The hamstrings consist of the biceps femoris long heads and biceps femoris short heads (BFlh and BFsh, respectively), semitendinosus (ST) and semimembranosus (SM), working as hip extensors and/or knee flexors. It is known that there are morphological and functional differences among the individual hamstring muscles. For example, BFlh and SM possess a large physiological cross-sectional area, whereas ST has long muscle fibers (15). Thus, BFlh and SM have a potential to generate large force, whereas ST has a potential to generate high contraction velocity (11). The individual hamstring muscles were demonstrated to be important for sports, such as sprint running and football (1,4). For instance, muscle volume of ST in sprinters was 54% greater than in nonsprinters, whereas these differences were 26 and 20% in BFlh and SM, respectively (6). In addition, ST showed higher activation than that of BFlh as increasing running speed (8). These results suggest that ST may play a more significant role in sprint running than BFlh. Meanwhile, muscle strain injuries often occurred during the maximal sprinting and kicking in the hamstrings, especially in BFlh (5,14,15,18). The injury of BFlh was shown to be prevented by resistance training for strengthening the hamstring muscles as a whole (3). It may be possible that training program aimed at selectively strengthening BFlh is more effective for preventing the injury of the muscle compared with the training for strengthening the whole hamstring muscles.
Muscle activity level during training exercises has been evaluated by surface electromyogram (EMG). The EMG activity of hamstrings was shown to be influenced by the hip joint position (21). The activity level of BFlh during maximum isometric knee flexion was lower with the hip flexed at 135° than with 0 or 45° of hip flexion (12). It should be noticed here that the hip joint is a ball-and-socket joint, which allows three-dimensional motions, such as internal/external rotation and adduction/abduction in addition to flexion/extension. A previous study examined the activation ratio of the medial-lateral hamstrings during single-leg deadlift in an internally and an externally rotated positions of the hip joint (13). The results showed that the medial-lateral hamstrings activation ratio in the internally rotated position (approximately 3.2) was higher than that in the externally rotated position (approximately 0.8) over the full range of the motion. This finding suggests that the internal/external rotation of the hip joint affects the muscle activity level of the lateral and medial hamstring muscles. Within the medial hamstrings, the activity level was demonstrated to be different between ST and SM during deadlift (16). However, BFlh crosses both the hip and knee joints, whereas BFsh does not cross the hip joint. Therefore, it is possible that the influence of the hip joint position on the muscle activity level differs among the individual hamstrings. However, Lynn and Costigan (13) did not separate the lateral hamstrings into BFlh and BFsh, or the medial hamstrings into ST and SM. Meanwhile, it was shown that activity level during stiff-leg deadlift was nonuniform along the length of BFlh and ST (7). However, it remains unclear whether the muscle activity level is different between the proximal and distal regions within the individual hamstrings during stiff-leg deadlift in various hip joint positions.
The purpose of this study was to investigate the effect of hip joint position on the level of activity of the individual hamstring muscles during stiff-leg deadlift.
Experimental Approach to the Problem
To determine the effect of hip joint position on the individual hamstring muscles, stiff-leg deadlift was chosen. This is because stiff-leg deadlift has been widely prescribed by strength and conditioning professionals to strengthen the hamstring muscles, and the hip joint positions (internal/external rotation and adduction/abduction) can be easily manipulated by adjusting foot placement and angle during the exercise. Subjects performed stiff-leg deadlift in 6 hip joint positions in a random order after maximal voluntary contraction (MVC) of isometric knee flexion. During the deadlift, EMG signals were recorded from the proximal and distal regions of BFlh, ST, and SM for the analyses. Root mean square values of EMG data (RMS-EMG) were calculated for each of the concentric and eccentric phases of the deadlift and were normalized by RMS-EMG during MVC as % MVC.
