Thus, the hamstrings may be susceptible to a late swing injury as a result of repetitive strides of high-speed running (23,41). However, others have argued that the hamstrings are more likely to be injured during stance phase when the limb is subjected to external loading via foot-ground contact (33,38). Indeed, EMG activity indicates that the hamstrings are active into stance phase (27,44,50). In addition, a recent study suggests that the hamstrings may also undergo a lengthening phase during late stance (52). However, to date, there have been limited attempts to measure ground reactions during sprinting (28,34,37) and no prior studies that have used such data to analyze hamstring kinetics during stance.
The purpose of this study was to directly compare hamstring kinetics between stance and swing phases of a sprinting gait cycle. We hypothesized that peak loading and negative work would be greater during swing than stance and that these measures would increase significantly with speed. The information obtained in this study provides a more complete understanding of hamstring demands during normal sprinting gait, which is important for scientifically assessing how specific exercises may be most effective in reducing injury rates.
Twelve subjects (9 men and 3 women) volunteered to participate in this study (mean ± SD: age = 24.5 ± 4.1 yr, height = 176.1 ± 5.0 cm, mass = 70.2 ± 8.8 kg). The average maximum treadmill running speeds were 8.0 and 7.1 m·s−1 for men and women, respectively. All subjects had no prior surgical history in their lower extremity and had no lower extremity injury and/or pain in the 3 months before testing. The testing protocol was approved by the University of Wisconsin institutional review board, and all subjects provided informed consent in accordance with institutional policies.
Whole-body kinematics were recorded using 42 reflective markers placed on the body segments of each subject; 23 of which were located on anatomical landmarks (Fig. 1A). EMG activities were recorded from single differential surface electrodes (DE-2.1; DelSys Inc., Boston, MA) placed on seven muscles of the right lower limb: biceps femoris (BF), medial hamstrings (ST-semitendinosus and SM-semimembranosus, vastus lateralis, rectus femoris, and medial gastrocnemius. Each subject ran on an instrumented treadmill at a preferred speed to warm-up before sprinting at 80%, 85%, 90%, 95%, and 100% of his or her maximum speed, with a minimum of five strides (∼5 s) collected at each speed. Two additional trials were collected to scale a biomechanical model to each subject: 1) a quiet standing position to establish segment lengths, joint centers, and joint coordinate systems, and 2) a hip circumduction movement (both right and left sides) to determine functional hip joint centers (39).
Three-dimensional kinematics were collected at 200 Hz using an eight-camera passive marker system (Motion Analysis Corporation, Santa Rosa, CA). Kinematic data were then low-pass filtered using a bidirectional, fourth-order Butterworth filter with a cutoff frequency of 12 Hz. Three-dimensional ground reaction forces and moments were also recorded at 2000 Hz using an instrumented treadmill (Bertec Corporation, Columbus, OH). These ground reactions were then low-pass filtered using a bidirectional sixth-order, low-pass Butterworth filter with a cutoff frequency of 25 Hz. Foot contact and toe-off times were identified as the times when the vertical ground reaction force went above and then fell below 50 N.
EMG activities were interfaced to an amplifier unit (Bagnoli-16, DelSys Inc.; CMRR > 84 dB at 60 Hz, input impedance > 100 MΩ) and recorded at 2000 Hz. The EMG signals were subsequently full-wave rectified and low-pass filtered using a bidirectional, sixth-order Butterworth filter with a cutoff frequency of 20 Hz. Ensemble-averaged EMG activities were generated over the sprinting gait cycle by aligning foot strikes, normalizing to peak EMG activity, and averaging activation patterns across multiple strides and subjects. For each muscle, the distribution of EMG activity over the gait cycle was quantified by computing the percentage of activity seen in three phases: stance phase (foot contact to toe off), first half of swing phase (swing 1, toe off to peak knee flexion), and second half of swing phase (swing 2, peak knee flexion to foot contact).
