Following lower extremity injury, rehabilitation is critical in returning the patient to full function. Special attention has been given to rehabilitation of knee joint injuries, particularly injuries to the anterior cruciate ligament (ACL) (9,15,22). Knee rehabilitation has often focused on open kinetic chain exercises such as straight leg raises and knee flexion/extension exercises in which the distal segment is “free” (13). More recent protocols, however, emphasize the use of closed kinetic chain exercises, such as the squat, the leg press, and the lateral step-up, in which the distal segment is “fixed”(2,9,11,15,21,22).
Both open and closed kinetic chain exercises are used in knee rehabilitation. Several limitations, however, exist with open kinetic chain exercises, including increased patellofemoral compression(18), increased tibial shear forces(6,11,12,22), increased ACL strain(9,17,22) and nonfunctional muscle recruitment patterns (15). While closed kinetic chain exercises may result in decreased isolation of the muscle of interest and possibly muscle substitution, many studies suggest that closed kinetic chain exercises offer distinct advantages over open kinetic chain exercises by providing co-contraction of the quadriceps and hamstring muscle groups,(11-13) thus decreasing shear forces(6,11,13) and providing functional muscle recruitment patterns through which integrated multi-joint movement is produced(4). However, few studies exist that quantify hamstring and quadriceps muscle activity during a squat activity(5,12). To the contrary, several studies report low hamstring activity during an unloaded squat (5) and lateral step-up, (2,3,19) which questions the ability of the hamstrings to counteract the anterior tibial forces imparted by the quadriceps at the knee (13).
Several studies have reported electromyographic (EMG) analyses of lower extremity muscle activity during a variety of closed kinetic chain activities(2-5,9,12,21). Of these activities, the unloaded squat exercise has been considered a safe and effective closed kinetic chain activity to be used in the early stages of ACL rehabilitation, in large part because of the purported stabilizing effects of hamstring-quadriceps co-contraction(5,11,12,22). Ohkoshi et al.(12) reported the effect of hip angle on quadriceps and hamstring recruitment and found increasing hamstring activity with corresponding increases in hip flexion angle. Gryzlo et al.(5) reported the effect of knee angle on quadriceps and hamstring recruitment, noting significant quadriceps activity but only minimal hamstring activity throughout all phases of the squat.
Studies examining EMG activity during squatting movements tend to isolate attention specifically on the quadriceps and hamstrings, but the response of other muscles may also be relevant. For example, the gluteus maximus has been shown to be active with increasing amounts of hip flexion(1). In addition, the gastrocnemius has potential function at the knee. As a result of its origin on the posterior femoral condyles, it may, quite possibly, limit posterior translation of the tibia relative to the femur. Therefore, an analysis of gluteus maximus and gastrocnemius EMG activity during an unloaded squat is warranted to provide insight as to any implied functional consequences. Therefore, the purposes of this study were: 1) to determine the muscle recruitment patterns of the gluteus maximus, hamstrings, quadriceps, and gastrocnemius during an unloaded squat exercise via EMG; and 2) to describe the amount of hamstring-quadriceps co-contraction during an unloaded squat exercise.
Subjects. Forty-one healthy subjects (20 males, 21 females; mean age, 25.5 yr, range, 21-39 yr; mean height, 174.4 cm, range, 157.5-195.6 cm; mean weight, 72.7 kg, range, 51.4-115.9 kg) without history of knee injury, knee pain, or surgery participated in this study. Each test leg was randomly chosen (18 right legs, 23 left legs). Prior to participation, subjects read and signed an informed consent form that was approved by a University Institutional Review Board.
Skin preparation. Each electrode site was shaved, abraded, and cleaned with alcohol to facilitate electrode adherence and conduction of EMG signals. Bipolar pre-amplified surface electrodes (Therapeutics Unlimited, Iowa City, IO) were placed over the bellies of the gluteus maximus, hamstring, vastus medialis oblique, rectus femoris, vastus lateralis, and gastrocnemius muscles.
Electrode placement was determined by placing subjects in the appropriate test positions and identifying the muscle bellies of interest via isometric contraction. For this study we selected the common superior hamstring muscle belly as the placement site for the hamstring electrode. Several studies have evaluated the EMG muscle activity of the medial and lateral hamstrings separately but have not shown significant differences between the two groups (2,3,5). Therefore, we did not feel that it was necessary to evaluate the EMG muscle activity of the medial and lateral hamstring groups separately.