Fourteen male collegiate sprinters (age: 19.6 ± 1.0 years, body mass: 64.5 ± 4.1 kg, height: 175.4 ± 5.6 cm, personal best time of a 100-m race: 11.03 ± 0.25 seconds, mean ± SD) participated in this study. The age range of these subjects was from 18 to 21 years. They had a minimum of 3 years of experience of sprint running and no injury in the hamstrings at the time of experiment. All subjects were fully informed of the purpose and procedure of this study and gave written informed consent before participation in the study. All protocols of this study were approved by the Doshisha University Research Ethics Review Committee regarding Human Subject Research [number: 18002].
Bipolar electrodes (DL-141; S&ME, Tokyo, Japan, interelectrode distance: 12 mm) were placed over 40 and 60% of the thigh length (the distance between the greater trochanter [0%] and the popliteal crease [100%]) for BFlh; 30 and 50% for ST and 50 and 70% for SM to evaluate the activity in the proximal and distal regions of each muscle. The electrode placements were determined on the basis of innervation patterns of the individual hamstring muscles (19,20). The border of each muscle was carefully identified with an ultrasonic apparatus (PROSOUND α7; Hitachi Aloka Medical, Tokyo, Japan). After careful preparation of the skin by shaving hair, abrading, and cleaning with alcohol, the electrodes (DL-141, S&ME, interelectrode distance: 12 mm) were attached on the right thigh and fixed with surgical tape. A ground electrode was attached on the right medial malleolus.
Subjects lay prone on a bed of a dynamometer (BIODEX SYSTEM4; Biodex Medical Systems, Upton, NY) with the right knee joint at 45° (full extension = 0°). The pelvis and right ankle were fixed with belts. They held the bed with their arms. After a warm-up session including 3 submaximal contractions, they performed 2 MVCs of isometric knee flexions for 2 seconds.
After MVC measurement, the subjects performed stiff-leg deadlift in the following 6 hip joint positions (Figure 1): (a): internally rotated by 20° (IN20), (b) neutral (NT), (c) externally rotated by 20° (EX20) and (d) by 40° (EX40), (e) adducted (ADD), and (f) abducted (ABD) positions. In IN20, EX20, and EX40, the distance between the middle of the heels was set at 20% of their height. In ADD, NT, and ABD, the feet were placed parallel with a distance between the middle of the heels set at 5, 20, and 40% of their height, respectively. They put their feet on white tapes preset to adjust the hip joint positions. The deadlift was started from the upright position with a barbell positioned in front of the thighs. They slowly lowered the barbell to their tibial tuberosities with the knee joints extended and then lifted the barbell to the start position. As a familiarization session, subjects performed a few repetitions of deadlift in each of the hip joint positions with a load of 60% of their body mass with particular attention to the range of motion and body position during the deadlift. After this session, they performed 2 sets of 2 repetitions of deadlift with the same load as the familiarization session. Each of the concentric and eccentric phases of deadlift was completed in 2 seconds with a metronome set to 60 b·min−1. The order of the hip joint positions was randomized across the subjects. Sufficient rest was allowed between the different hip joint positions to avoid an influence of muscular fatigue on EMG data. When they were unable to perform the deadlift at the prescribed pace or with proper technique, an additional set was requested to perform after 3-minute rest.
Raw EMG signals were preamplified and filtered at 5–500 Hz with the electrodes and recorded on a computer at a sampling frequency of 1000 Hz via an A/D converter (PowerLab 16SP; AD Instruments, Sydney, Australia). The EMG signals were high-pass filtered at a cutoff frequency of 5 Hz with computer software (LabChart ver.8; AD Instruments). The knee flexion torque signal was sampled with the A/D converter (PowerLab 16SP; AD Instruments) at 1000 Hz and transferred to the computer. The RMS-EMG during MVC was determined for 0.5 seconds around the peak torque. The higher RMS-EMG value of the 2 MVC trials was used for the normalization of EMG. The RMS-EMG during the stiff-leg deadlift was calculated from the EMG data of each of concentric (from the start to the end of upward movement) and eccentric phases (from the start to the end of the downward movement). To determine the concentric and eccentric phases, the motion of the subject during the deadlift was recorded with a digital video camera (EX-100; Casio, Tokyo, Japan) at 30 Hz. The video was synchronized with the EMG data by using an electrical signal from a synchronizer (PH-1250A-6SW; DKH, Tokyo, Japan). The RMS-EMG values during deadlift excluding the largest and the smallest values among the 4 data sets (2 sets × 2 repetitions) of each hip joint position were extracted, and the mean of the 2 values was normalized by RMS-EMG during MVC as %MVC. The intraclass correlation coefficients (ICC) of the individual hamstring muscles were analyzed by the 2 extracted RMS-EMG values in each hip joint position. Based on the previous study (10), ICC was evaluated as “almost perfect” (ICC > 0.81), “substantial” (ICC = 0.61–0.80), and “moderate” (ICC = 0.41–0.60). All EMG data of the individual hamstring muscles were evaluated as almost perfect (ICC [1, 2] = 0.89–0.99) except for the proximal region of SM during eccentric phase in NT (ICC [1, 2] = 0.67).