The body was modeled as a 14-segment, 31 degree-of-freedom (DOF) articulated linkage (Fig. 1A). Anthropometric properties of body segments were scaled to each individual using the subject's height, mass, and segment lengths (12). The functional hip joint centers were used to scale the mediolateral width of the pelvis. The hip joint was modeled as a ball and socket with three DOF. The knee joint was represented as a one DOF joint, in which the tibiofemoral translations and the nonsagittal rotations were constrained functions of the knee flexion-extension angle (49). The ankle-subtalar complex was represented by two revolute joints aligned with anatomical axes (13). The lumbar spine was represented as a ball and socket joint at approximately the third lumbar vertebra (1). For each trial, joint angles were computed at each time step using a global optimization routine to minimize the weighted sum of squared differences between the measured and the model marker positions (32).
Musculotendon actuators were represented by a series of line segments connecting the origin to the insertion, with wrapping about joints and other structures accounted for with wrapping surfaces (4). The input to each musculotendon actuator was an idealized excitation level that varied between zero and one (full excitation). Muscle excitation-to-activation dynamics was represented by a first-order differential equation that had a faster time constant during activation (10 ms) than deactivation (30 ms). A Hill-type lumped parameter model (Fig. 1C) of musculotendon contraction dynamics was used to relate activation to force, with muscle fibers assumed to be in series with an elastic tendon (40,53). Force produced by the musculotendon actuator was applied to the segment in which the tendon was attached.
Passive joint torques (equation 1) were included to account for uniarticular, passive-elastic structures at the hip, knee, and ankle, which is defined as a function of the joint angle (q) with parameters (k 1, k 2, r 1, r 2, φ1, φ2) taken from Silder et al. (43).
The equations of motion of the musculoskeletal model were derived using the SDFast (Parametric Technology Corporation, Waltham, MA) and the SIMM Pipeline (Musculographics, Inc., Chicago, IL).
Forward dynamic simulations.
We generated muscle-actuated forward dynamic simulations of sprinting to characterize hamstring stretch, force, and work. A total of 52 musculotendon actuators (26 actuators on each limb) were used to actuate two DOF on each limb (hip flexion/extension and knee/flexion extension). All other DOF were prescribed to follow the measured kinematic trajectories, thereby accounting for intersegmental dynamics (54). Simulations were generated for a minimum of five strides at each speed for each subject (Fig. 1D).
A computed muscle control algorithm was used (10,45) to generate muscles excitations that drove the forward dynamic model to closely replicate experimental hip and knee kinematics. Muscle redundancy was resolved by using numerical optimization (20) to minimize the sum of the squared weighted normalized contractile element forces (equation 2 and Fig. 1B):
where the weighting factor, wi, is the volume of muscle i, Fi CE is the contractile element force of muscle i, fi (li M) is an active force-length scaling factor, and Fi max is the maximum isometric force of muscle i. We included constraints in the optimization problem (no hamstring excitation from 20% to 40% gait cycle, and no rectus femoris excitation from 15% to 35% gait cycle) to ensure that the hamstring and rectus femoris muscles were not excited when EMG data indicate they are not normally active (10).
Simulations were used to characterize the musculotendon stretch, force, and power development of the biarticular hamstrings: BF, ST, and SM. Musculotendon stretch was defined as the change in length of the musculotendon unit from an upright posture. Upright musculotendon lengths were computed by setting all joint angles to zero in the subject-specific scaled model. The musculotendon power generated (absorbed) was computed as the product of the force and musculotendon velocity. The negative and positive musculotendon work was computed by integrating the respective negative and positive portions of the power curves. The three biarticular hamstring musculotendon work quantities were then added to obtain a measure of net-negative and net-positive work.
A two-way repeated-measures ANOVA was used to determine the effects of gait cycle phase and normalized speed (80%, 85%, 90%, 95%, and 100%) on the magnitude of peak force. A one-way repeated-measures ANOVA was then used to determine the effect of speed on net-negative and net-positive musculotendon work. Tukey's post hoc test was used to analyze significant main effects. The statistical analyses were completed using Statistica (Version 6.0; StatSoft, Inc., Tulsa, OK), with a significance level of 0.05 for all comparisons.
The computed muscle control algorithm generated simulations that closely tracked the experimental kinematics. The root mean squared error values for the hip and knee flexion-extension angles were 1.5° ± 0.5° for hip flexion/extension and 2.8° ± 0.8° for knee flexion/extension during the time when the hamstrings were active. The pattern of the predicted muscle excitations was generally similar to the measured EMG activities (Fig. 2). Mean hamstring excitations tended to be slightly less than the measured EMG activities during the second half of swing phase but were then greater than the measured EMG magnitudes during stance (Fig. 2, Table 1).