All electrodes were positioned parallel to the muscle fibers in such a way that this alignment was maintained throughout the entire arc of movement. A common ground electrode was placed anteriorly on the proximal tibia. Elastic wraps were applied to the test extremity to maintain electrode placement. All test sites were identified and prepared by the same researcher.
Instrumentation. Raw EMG signals were monitored with pre-amplified surface electrodes. Amplified signals were then processed by the main amplifier (Therapeutics Unlimited, Iowa City, IO) to produce root mean square (RMS) signals with a time constant of 11.75 ms. During testing, RMS EMG signals were monitored by an oscilloscope. The main amplifier was linked to a Macintosh computer (Apple Computer Inc., Cupertino, CA) running Acknowledge Software (Biopac Systems Inc., Goleta, CA) to sample digitally each muscle's RMS EMG signals at a rate of 50 samples/s/channel.
Pre-trial open kinetic chain testing procedures. Pain-free maximal voluntary isometric contractions (MVIC) were recorded for each muscle tested and were used as references for comparison of muscle activity during the squat exercise. Subjects performed three warm-up isometric contractions at 50, 75, and 100% of perceived maximum. Each contraction was held for 4 s with a 30-s rest period between practice repetitions. Then, three 4-s MVICs were performed against a fixed resistance for each muscle group in the following positions: gastrocnemius- prone with the hip and knee in 0° of flexion and the ankle in 0° of plantar flexion; vastus medialis oblique, rectus femoris, and vastus lateralis- -seated with the hip at 90° and knee at 60° of flexion; hamstrings-prone with the hip and knee in 0° and 30° of flexion, respectively; and gluteus maximus - prone with the hip in 20° of extension and the knee in >90° of flexion. The largest of the three MVICs was used for normalization. Subjects received a 5-min rest period between open and closed kinetic chain testing sessions.
Closed kinetic chain testing procedures. Prior to the squat trials, electrogoniometers (ELGON) were applied to the knee and hip joints of the test extremity. The knee ELGON axis of rotation was centered directly over the lateral joint line. The proximal arm was placed on the lateral thigh and was aligned with the lateral midline of the femur, using the greater trochanter for reference. The distal arm was placed on the lateral aspect of the lower leg and was aligned with the lateral aspect of the fibula, using the lateral malleolus for reference. The knee ELGON (femur relative to tibia) was used to define the arcs of motion used for data analysis. The hip ELGON was applied to the hip such that the axis of rotation was centered directly over the greater trochanter. The proximal arm was placed on the pelvis and trunk and was aligned with the lateral midline of the pelvis. The distal arm was placed on the lateral thigh and was aligned with the lateral midline of the femur, using the greater trochanter for reference. The hip ELGON was used to monitor the relationship between the femur and pelvis during squat testing. Elastic wraps were applied to the goniometers to maintain proper positioning during dynamic activities.
Subjects were then placed in the squat station. Test position was standardized for subjects based on the angle of ankle dorsiflexion achieved during the hold position (i.e., transition between eccentric and concentric phases). Dorsiflexion angle was standardized for all subjects to 20° as measured in the closed kinetic chain. From the standing position with feet approximately shoulder-width apart, subjects were asked to squat down until the anterior aspects of both tibias contacted a horizontal bar that controlled for the desired angle of ankle dorsiflexion. Likewise, buttocks contact with a bench regulated the maximum degree of hip and knee flexion, a position that placed the femurs in a plane parallel with the floor (Fig. 1). Upon contact with the bars, each subject was asked to hold the position and then return to the starting position. During warm-up and test repetitions, subjects were instructed not to “rest” or “unload” on the bench and horizontal bar. In addition, the hold position was visually monitored by one of the researchers to ensure that unloading did not occur.
Subjects were permitted to perform as many warm-up repetitions as needed until they demonstrated proper squat technique in cadence with a metronome that was set at 50 beats·min-1. The pace of the metronome was set such that one complete squat was performed in 3.6 s, with each phase of the squat (i.e., eccentric, hold, and concentric) lasting 1.2 s. The mean starting/ending position (± SD) for all subjects was 5 ± 7.5° (range, 8-0-22°) at the knee and 5 ± 6.4° (range, 12-0-16°) at the hip. The mean hold position (± SD) was 94 ± 4° (range, 85-100°) at the knee and 89 ± 12° (range, 70-126°) at the hip. Subjects performed three test trials, each consisting of four consecutive squat repetitions, during which time EMG activity of the gluteus maximus, hamstrings, vastus medialis oblique, rectus femoris, vastus lateralis, and gastrocnemius was recorded.
