To move the body forward during running, humans have to produce lower limb joint moments generated by muscle force. Therefore, it is critical to explore which muscles are important for achieving fast running for athletes and coaches.
It has been previously demonstrated that the force/work generated by shank muscles and their contributions to vertical ground reaction force increased with running speed, but the values reached plateau at medium-paced running (1,2). Because maximum force generated by a muscle is associated with its size (3), the relationship between the sprinters’ hip and thigh muscularity and their maximal running time would be of particular interest. Among the muscles, muscle sizes of hip extensor, hamstrings and gluteus maximus, and flexor psoas major (PM) were associated with the 100-m race time (4). In contrast, regarding the muscularity of the knee extensor quadriceps femoris, the largest muscle groups in the thigh (5), its impact on running performance remains controversial (4,6,7). For example, a negative correlation was reported between the muscle thickness of the quadriceps femoris and the 100-m race time, that is, a better 100-m race time was associated with thicker muscles (6). In contrast, an opposite effect of the quadriceps femoris size on the 100-m race time was reported (4,7). The limitations of previous studies are as follows: 1) With the exception of one study (8), previous work did not distinguish the biarticular (hip flexor as well as knee extensor) rectus femoris (RF) and monoarticular knee extensor vasti (vastus lateralis [VL], vastus medialis [VM], vastus intermedius [VI]) when evaluating the muscle sizes, and 2) the interrelations among sprinters’ muscularity, their joint kinematics/kinetics, and running speed during maximal sprint running have never been examined. These are critical considering the different roles between RF and vasti as well as the substantial importance of hip flexor muscles during the swing phase at fast speeds, such as >7 m·s−1 of center of gravity (CG) velocity (1). Moreover, although size of the hip flexor PM was associated with the 100-m race time (4,7), the RF volume, which was previously ignored, was greater than the PM volume (5). Because muscle volume is strongly associated with joint moment (9) and power (10), it is possible that RF plays as substantial a role as the hip flexor muscle during the swing phase to achieve a fast running speed.
The aims of the present study were 1) to identify the muscularity of the individual thigh muscles (i.e., muscles existing along the femur) and PM of sprinters through comparison with those of nonathletes and 2) to examine the interrelations among the muscularity, hip joint mechanics, and running speed during maximal running. A simulation study (1) demonstrated that the muscle force generated by the vasti showed no speed effects during stance. Therefore, we hypothesized that 1) PM and knee- and hip-crossing thigh muscles, but not the vasti, are hypertrophied in sprinters and that 2) RF volume is associated with hip joint kinetics and running speed.
Fifteen male sprinters (21 ± 2 yr, 174 ± 4 cm, 66 ± 5 kg; mean ± SD) and 12 physically and age-matched untrained men (21 ± 2 yr, 173 ± 5 cm, 66 ± 6 kg) participated in this study. The sprinters had performed competitive sprinting for >4 yr. Their best 100-m sprint times ranged from 10.63 to 11.57 s (11.09 ± 0.27 s). None of the untrained men habitually undertook resistance training or sport activities for at least 2 yr. This study was approved by the Ethics Committee on Human Research of Waseda University. Before the experiment, the participants were informed of the purpose and risks of the study, and they provided written informed consent.
Magnetic resonance imaging measurement
To examine the sprinters’ muscularity by comparison with untrained controls, using magnetic resonance (MR) scanner (Signa 1.5 T, GE Healthcare, USA), T1-weighted MR images of each thigh (echo time, 10 ms; repetition time, 520 ms; matrix, 256 × 192; field of view, 240 mm; slice thickness, 10 mm; gap, 0 mm) and the trunk (echo time, 15.6 ms; repetition time, 860 ms; matrix, 256 × 192; field of view, 450 mm; slice thickness, 10 mm; gap, 0 mm) were obtained, according to a previous study protocol (5). We analyzed images of the side on which the lower limb was grounded on the force plate (right, 13; left, 2) in sprint performance measurement (see below) for male sprinters and the right side for the untrained men. Digitization of the outlines of each muscle of the quadriceps femoris and hamstrings (biceps femoris short head [BFs], biceps femoris long head [BFl], semitendinosus [ST], and semimembranosus [SM]), total adductors (sum of the adductor magnus, adductor longus, adductor brevis, and pectineus), sartorius (Sar), tensor fasciae latae (TFL), gracilis (Gr), and PM were performed, and the anatomical cross-sectional areas (ACSA) of the muscles were determined with ImageJ software (National Institutes of Health, USA). We determined the total adductor ACSA because of difficulty in identifying individual adductor muscles along the thigh. Digitization was performed twice for each image by an investigator without knowledge of the participant’s group or kinetics/kinematics data, and the mean values were used for later analysis. Intertester repeatability of digitization has been reported previously (5). Briefly, the coefficient of variation (CV) and intraclass correlation coefficients of thigh muscles’ ACSA measurements were 2.8% ± 2.0% and 0.997, respectively (5). Muscle volume was calculated by summing ACSA by slice thickness (1 cm) and was normalized to body mass to compare the muscularity between different participants’ groups (11).