All statistical analyses were performed using a statistical software package (IBM SPSS Statistics, ver. 25.0, IBM). A three-way analysis of variance (ANOVA) with repeated measures was used to examine the effects of muscle (BFlh, ST, and SM), region (proximal and distal) and hip joint position on RMS-EMG of the concentric and eccentric phase, separately. The ANOVA was separately performed for the internal/external rotation (IN20, NT, EX20, and EX40) and adduction/abduction (ADD, NT, and ABD) of the hip joint. If a significant interaction was found, two- and one-way ANOVAs with Bonferroni multiple comparison tests were performed to test for any differences in RMS-EMG among the hip joint positions. The statistical power, effect size (partial η2 and Cohen's d), and confidence interval (CI) were calculated for these tests. The statistical significance level was set at p ≤ 0.05.
Three-way ANOVA for the internal/external rotation showed no significant interaction of muscle × region × hip joint position for RMS-EMG in the concentric phase (p = 0.165, partial η2 = 0.108, statistical power = 0.575). Because no significant interactions of muscle × region (p = 0.689, partial η2 = 0.028, statistical power = 0.104) or region × hip joint position (p = 0.996, partial η2 = 0.064, statistical power = 0.053) were found, RMS-EMG values in the proximal and distal regions were pooled for each muscle. A significant interaction of muscle × hip joint position was found for the pooled RMS-EMG in the concentric phase (p < 0.001, partial η2 = 0.563, statistical power = 0.999, Figure 2A). The post hoc test revealed that RMS-EMG of BFlh was significantly higher in EX20 than that in NT (p = 0.008, 95% CI: 1.4–11.5, d = 0.446, difference = 6.3%). The RMS-EMG of BFlh in EX40 was significantly higher than that in NT (p = 0.001, 95% CI: 4.1–14.7, d = 0.627, difference = 9.4%) and IN20 (p = 0.021, 95% CI: 1.0–13.9, d = 0.497, difference = 7.4%). The RMS-EMG of ST was not significantly different among IN20, NT, EX20, and EX40 (p = 0.051, partial η2 = 0.179, statistical power = 0.634). The RMS-EMG of SM was significantly higher in IN20 than that in EX40 (p = 0.004, 95% CI: 1.2–7.3, d = 0.407, difference = 4.3%).
Three-way ANOVA for the adduction/abduction showed no significant interaction of muscle × region × hip joint position for RMS-EMG in the concentric phase (p = 0.770, partial η2 = 0.034, statistical power = 0.148). There were no significant interactions of muscle × region (p = 0.911, partial η2 = 0.007, statistical power = 0.063) or region × hip joint position (p = 0.749, partial η2 = 0.022, statistical power = 0.091). Hence, RMS-EMG values in the proximal and distal regions were pooled for each muscle. Although no significant interaction of muscle × hip joint position was found for the pooled RMS-EMG in the concentric phase (p = 0.792, partial η2 = 0.031, statistical power = 0.140), there was a significant main effect of hip joint position (p = 0.001, partial η2 = 0.434, statistical power = 0.972, Figure 2B). The post hoc analysis showed that the RMS-EMG in ABD was significantly higher than that in NT for BFlh (p = 0.015, 95% CI: 0.9–8.4, d = 0.320, difference = 4.6%), ST (p = 0.047, 95% CI: 0.0–7.5, d = 0.444, difference = 3.8%), and SM (p = 0.005, 95% CI: 1.2–6.6, d = 0.367, difference = 3.9%).