Data processing procedures. Specific arcs of the squat were manually identified for each subject by the ELGON range-of-motion display on the computer screen (Fig. 2). Arc values were derived by using the mean RMS signal value that was elicited during each arc of motion throughout each complete squatting motion. MVIC values were derived by using the mean RMS signal value that was elicited during each 4-s MVIC. Percent MVIC values were derived by scaling the mean arc values with the corresponding mean MVIC values. Figure 2 represents the Acknowledge Software display of EMG data and knee and hip range of motion, illustrating the temporal coordination of muscle activity and knee and hip joint kinematics.
Statistical analyses. Intraclass correlation coefficients (ICC 2,1) were used to determine reliability during MVIC and squat testing(19). Based on these results, the largest EMG values recorded during MVIC testing were used to normalize the values obtained in repetition two of squat trial two. The resultant values were expressed as percent MVIC.
Percent MVIC values were used to compare muscle activity during arcs(0-30°, 30-60°, 60-90°, hold, 90-60°, 60-30°, 30-0°) using a two-way repeated measures ANOVA (6 muscles × 7 arcs). The probability level was set at P ≤ 0.05. Tukey post-hoc comparisons were conducted for the simple main effects(7). Statview II (Abacus Concepts, Berkeley, CA) was used for reliability data analysis, and SuperAnova was used for the two-way ANOVA.
Intraclass correlation coefficient (ICC) values for MVIC testing ranged from 0.89-0.98. Between trial comparison of the squat produced ICC values ranging from 0.88-0.97. Furthermore, comparison of repetitions within Trial 2 revealed ICC values ranging from 0.75-0.95 (Table 1).
Two-way repeated measures ANOVA (6 muscles × 7 arcs) revealed significant main effects for muscles and arcs. In addition, a significant interaction was found between muscles and arcs. The significant interaction finding precludes an evaluation of muscle or arc main effects.Figure 3 illustrates the interaction of the muscle activation response across arcs. Group means, SDs, and ranges for% MVIC values for each muscle/arc combination are presented for reference inTable 2A. Table 2B represents the results of Tukey post-hoc analyses that were used to quantify and interpret the simple main effects of the muscles.
Palmitier et al. (13) theorized that during an unloaded squat the hamstrings stabilize the knee joint by counter-acting the anterior shear forces imparted by the quadriceps. They did not, however, substantiate this theory with experimental data. Ohkoshi et al.(12) conducted a study reporting hamstring and quadriceps EMG activity at various angles of knee and hip flexion during an isometric unloaded squat activity. They reported simultaneous contraction of the quadriceps and hamstrings at all angles with hamstring activity increasing as trunk flexion angle increased. Also, they reported increasing posterior shear forces on the tibia with increasing trunk flexion angles, which they attributed to the increased hamstring activity. Ohkoshi et al.(12) did not, however, provide normalized hamstring or quadriceps muscle activity (i.e., percent MVIC). Therefore, we were unable to compare hamstring and quadriceps activity with respect to their capacities.
In our study, post-hoc analyses of simple main effects revealed that during an unloaded squat the vastus medialis oblique and vastus lateralis were significantly more active than the hamstrings across all arcs(Table 2B), with the greatest activity for each muscle occurring during the 90-60° arc (Table 2B, Fig. 3). During this arc, the vastus medialis oblique and vastus lateralis demonstrated EMG activity of ≈68 and 63% of MVIC, respectively. The hamstrings, however, displayed EMG activity of only ≈12% of MVIC. Gryzlo et al. (5) reported similar hamstring and quadriceps EMG activity trends during an unloaded squat. In addition, the lateral step-up exercise has been shown to involve quadriceps and hamstring recruitment patterns similar to that of an isometric squat exercise(14). Several studies(2,3,21) have reported comparably meager hamstring recruitment levels during a lateral step-up exercise with and without additional weight load applied.