Sprint performance measurement
For sprinters, to investigate the interrelations among the muscularity, hip joint mechanics, and running speed, three-dimensional whole body kinematics and kinetics data were collected during maximal sprinting. All sprinters were asked to perform maximal-effort 50-m sprinting from standing position, with at least two trials (mean, two trials; range, two to four trials), as with the previous similar data collection setting (12). At approximately 35-m point, three-dimensional reflective marker positions and ground reaction forces were simultaneously recorded with an 8-camera motion capture system (Motion Analysis Corporation, USA) and two 120 cm × 60 cm force plates (Bertec Corporation, USA) in series sampling at 240 and 2400 Hz, respectively. The camera was calibrated to define a capture volume (4.5 m length, 1.2 m width, 2.0 m height), and the average three-dimensional residual was <0.5 mm.
A total of 68 reflective markers were attached to the participants (Fig. 1). Three to six markers per segment were used to track the position and orientation of the segments. For the foot segment, reflective markers were placed on the upper shoe heel, lower shoe heel, and lateral side of the shoe, below the lateral malleoli. To minimize soft tissue artifact as much as possible, four-marker clusters were placed on rigid shells over the distal part of the shank and thigh of each leg and the upper arm segment. For the pelvis, six markers were attached to the following anatomical positions: anterior superior iliac spine, posterior superior iliac spine, and iliac crest of both sides. For the head segment, four markers were attached to the anterior and posterior parts of the head at the level of ear on both sides. For the chest segment, markers were attached to the sternum notch, xiphoid process, and the spinous processes of C7 and T8.
Before the maximal running trials, a static standing trial was recorded to identify the tracking marker positions relative to the anatomical segment coordinate system and joint centers for each segment. The participants stood with their feet hip-width apart, with the hip and knee joints fully extended. When recording this trial, additional anatomical markers were placed on the medial and lateral malleolus, medial and lateral epicondyle of each leg to determine the joint centers and segment coordinate systems. Participants were then asked to perform hip circumduction and knee flexion–extension movements in a standing position 10 continuous times for each leg to identify the participant-specific joint centers. After recording the functional movement trials, the anatomical markers attached to the medial bony landmarks were removed. Next, all participants completed a warm-up comprising slow jogging and submaximal running at 3 × 50%, 3 × 60% to 70%, and 3 × 85% of maximal effort to familiarize themselves with the experimental conditions. In maximal running trials, we asked the participants to ground the force plates without conscious thought. We adjusted the starting position ± 1 m to enable the participants to hit the force plates with their natural running style.
The joint center mislocation has been shown to substantially affect the joint moments (13); hence, the participant-specific joint center was determined using the symmetrical center of rotation estimation method (14) proposed for ball-and-socket joints to identify the hip joint center from hip circumduction movements. Similarly, the functional tibiofemoral axis of rotation was determined from the knee flexion–extension movements using the symmetrical axis of rotation analysis (15). The medial and lateral epicondyle markers were then projected onto the estimated tibiofemoral axis, and the knee joint center was determined as the midpoint between the projected medial and lateral epicondyle markers. For these procedures, the marker position data were low-pass filtered using a fourth-order zero lag Butterworth filter with a cutoff frequency of 12 Hz, and then the method proposed by Söderkvist and Wedin (16) was used to easily obtain the technical frames of pelvis, thigh and shank segments and to minimize soft tissue artifacts of the markers used to track the position and orientation of the segments during functional movements. Once the joint centers were determined, right-handed orthogonal anatomical segment coordinate systems were defined for the pelvis segment and the foot, shank, and thigh segments of each leg based on the anatomical markers and joint centers determined in the neutral trial (Fig. 1), as per the method used in our previous study (17).