The hamstring musculotendons lengthened from approximately 50% to 90% of the gait cycle, with negative work done only during swing phase (Fig. 3). The hamstrings shortened and did positive work from 90% of the gait cycle throughout the subsequent stance phase. Both net-negative and net-positive musculotendon work increased significantly (P < 0.001) with speed (Fig. 4); however, net-negative hamstring work increased at a faster rate than net-positive work as speed increased (Fig. 4).
Two distinct loading peaks were present for the hamstrings (Fig. 3): first, during late swing phase (between 85% and 95% of the gait cycle), and second, during early stance phase (between 0% and 15% of the gait cycle). A significant (P < 0.001) speed × phase interaction was present for peak load of the BF and SM. Peak musculotendon force during swing increased significantly (P < 0.001) with speed (Fig. 3) for all three muscles (Table 2), whereas peak force during stance was independent of speed (Table 2, Fig. 4). Peak swing phase force at the fastest speed exceeded peak stance phase force for the BF. At all speeds, the SM swing phase force exceeded the stance phase peak force. ST peak forces were not significantly different between swing and stance phase at the fastest speed (Table 2).
In this study, we extended our prior analysis of hamstring kinetics during swing (10,46,48) to stance phase, thereby providing insights into when the hamstrings seem most susceptible to injury. The sprinting simulations predict that BF loading increases with speed during swing but not stance and further that peak stretch and negative work demands occur exclusively during swing. These data lend further evidence to the belief that inertial loads put the lengthening hamstrings at risk for injury during high-speed forward running (10,23,41). We do note that our analyses are limited to constant speed forward running, similar to the types of hamstring loading observed in a track sprinter. The hamstring loads that occur during sudden accelerations and turns could certainly factor into injury risk and should be considered an important area of further study.
Prior studies of sprinting kinetics have been limited to joint level analyses (28,34,44). However, characterizing injury risk solely on the basis of joint analysis is challenging because of the biarticular nature of the hamstrings. For example, peak hip and knee angles vary significantly with running speed (28,47), whereas peak hamstring stretch remains invariant as sprinting speed increases (10). Although increased hip flexion causes the hamstrings to be stretched, this stretch can be offset by shortening from greater knee flexion at fast speeds (47). Similarly, it has been shown that hip and knee joint moments and powers increase with speed (28), but the associated biomechanical demands on the hamstrings are more difficult to determine from joint level analyses.
Our results suggest that the hamstrings are substantially loaded during both the stance and the swing phases of high-speed treadmill running. However, we found that the hamstrings only undergo a lengthening contraction during late swing phase, and thus, negative muscle work is constrained to this phase. This result is similar to several prior kinematic studies (10,21,47,50) but contrasts with Yu et al. (52), who reported that the hamstrings undergo a second lengthening period during the late stance phase of overground running. Conceptually, hamstring lengths are primarily a function of hip and knee flexion angles. For example, the hamstrings are lengthening during late swing when the hip is flexing and the knee is extending because both of these motions contribute to hamstring stretch. However, the hip starts to extend before foot contact and continues to extend throughout stance (47); the knee flexes until mid stance and then extends until toe off. Therefore, if hamstring stretch is to occur during the second half of stance, the hamstring lengthening because of knee extension would have to exceed the hamstring shortening because of hip extension. Further, because the hamstring moment arms at the hip are greater than at the knee (13,47), the knee extension velocity would have to substantially exceed the hip extension velocity. However, sagittal hip and knee angular velocities during stance tend to be of comparable magnitude (28), making it unlikely that hamstring stretch is occurring during stance.
Animal models of muscle injury show that active lengthening contractions are linked with injury (31) and that the degree of injury seems to be associated with the negative work done by a maximally activated muscle (8). In this study, we have shown that negative work performed by the hamstrings occurs exclusively during the swing phase, suggesting that the likelihood of a strain injury is greater during swing than stance. Further, hamstring loading was found to increase significantly with speed during swing but not stance, providing further support for injury susceptibility during the swing phase of high-speed running. Repeated stretch shortening cycles have been shown to increase risk for injury in animal models (9), such that repeated strides of high-speed running may contribute to the injury risk susceptibility.