Three-way ANOVA for the internal/external rotation showed no significant interaction of muscle × region × hip joint position for RMS-EMG in the eccentric phase (p = 0.806, partial η2 = 0.037, statistical power = 0.192). Because there were no significant interactions of muscle × region (p = 0.604, partial η2 = 0.038, statistical power = 0.125) or region × hip joint position (p = 0.808, partial η2 = 0.024, statistical power = 0.107), RMS-EMG values in the proximal and distal regions were pooled for each muscle. There was a significant interaction of muscle × hip joint position for the pooled RMS-EMG in the eccentric phase (p < 0.001, partial η2 = 0.444, statistical power = 0.999, Figure 3A). The post hoc analysis demonstrated that RMS-EMG of BFlh was significantly higher in EX40 than that in IN20 (p = 0.038, 95% CI: 0.3–13.4, d = 0.684, difference = 6.8%). The RMS-EMG of ST was not significantly different among the hip joint positions (p = 0.660, partial η2 = 0.048, statistical power = 0.176). The RMS-EMG of SM was significantly higher in IN20 than in EX20 (p = 0.019, 95% CI: 0.5–6.3, d = 0.440, difference = 3.4%) and EX40 (p = 0.023, 95% CI: 0.5–7.8, d = 0.548, difference = 4.1%).
Three-way ANOVA for the adduction/abduction showed no significant interaction of muscle × region × hip joint position for RMS-EMG in the eccentric phase (p = 0.983, partial η2 = 0.007, statistical power = 0.068). There were no significant interactions of muscle × region (p = 0.804, partial η2 = 0.017, statistical power = 0.081) or region × hip joint position (p = 0.502, partial η2 = 0.052, statistical power = 0.157). Thus, RMS-EMG values in the proximal and distal regions were pooled for each muscle. There was no significant interaction of muscle × hip joint position (p = 0.650, partial η2 = 0.046, statistical power = 0.191) or significant main effect of hip joint position (p = 0.672, partial η2 = 0.030, statistical power = 0.108, Figure 3B) for the pooled RMS-EMG in the eccentric phase.
The results of this study showed that the level of muscle activity of BFlh in EX20 and EX40 was higher than that in NT during the concentric phase of stiff-leg deadlift, whereas that of SM was higher in IN20 than in EX40 during the concentric and eccentric phases. In addition, the activity levels of the individual hamstring muscles in ABD were higher than that in NT during the concentric phase. However, no significant differences were found in the activity level of ST among the hip joint positions during the concentric or eccentric phases except for the difference between NT and ABD during concentric phase. It has been reported that the internal/external rotation of the hip joint affects the muscle activity level of the hamstrings during single-leg deadlift (13). However, the previous study (13) did not separate the lateral or medial hamstrings into the individual hamstrings. This study first revealed that the external/internal rotation and abduction of the hip joint affect the muscle activity levels of the individual hamstring muscles during stiff-leg deadlift.
The muscle activity level of BFlh was higher in EX20 and EX40 than in NT during the concentric phase of stiff-leg deadlift. On the contrary, the activity level of SM was higher in IN20 than in EX40 during the concentric and eccentric phases. As the stiff-leg deadlift is a hip-dominant exercise, the hip extensors including the individual hamstring muscles need to generate hip extension torque. The magnitude of muscle torque is the product of the muscle force and its moment arm. It is possible that the moment arm of BFlh and SM at the hip joint in the sagittal plane increased when these muscles were located at the most posterior position in the externally and internally rotated position, respectively (Figure 4A). If so, the long moment arm greatly contributes to generating muscle torque, and thus high activity level of the muscle with the long moment arm could be efficient for exerting the required torque. Taken together, the possible difference in the moment arm at the hip joint might be related to the difference in the activity level of BFlh and SM in the internally/externally rotated positions, respectively.