To minimize possible errors associating with the limited capture volume size, the data for 1 step from the foot strike of the stance leg to that of the swing leg were analyzed. The marker position data and ground reaction force were low-pass filtered using a fourth-order zero lag Butterworth filter with a cutoff frequency of 12 and 100 Hz, respectively. The method proposed by Söderkvist and Wedin (16) was utilized to optimally determine the anatomical coordinate systems and joint centers during running trials. Foot strike/off was defined when the ground reaction force was above/below 20 N. The joint moments were determined with a Newton–Euler inverse dynamics algorithm (18), and the resultant joint moment was expressed in the joint coordinate system. The total hip flexion angular impulse was calculated as the sum of the integral of the hip flexion moment for the stance and swing legs. The moment and angular impulse were normalized to body mass. The inertial properties of the segments were estimated based on the participants’ masses and heights (19). The CG velocity was calculated as the average value over the step. The best sprinting trial based on the CG velocity was used for further analysis.
Descriptive data are presented as means ± SD. The significance level was set at P < 0.05. A two-way analysis of variance was performed to determine the significance of participants’ group (sprinters and untrained men) effect and muscle (VL, VM, VI, RF, PM, Sar, TFL, BFs, BFl, ST, SM, and Gr) or muscle group (total quadriceps femoris, hamstrings, and adductors) effect on normalized volume to body mass. When a significant interaction or main effect of participants’ group was detected, post hoc tests with Bonferroni multiple comparisons were performed. The CV among the participants was calculated to describe the variability when needed. The relationships between normalized muscle volume that was greater in sprinters and CG velocity, between hip joint kinetics and CG velocity, and between normalized RF or PM volumes and hip joint kinetics were tested using Pearson product moment correlation coefficient (r). The P values of correlations were corrected with Benjamini and Hochberg method (20). The smallest practically important correlation was assumed to be 0.1 (21). According to a previous suggestion (21), the level of confidence interval (CI) of r was determined at 90%. The 90% CI for the correlations were adjusted using the corrected P values (22). We considered the relationship to be substantial if a lower limit of 90% CI was ≥0.10 for a positive r value and an upper limit of 90% CI ≤−0.10 for a negative r value, in addition to P < 0.05.
A significant participants’ group–muscle interaction was found for normalized individual muscle volume to body mass (P = 0.015). The normalized thigh muscle volumes for the hip flexors (RF, Sar, and TFL), hip extensors (BFl, ST, and SM), Gr, and PM (Fig. 2) were significantly greater in the sprinters than in the untrained men (P ≤ 0.001–0.038), whereas those of the monoarticular knee extensor (VL, VM, and VI) and knee flexor (BFs) thigh muscles were not significantly different between the two groups. A significant main effect of the participants’ group (P < 0.001) without any interaction was shown for normalized total muscle volume. The male sprinters had significantly greater volumes of the total hamstrings (P < 0.001) and adductors (P < 0.001) but not of the quadriceps femoris (Fig. 2).
The ground reaction force and hip joint moment data in sprinters are presented in Figure 3. The CG velocity was 8.8 ± 0.3 m·s−1 (CV = 3.4%). Figure 4 shows the relationships between the parameters. There were substantial correlations between the CG velocity and normalized volumes of RF (P = 0.045), between the CG velocity and normalized hip flexion angular impulse (P = 0.045), and between the normalized peak value of hip flexion moment for the swing leg and normalized RF volume (P = 0.045). In contrast, CG velocity was not correlated with other normalized muscle volumes (Table 1). No significant relationships were found between the hip joint kinetics and PM volume.
To address the unanswered question of which muscles are the most important for sprinters, we focused on the thigh and PM muscles, with special interest on the hip flexor and knee extensor individual thigh muscles and PM. It has been suggested that the hip flexor and extensor muscles play important roles in achieving a faster running speed, particularly during the swing phase (1). In agreement with the suggestions of the above study (1), the male sprinters in the present study had greater volumes for the hip flexor and extensor thigh muscles compared with untrained men. The present study is a cross-sectional study and thus causality between the muscularity and running speed remains unclear. However, the RF volume was associated with the hip joint kinetics as well as running speed. Overall, these findings supported our hypotheses and suggested that RF plays a significant role in fast running.