There are several assumptions that are important when interpreting the model-based results in this study. First, we relied on literature-derived estimates for parameters such as tendon compliance, maximum isometric force, optimum fiber length, and musculoskeletal geometry (4,13). We have previously shown that tendon compliance can substantially influence estimates of muscle fiber stretch and work (46). For this reason, we limited our comparisons in this study to net musculotendon stretch and work, which are less dependent on specific musculotendon parameters (46). In addition, we note that the speed-dependent changes in peak loading and negative work were similar (Table 2) for all three biarticular hamstrings (which have different force capabilities, fiber lengths, and geometries), which supports the idea that specific parameter assumptions did not predetermine the observed results.
A numerical objective function, the sum of weighted squared normalized forces (11), was used to resolve muscle redundancy. Although predicted excitations were of comparable magnitude with normalized EMG activities during late swing (Table 1), the predicted BF excitation onsets were slightly delayed relative to EMG phasing (Fig. 2). We used sensitivity analyses to determine how dependent the BF musculotendon force predictions are on excitation onset. We found that shifting the BF excitation onset time to 5% earlier in the gait cycle increased the swing phase peak force by 4% while having a small (<0.5%) effect on stance phase forces. These differences are less than the force discrepancies seen between stance and swing at the fastest running speed (Fig. 3). Predicted hamstring excitation magnitudes during stance tended to be larger than the measured EMG activities, whereas gastrocnemius excitations tended to be underpredicted in this phase (Fig. 2, Table 1). This could be a result of not including active ankle actuation in the simulations, such that gastrocnemius excitation was not required to actively generate an ankle plantarflexion moment. Reduced gastrocnemius excitation also diminishes its contribution to the knee moment, and as a result, hamstring excitation in early stance could have been overestimated.
Measured ground reactions were directly applied to the feet rather than dealing with the complexity of modeling the foot-floor interactions (2,18,19,35,36). We low-pass filtered the ground reactions at 25 Hz, which is lower than what is often done when analyzing running mechanics (15,26). This decision was based on a recent study by Bisseling and Hof (6), who suggested that using substantially different cutoff frequencies for kinematic and force data can introduce artificial peaks into computed joint kinetics (6). The 25-Hz cutoff frequency adopted for this study was selected to mitigate this effect while retaining the high-frequency information present in the ground reactions at contact. It is also important to recognize that one of our primary outcome measures (musculotendon work) is an integrated quantity rather than an instantaneous quantity and thus is less susceptible to choice of filter cutoff frequencies. However, we do note that the calculation of peak biomechanical loads during impact remains a challenging problem, and additional work on the appropriate selection of kinematic and kinetic filters (6) is warranted.
Defining the musculotendon demands during sprinting at the likely time of injury provides direction as to the type of resistance training that may prove most beneficial in preventing an initial or recurrent hamstring injury (24). For example, if injury occurs during the swing phase, which we suggest, then resistance training of the hamstrings may be most effective when performed with lower extremity postures that place the muscle in an elongated position (e.g., hip flexion with knee extension) (22). Further, lengthening (eccentric) contractions under inertial load would likely be an essential component of the prevention and rehabilitation strategy. This contraction specificity with respect to hamstring injury prevention is supported in prior studies that demonstrated the success of an eccentric-focused resistance training program (3,5,7). Finally, characterizing the load during the sprinting gait cycle enables the relative intensity of the resistance training to be more scientifically determined.
In conclusion, our data suggest that the hamstrings are at greater risk for injury during the late swing phase of high-speed running when compared with stance phase. Late swing phase is when the biomechanical demands placed on the hamstrings seem most consistent with muscle injury mechanisms. On the basis of the results from this study, a rehabilitation program incorporating eccentric focused exercises is likely to have greater benefit over programs that focus on concentric loading of the hamstrings.
The authors acknowledge support from the National Institutes of Health (grants AR 56201 and 1UL1RR025011), the NFL Charities, and an American Association of University Women Dissertation Fellowship to E. Chumanov.
The results of this study do not constitute endorsement by the American College of Sports Medicine.