The activity levels of the individual hamstring muscles were higher in ABD than in NT during the concentric phase of the deadlift. To lift the barbell, the hip extensors including the hamstring muscles in ABD must generate the hip extension torque equivalent to that in NT. However, the contribution of the component of the force generated by the individual hamstrings to the production of hip extension torque was supposed to be smaller in ABD than in NT (Figure 4B). Thus, higher activation of the hamstring muscles may be required in ABD as compared to NT.
There was no difference in the muscle activity level of ST among the hip joint positions during concentric or eccentric phases of the deadlift except for the difference between NT and ABD during concentric phase. A previous study reported that muscle activity level of ST during stiff-leg deadlift was lower than that of BFlh and SM (16). Similarly, the average EMG activity across the hip joint positions in this study was lower in ST (19 and 8%) than in BFlh (32 and 15%) and in SM (33 and 20%) during the concentric and eccentric phases, respectively. These results imply that a contribution of ST to the production of hip extension torque would be small during stiff-leg deadlift among the bi-articular hamstring muscles. This might be related to the lack of difference in the activity level of ST among the hip joint positions. The current result suggests that the internal/external rotation and adduction of the hip joint do not have a great influence on the activity level of ST during stiff-leg deadlift.
The activity level was not different between the proximal and distal regions within the individual hamstring muscles during stiff-leg deadlift. This result was inconsistent with the previous finding that the distal region of BFlh showed a higher activity level than the proximal region during stiff-leg deadlift (7). There are some differences between the present and previous studies. First, the neuromuscular activity of a region was determined from the data recorded from multiple electrodes using high-density electromyography in the previous study (7), while using a single electrode in this study. Second, range of motion during the deadlift differed between the 2 studies. In the previous study, the subjects lowered the bar from their thigh to the floor (7). However, the subjects of this study lowered the bar from their thigh to tibial tuberosities. Finally, the load during the deadlift might be heavier in the previous study (80% of one repetition maximum) than in this study (60% of body mass). Although these differences may be associated with the discrepancy between the 2 studies, it is unclear how these factors affect the activity level of the proximal and distal regions within the individual hamstring muscles. Further research is needed to scrutinize the regional difference in the activity within each of the hamstring muscles.
This study has several limitations. The surface EMG is known to be susceptible to cross-talk from adjacent muscles. To reduce the effect of cross-talk, we carefully identified the belly of each muscle using ultrasonography. Moreover, our subjects were male sprinters, who have larger hamstring muscles than other populations (9). Therefore, we assume that the effect of cross-talk did not have a substantial influence on the main finding of this study. Further research is warranted to clarify the applicability of the present findings to other populations.
Stiff-leg deadlift is commonly used for preventing strain injury of the hamstring muscles and enhancing the strength of hip extensor muscles in rehabilitation and training fields. When performing the stiff-leg deadlift, athletes and their coaches may not pay attention to the hip joint position. Meanwhile, this study revealed the effects of hip joint position on the muscle activity levels of the individual hamstrings during the stiff-leg deadlift. The activity level of BFlh was higher in the externally rotated and abducted positions than in the neutral position. It is known that BFlh is the most frequently injured muscle among the hamstrings, probably due to the weakness of the muscle strength (17). Thus, the stiff-leg deadlift in the externally rotated and abducted positions may have a great potential to prevent the injury of BFlh. However, SM showed a high level of activity in the internally rotated or abducted positions. The strain injury of SM frequently occurred during stretching exercises executed at an extreme joint position (2). It is possible that the injury of SM is prevented by performing the stiff-leg deadlift in the internally rotated and abducted positions. Taken together, it is recommended that athletes and their coaches carefully select the hip joint position during the stiff-leg deadlift considering the different activity levels of the individual hamstring muscles.
The authors show their greatest appreciation to the participations in this study and all members of their laboratory. The results of this study do not constitute endorsement of any product by the authors or the National Strength and Conditioning Association.
The authors have no conflict of interests to disclose.
1. 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 18: 40–48, 2008.