The sprinters had greater RF muscularity compared with untrained controls, and this muscularity was related to running performance. In addition, the CG velocity and RF volume were positively correlated with the hip flexion angular impulse or peak value of the hip flexion moment for the swing leg. During the swing phase where hip flexion moment was produced, hip extensor muscles are hardly activated (1). Therefore, the substantial effect of hip extension moment on the relationship between RF volume and hip joint kinetics is unlikely. A greater hip flexion moment and thus angular impulse would result in a greater angular momentum of the thigh, which vigorously accelerates the thigh through the air and provides the fast swing movement needed to achieve high stride frequency. Regarding another large hip flexor PM, its volume was greater in the sprinters than in the controls, but the volume was not correlated with either the CG velocity or hip joint kinetics. The peak value of the hip flexion moment for the swing leg occurred approximately at the instant of foot strike for the stance leg (Fig. 3). A simulation study (1) demonstrated that towards the instant of stance leg’s foot strike, muscle force generated by RF increases, whereas the force generated by PM decreases, and the RF force reaches almost maximum at the instant. During this early phase of the swing, based on the previous study (1), RF and PM could produce force eccentrically and concentrically, respectively. Therefore, the velocity-dependent force-generating capability of RF may be hardly limited by increasing running speed, whereas that of PM would be diminished with increase in speed. Indeed, the magnitude of the maximal RF force during the swing phase was greater at faster running speeds (1). Additionally, muscle volume was a predictor of the strength during slow but not fast concentric contractions (23). It is thus possible that such intermuscle differences in velocity-dependent force potential during the swing phase are related to the significant (RF) or insignificant (PM) correlations between the muscle volume and kinetic variable. Furthermore, RF can act in a stretch-shortening manner during the swing phase, likely resulting in a greater muscular power. Taken together, it is likely that the greater RF in the sprinters substantially contributes to higher running speed by generating a greater hip flexion moment and angular impulse during the swing phase.
We did not find greater total quadriceps femoris muscularity in the sprinters than in the controls (Fig. 2). In contrast, faster sprinters had a greater thickness in the proximal anterior portion of the thigh (probably the combined RF and VI thickness), and this thickness was negatively correlated with the 100-m sprint time (6). The proportional contribution of RF thickness to the previously measured thickness (6) would be relatively high, considering the distribution of ACSA along the thigh (5). Taking into account the lack of VI hypertrophy in the sprinters (Fig. 2), the aforementioned results (6) may result from the hypertrophied muscularity of RF in faster sprinters. Regarding the role of the vasti during sprinting, according to a simulation study (1), the force generated by the vasti and its contribution to the vertical ground reaction force are not affected by running speed. This has been indirectly supported by actual joint kinetics data (2). At a running speed of 9.0 m·s−1, which is a slightly higher value than that in the current study (8.8 ± 0.3 m·s−1), the muscular peak force of the vasti during stance can reach 4.9 times the body weight (1). The magnitude of the generated force would be small compared with that generated during resistance exercises such as a squat (~7 kN, (24). Moreover, the duration of vasti contraction for each stance phase is at most 100 ms (25), indicating that the total contraction duration during a 100-m sprint is <3 s (~30 steps per leg). Therefore, the mechanical stress for the vasti during sprint running may be insufficient to induce substantial muscular hypertrophy. Furthermore, the knee extension moment generated by RF during stance can compensate for the vasti action (1). Accordingly, it may be the case that hypertrophied vasti are not necessary to achieve running speeds up to the 9 m·s−1 that were apparent in the current study.
In addition to the preceding argument, the sprinters’ lack of hypertrophied musculature of the vasti may be related to performance level. We observed that male sprinters with a personal best for the 100-m race of 11.09 ± 0.27 s had similar vasti volumes to untrained controls. A previous study (4) suggested that excessive hypertrophy of the total quadriceps femoris could be disadvantageous for sprint ability, and the participants’ season-best for the 100-m race was 10.94 ± 0.39 s. In another study (26), there was no significant difference in the total quadriceps femoris ACSA between male sprinters and controls, with the sprinters’ personal best time for the 100-m race being 11.10 ± 0.44 s. Contrary to these findings, a group of National College Athletic Association Division I sprinters in the US had approximately 30% greater vasti volumes compared with controls (8), and Kumagai et al. (6) reported that the personal best of faster sprinters who had greater proximal anterior thigh thickness was 10.58 ± 0.23 s. Top-level sprinters and competition level sprinters, therefore, may have different vasti muscularity. Alternatively, it may be that resistance training regimen among top-level sprinters contributes to hypertrophy of the vasti where this increased muscle mass may not actually improve sprint performance; more work on this question is needed. Similar concepts may be applied to BFs, another nonhypertrophied muscle, in the present study, because top level sprinters reportedly had 26% greater BFs volume than nonsprinters (8), and a large increase in biomechanical load by knee flexor muscles was shown when running speed progressed toward maximal sprint (2). Moreover, BFs, but not the other three biarticular hamstrings, performed negative work during stance at a similar running speed to that observed in the current study (27). It might be possible that the vasti and BFs activations contribute to postural stability during fast running.