1. Anderson FC, Pandy MG. A dynamic optimization solution for vertical jumping in three dimensions. Comput Methods Biomech Biomed Engin
2. Anderson FC, Pandy MG. Dynamic optimization of human walking. J Biomech Eng
3. Arnason A, Andersen TE, Holme I, Engebretsen L, Bahr R. Prevention of hamstring strains in elite soccer: an intervention study. Scand J Med Sci Sports
4. Arnold AS, Salinas S, Asakawa DJ, Delp SL. Accuracy of muscle moment arms estimated from MRI-based musculoskeletal models of the lower extremity. Comput Aided Surg
5. Askling C, Karlsson J, Thorstensson A. Hamstring injury occurrence in elite soccer players after preseason strength training with eccentric overload. Scand J Med Sci Sports
6. Bisseling RW, Hof AL. Handling of impact forces in inverse dynamics. J Biomech
7. Brooks JH, Fuller CW, Kemp SP, Reddin DB. Incidence, risk, and prevention of hamstring muscle injuries in professional rugby union. Am J Sports Med
8. Brooks SV, Zerba E, Faulkner JA. Injury to muscle fibres after single stretches of passive and maximally stimulated muscles in mice. J Physiol
9. Butterfield TA, Herzog W. Quantification of muscle fiber strain during in-vivo repetitive stretch-shortening cycles. J Appl Physiol
10. Chumanov ES, Heiderscheit BC, Thelen DG. The effect of speed and influence of individual muscles on hamstring mechanics during the swing phase of sprinting. J Biomech
11. Crowninshield RD, Brand RA. A physiologically based criterion of muscle force prediction in locomotion. J Biomech
12. de Leva P. Adjustments to Zatsiorsky-Seluyanov's segment inertia parameters. J Biomech
13. Delp SL, Loan JP, Hoy MG, Zajac FE, Topp EL, Rosen JM. An interactive graphics-based model of the lower extremity to study orthopaedic surgical procedures. IEEE Trans Biomed Eng
14. Feeley BT, Kennelly S, Barnes RP, et al. Epidemiology of National Football League training camp injuries from 1998 to 2007. Am J Sports Med
15. Ferber R, Davis IM, Williams D Sr. Gender differences in lower extremity mechanics during running. Clin Biomech
16. Gabbe BJ, Finch CF, Bennel KL, Wajswelner H. Risk factors for hamstring injuries in community level Australian football. Br J Sports Med
17. Garrett WE Jr, Safran MR, Seaber AV, Glisson RR, Ribbeck BM. Biomechanical comparison of stimulated and nonstimulated skeletal muscle pulled to failure. Am J Sports Med
18. Gerritsen K, van den Bogert AJ, Nigg BM. Direct dynamics simulation of the impact phase in heel-toe running. J Biomech
19. Gilchrist LA, Winter DA. A two-part, viscoelastic foot model for use in gait simulations. J Biomech
20. Happee R. Inverse dynamic optimization including muscular dynamics, a new simulation method applied to goal directed movements. J Biomech
21. Hay JG, Thorson EM, Kippenhan BC. Changes in muscle-tendon length during the take-off of a running long jump. J Sports Sci
22. Heiderscheit B, Sherry M, Silder A, Chumanov E, Thelen D. Acute hamstring injuries: recommendations for diagnosis rehabilitation and injury prevention. J Orthop Sports Phys Ther
23. Heiderscheit BC, Hoerth DM, Chumanov ES, Swanson SC, Thelen BJ, Thelen DG. Identifying the time of occurrence of a hamstring strain injuring during treadmill running: a case study. Clin Biomech
24. Hortobágyi T, Hill J, Houmard J, Fraser D, Lambert N, Israel R. Adaptive responses to muscle lengthening and shortening in humans. J Appl Physiol
25. Hoskins WT, Pollard H. Injuries in Australian rules football: a review of the literature. Australas Chiropr Osteopathy
26. Hunter JP, Marshall RN, McNair PJ. Interaction of step length and step rate during sprint running. Med Sci Sports Exerc
27. Jonhagen S, Ericson MO, Nemeth G, Eriksoon E. Amplitude and timing of electromyographic activity during sprinting. Scand J Med Sci Sports
28. Kuitunen S, Komi PV, Kyrolainen H. Knee and ankle joint stiffness in sprint running. Med Sci Sports Exerc
29. Kujala UM, Orava S, Jarvinen M. Hamstring injuries: current trends in treatment and prevention. Sports Med
30. Lieber RL, Friden J. Muscle damage is not a function of muscle force but active muscle strain. J Appl Physiol
31. Lieber RL, Friden J. Mechanisms of muscle injury gleaned from animal models. Am J Phys Med Rehabil
. 2002;81(11 suppl):S70-9.