2. Askling CM, Tengvar M, Saartok T, Thorstensson A. Proximal hamstring strains of stretching type in different sports: Injury situations, clinical and magnetic resonance imaging characteristics, and return to sport. Am J Sports Med 36: 1799–1804, 2008.
3. Askling CM, Karlsson J, Thorstensson A. Hamstring injury occurrence in elite soccer players after preseason strength training with eccentric overload. Scand J Med Sci Sports 13: 244–250, 2003.
4. Askling CM, Tengvar M, Saartok T, Thorstensson A. Acute first-time hamstring strains during high-speed running: A longitudinal study including clinical and magnetic resonance imaging findings. Am J Sports Med 35: 197–206, 2007.
5. Garrett JW, Rich FR, Nikolaou PK. Computed tomography of hamstring muscle strains. Med Sci Sports Exerc 21: 506–514, 1989.
6. Handsfield GG, Knaus KR, Fiorentino NM, et al. Adding muscle where you need it: Non-uniform hypertrophy patterns in elite sprinters. Scand J Med Sci Sports 27: 1050–1060, 2017.
7. Hegyi A, Peter A, Finni T, Cronin NJ. Region-dependent hamstrings activity in Nordic hamstring exercise and stiff-leg deadlift defined with high-density electromyography. Scand J Med Sci Sports 28: 992–1000, 2018.
8. Higashihara A, Ono T, Kubota J, Okuwaki T, Fukubayashi T. Functional differences in the activity of the hamstring muscles with increasing running speed. J Sports Sci 28: 1085–1092, 2010.
9. Hoshikawa Y, Muramatsu M, Iida T, et al. Event-related differences in the cross-sectional areas and torque generation capabilities of quadriceps femoris and hamstrings in male high school athletes. J Physiol Anthropol 29: 13–21, 2010.
10. Landis JR, Koch GG. The measurement of observer agreement for categorical data. Biometrics 33: 159–174, 1977.
11. Lieber RL, Fridén J. Functional and clinical significance of skeletal muscle architecture. Muscle Nerve 23: 1647–1666, 2000.
12. Lunnen JD, Yack J, LeVeau BF. Relationship between muscle length, muscle activity, and torque of the hamstring muscles. Phys Ther 61: 190–195, 1981.
13. Lynn SK, Costigan PA. Changes in the medial-lateral hamstring activation ratio with foot rotation during lower limb exercise. J Electromyogr Kinesiol 19: 197–205, 2009.
14. Lysholm J, Wiklander J. Injuries in runners. Am J Sports Med 15: 168–171, 1987.
15. Makihara Y, Nishino A, Fukubayashi T, Kanamori A. Decrease of knee flexion torque in patients with ACL reconstruction: Combined analysis of the architecture and function of the knee flexor muscles. Knee Surg Sports Traumatol Arthrosc 14: 310–317, 2006.
16. Ono T, Higashihara A, Fukubayashi T. Hamstring functions during hip-extension exercise assessed with electromyography and magnetic resonance imaging. Res Sports Med 19: 42–52, 2011.
17. Orchard J, Marsden J, Lord S, Garlick D. Preseason hamstring muscle weakness associated with hamstring muscle injury in Australian footballers. Am J Sports Med 25: 81–85, 1997.
18. Pomeranz SJ, Heidt RS Jr. MR imaging in the prognostication of hamstring injury. Work Prog Radiol 189: 897–900, 1993.
19. Rha DW, Yi KH, Park ES, Park C, Kim HJ. Intramuscular nerve distribution of the hamstring muscles: Application to treating spasticity. Clin Anat 29: 746–751, 2016.
20. Woodley SJ, Mercer SR. Hamstring muscles: Architecture and innervation. Cells Tissues Organs 179: 125–141, 2005.
21. Worrell TW, Karst G, Adamczyk D, et al. Influence of joint position on electromyographic and torque generation during maximal voluntary isometric contractions of the hamstrings and gluteus maximus muscles. J Orthopaedic Sports Phys Ther 31: 730–740, 2001.