There are some limitations associated with the present study. First, kinetics and kinematics data were obtained at the 35-m point during the acceleration phase. If the data were collected at a constant speed and/or the deceleration phase, the results may be different. However, for competition level sprinters, such as the current participants, the running speed can reach nearly maximum at this point (28); therefore, a substantial effect of the measurement point on the main findings is unlikely. Second, it is unclear whether the current findings are applicable to faster sprinters (e.g., CG velocity > 10 m·s−1). The 40% greater RF volume in the National College Athletic Association Division I sprinters than in the controls (8) may suggest that the present findings are further highlighted in top level sprinters. Third, we did not evaluate iliacus and gluteus maximus musculature, and thus our data do not provide complete picture of the relationships between hip flexor/extensor muscularity and running performance in sprinters. Finally, the relatively small sample size (n = 15) may have affected the correlation results. The present study involves the exploratory nature and the risk of exaggerated correlation magnitude in low power settings; a larger definitive study would be needed in future. However, regarding the significant associations between the CG velocity or hip joint kinetics and RF muscularity, the lower limit of adjusted 90% CI of the correlation coefficients was above 0.1, smallest practically important threshold, with corrected P value < 0.05. In contrast, although statistical significance was not observed, the 90% CI of the correlation coefficients between the CG velocity and volumes of some muscles did not include negative/positive regions (i.e., below/above ± 0.1). Therefore, considering the uncertainty, it is possible that significant relationships could be found in these variables with more data. In any case, the substantial correlations observed are expected to be supporting evidence for practical application in training and coaching situations.
The significant practical implication of the current study is that improvement of the mechanical contribution of RF during swing could be important for sprinting, considering the interrelationships among the RF muscularity, hip flexion kinetics during swing and running speed (Fig. 4). A possible way to accomplish this enhancement is by performing resistance training to induce neuromuscular adaptation of RF. While the optimal exercises for stimulating RF adaptation are not known, because the RF is a biarticular muscle, we suggest knee extension training as well as hip flexion training as a training modality in terms of joint actions during training. A recent study showed that hip flexion training but not knee extension training induced a substantial increase in neural activation of RF during hip flexions, whereas mean the hypertrophic response of RF was similar between the two training modalities (29). In addition, hip flexion training improved sprint ability, even though changes in RF neuromuscular function was not evaluated (30). Therefore, compared with knee extension training, hip flexion training is likely to be more appropriate to incorporate into sprinters’ training activities.
The current study revealed that compared with untrained controls, sprinters have greater hip flexor and extensor thigh muscles, PM, total adductors, and Gr, but both groups have similar monoarticular knee extensor and flexor thigh muscles. There were relationships among the RF volume, hip flexion kinetics, and running speed. The findings suggested that RF plays an important role for fast sprinting through generating hip flexion moment and thus angular impulse during swing.
This study was partly supported by JSPS KAKENHI grant JP15J08355 and JP16H01870. The authors gratefully acknowledge Mr. Shintaro Hanawa, Mr. Hikaru Miura, Ms. Hoshizora Ichinose, Mr. Shun Otsuka and Mr. Kazuma Honda for data acquisition. No conflicts of interest are directly relevant to this article. The present results do not constitute endorsement by the American College of Sports Medicine.
The authors declare that the results of the study are presented clearly, honestly, and without fabrication, falsification, or inappropriate data manipulation.
1. Dorn TW, Schache AG, Pandy MG. Muscular strategy shift in human running: dependence of running speed on hip and ankle muscle performance. J Exp Biol
. 2012;215(Pt 11):1944–56.