32. Lu TW, O'Connor JJ. Bone position estimation from skin marker co-ordinates using global optimisation with joint constraints. J Biomech
33. Mann R, Hagy J. Biomechanics of walking, running, and sprinting. Am J Sports Med
34. Mann RV. A kinetic analysis of sprinting. Med Sci Sports Exerc
35. Neptune RR, Kautz SA, Zajac FE. Contributions of the individual ankle plantar flexors to support, forward progression and swing initiation during walking. J Biomech
36. Neptune RR, Wright IC, van den Bogert AJ. A method for numerical simulation of single limb ground contact events: application to heel-toe running. Comput Methods Biomech Biomed Eng
37. Novacheck TF. Walking, running, and sprinting: a three-dimensional analysis of kinematics and kinetics. Instr Course Lect
38. Orchard J, Best TM. The management of muscle strain injuries: an early return versus the risk of recurrence. Clin J Sport Med
39. Piazza SJ, Erdemir A, Okita N, Cavanagh PR. Assessment of the functional method of hip joint center location subject to reduced range of hip motion. J Biomech
40. Sandercock TG, Heckman CJ. Force from cat soleus muscle during imposed locomotor-like movements: experimental data versus Hill-type model predictions. J Neurophysiol
41. Schache A, Wrigley T, Baker R, Pandy M. Biomechanical response to hamstring muscle strain injury. Gait Posture
42. Seward H, Orchard J, Hazard H, Collinson D. Football injuries in Australia at the elite level. Med J Aust
43. Silder A, Whittington B, Heiderscheit B, Thelen DG. Identification of passive elastic joint moment-angle relationships in the lower extremity. J Biomech
44. Swanson S, Caldwell G. An integrated biomechanical analysis of high speed incline and level treadmill running. Med Sci Sports Exerc
45. Thelen DG, Anderson FC. Using computed muscle control to generate forward dynamic simulations of human walking from experimental data. J Biomech
46. Thelen DG, Chumanov ES, Best TM, Swanson SC, Heiderscheit BC. Simulation of biceps femoris musculotendon mechanics during the swing phase of sprinting. Med Sci Sports Exerc
47. Thelen DG, Chumanov ES, Hoerth DM, et al. Hamstring muscle kinematics during treadmill sprinting. Med Sci Sports Exerc
48. Thelen DG, Chumanov ES, Sherry M, Heiderscheit BC. Neuromusculoskeletal models provide insights into the mechanisms and rehabilitation of hamstring strains. Exerc Sport Sci Rev
49. Walker PS, Rovick JS, Robertson DD. The effects of knee brace hinge design and placement on joint mechanics. J Biomech
50. Wood G. Biomechanical limitations to sprint running. In: van Gheluwe B, Atha J, editors. Medicine and Sport Science, Current Research in Sports Biomechanics
. Vol. 25: New York: Karger, Basel; 1987. p. 58-71.
51. Woods C, Hawkins RD, Maltby S, Hulse M, Thomas A, Hodson A. The football association medical research programme: an audit of injuries in professional football-analysis of hamstring injuries. Br J Sports Med
52. Yu B, Queen R, Abbey A, Liu Y, Moorman C, Garrett W. Hamstring muscle kinematics and activation during overground sprinting. J Biomech
53. Zajac FE. Muscle and tendon: properties, models, scaling and application to biomechanics and motor control. Crit Rev Biomed Eng
54. Zajac FE, Gordon ME. Determining muscle's force and action in multi-articular movement. Exerc Sport Sci Rev
Keywords:©2011The American College of Sports Medicine
ACUTE STRAIN INJURY; MOTION ANALYSIS; FORWARD DYNAMIC SIMULATION; MUSCULOSKELETAL MODEL; MUSCLE MECHANICS