2. Schache AG, Blanch PD, Dorn TW, Brown NA, Rosemond D, Pandy MG. Effect of running speed on lower limb joint kinetics. Med Sci Sports Exerc
3. Ikai M, Fukunaga T. Calculation of muscle strength per unit cross-sectional area of human muscle by means of ultrasonic measurement. Int Z Angew Physiol
4. Sugisaki N, Kobayashi K, Tsuchie H, Kanehisa H. Associations between individual lower limb muscle volumes and 100-m sprint time in male sprinters. Int J Sports Physiol Perform
5. Ema R, Wakahara T, Yanaka T, Kanehisa H, Kawakami Y. Unique muscularity in cyclists’ thigh and trunk: a cross-sectional and longitudinal study. Scand J Med Sci Sports
6. Kumagai K, Abe T, Brechue WF, Ryushi T, Takano S, Mizuno M. Sprint performance is related to muscle fascicle length in male 100-m sprinters. J Appl Physiol (1985)
7. Hoshikawa Y, Muramatsu M, Iida T, et al. Influence of the psoas major and thigh muscularity on 100-m times in junior sprinters. Med Sci Sports Exerc
8. Handsfield GG, Knaus KR, Fiorentino NM, Meyer CH, Hart JM, Blemker SS. Adding muscle where you need it: non-uniform hypertrophy patterns in elite sprinters. Scand J Med Sci Sports
9. Fukunaga T, Miyatani M, Tachi M, Kouzaki M, Kawakami Y, Kanehisa H. Muscle volume is a major determinant of joint torque in humans. Acta Physiol Scand
10. O’Brien TD, Reeves ND, Baltzopoulos V, Jones DA, Maganaris CN. Strong relationships exist between muscle volume, joint power and whole-body external mechanical power in adults and children. Exp Physiol
11. Ema R, Wakahara T, Kanehisa H, Kawakami Y. Inferior muscularity of the rectus femoris
to vasti in varsity oarsmen. Int J Sports Med
12. Sakaguchi M, Shimizu N, Yanai T, Stefanyshyn DJ, Kawakami Y. Hip rotation angle is associated with frontal plane knee joint mechanics during running. Gait Posture
13. Delp SL, Maloney W. Effects of hip center location on the moment-generating capacity of the muscles. J Biomech
14. Ehrig RM, Taylor WR, Duda GN, Heller MO. A survey of formal methods for determining the centre of rotation of ball joints. J Biomech
15. Ehrig RM, Taylor WR, Duda GN, Heller MO. A survey of formal methods for determining functional joint axes. J Biomech
16. Söderkvist I, Wedin PA. Determining the movements of the skeleton using well-configured markers. J Biomech
17. Sakaguchi M, Ogawa H, Shimizu N, Kanehisa H, Yanai T, Kawakami Y. Gender differences in hip and ankle joint kinematics on knee abduction during running. Eur J Sport Sci
. 2014;14(1 Suppl):S302–9.
18. Vaughan CL, Davis BL, O’Connor JC. Dynamics of Human Gait
. Champaign (IL): Human Kinetics; 1992. pp. 83–106.
19. Ae M, Tang H, Yokoi T. Estimation of inertia properties of the body segments in Japanese athletes. Biomechanism
20. FDR online calculator [Internet]. SDM Project web; [cited 2018 Apr 20]. Available from: https://www.sdmproject.com/utilities/?show=FDR
21. Hopkins WG, Marshall SW, Batterham AM, Hanin J. Progressive statistics for studies in sports medicine and exercise science. Med Sci Sports Exerc
22. Altman DG, Bland JM. How to obtain the confidence interval from a P value. BMJ
23. Blazevich AJ, Coleman DR, Horne S, Cannavan D. Anatomical predictors of maximum isometric and concentric knee extensor moment. Eur J Appl Physiol
24. Zheng N, Fleisig GS, Escamilla RF, Barrentine SW. An analytical model of the knee for estimation of internal forces during exercise. J Biomech
25. Nummela A, Rusko H, Mero A. EMG activities and ground reaction forces during fatigued and nonfatigued sprinting. Med Sci Sports Exerc
26. Miyake Y, Suga T, Otsuka M, et al. The knee extensor moment arm is associated with performance in male sprinters. Eur J Appl Physiol
27. Schache AG, Dorn TW, Blanch PD, Brown NA, Pandy MG. Mechanics of the human hamstring muscles during sprinting. Med Sci Sports Exerc
28. Delecluse C. Influence of strength training on sprint running performance. Current findings and implications for training. Sports Med
29. Ema R, Saito I, Akagi R. Neuromuscular adaptations induced by adjacent joint training. Scand J Med Sci Sports
30. Deane RS, Chow JW, Tillman MD, Fournier KA. Effects of hip flexor training on sprint, shuttle run, and vertical jump performance. J Strength Cond Res