The rectus femoris muscle demonstrated somewhat different EMG activity than that of the vastus medialis oblique and vastus lateralis with its greatest activity (≈48% of MVIC) occurring during the 60-90° and 90° arcs. Furthermore, from the 90° arc (Hold Phase) to the 90-60° arc, the rectus femoris demonstrated a sharp decrease in EMG activity while the vastus medialis oblique and vastus lateralis demonstrated sharp increases in EMG activity (Fig. 3). We speculate that this different EMG pattern represents the biarticular nature of the rectus femoris which also functions at the hip joint (i.e., the hip is extending from the 90° arc to the 30-0° arc). This requires the rectus femoris to decrease its activity since the rectus femoris is a hip flexor, not a hip extensor. In addition, from the 90-60° arc to the 30-0° arc, EMG activity of all quadriceps muscles decreased, which likely reflects the decreased demand for knee extensor force.
The hamstring and gluteus maximus muscles demonstrated similar EMG activity patterns across all arcs (Fig. 3). At no time during the squat, however, did the hamstrings and gluteus maximus demonstrate significantly greater EMG activity than the vastus medialis oblique and vastus lateralis (Table 2B) even though both muscles demonstrated their greatest EMG activity during the 90-60° arc. We speculate that the greater activity during this arc is needed to begin to propel the body (via hip extension) after the 90° arc (Hold Phase). In addition, the gastrocnemius muscle demonstrated a narrow level of EMG activity (≈7-11% of MVIC) during all arcs with its greatest activity occurring during the 60-90° arc (Fig. 3), but this activity was not significantly greater than that of any other muscle(Table 2B).
To further question the importance of the hamstrings during an unloaded squat, we reviewed Wickiewicz et al. (20) cadaveric study of lower extremity muscle architecture which demonstrated that the cross-sectional area of the hamstrings was approximately 40% smaller than that of the quadriceps. Thus, to offset the large forces imparted by the quadriceps, we believe that the hamstrings would need to demonstrate significantly more relative EMG activity than the quadriceps regardless of any advantages that might be realized as a result of line of pull or muscle length relationships. Our study, however, demonstrated hamstring activity of more than five times less than that of the quadriceps. Therefore, we believe these data provide adequate support from which to question the notion, as suggested by the co-contraction hypothesis, that hamstring muscle activity is sufficient to provide the necessary posterior tibial shear forces to counteract adequately the anterior tibial shear forces imparted by the quadriceps, thus providing a stabilizing force at the knee during an unloaded squat.
Since the hamstrings appear to be minimally active during closed kinetic chain activities such as an unloaded squat and lateral step-up, perhaps other factors, such as joint compressive forces (e.g., axial loading) and joint geometry (e.g., 9% posterior tilt of the proximal articulating surface of the tibia), play more integral roles in knee joint stability than does muscular co-contraction of the hamstrings and quadriceps. In an attempt to understand and explain these factors, we reviewed several cadaveric studies(8,10,16) that have attempted to simulate a closed kinetic chain environment by applying joint compressive forcesvia axial loading to their experimental models. Hsieh and Walker(8) investigated the effects of varying amounts of axial loading on applied anterior-posterior (A-P) and rotary forces to cadaveric knees at 0° and 30° of knee flexion. The researchers reported a reduction in anterior-posterior (A-P) and rotary laxity with increasing axial loads, citing the conformity of the condylar surfaces as being the most important factor contributing to this reduction of motion(8). Similarly, Markolf et al. (10) reported the effect of axial loading on A-P shear, medial-lateral shear, valgus-varus, and rotary laxity in loaded and unloaded knees at 0° and 20° of knee flexion before and after medial and lateral meniscectomy. They concluded that axial loading of the joint reduced laxity in all planes. Furthermore, medial and lateral meniscectomy significantly increased A-P laxity only, with a maximum force (925N) applied to the fully extended knee. Shoemaker and Markolf (16) studied the effect of joint load on ACL and/or medial collateral ligament (MCL)-deficient cadaveric knees. They reported that although joint load increased joint stability it was not able to control rotary instability commonly encountered during closed kinetic chain change of direction activities (i.e., pivot shift), when either the ACL or MCL was sectioned.
In a classic study with human subjects, Henning et al.(6) investigated the effect of various open and closed kinetic chain activities on the elongation of the ACL in subjects with grade II ACL sprains. They concluded that closed kinetic chain exercises, such as the single-leg half-squat, produced a marked decrease in ACL strain as compared with open kinetic chain exercises. In a similar study, Yack et al.(22) studied the effects of an open kinetic chain knee extension exercise and a closed kinetic chain unloaded squat on anterior tibial shear in patients with ACL-deficient knees. They reported decreased anterior tibial displacement in both the involved and uninvolved knees during the squat as compared with the knee extension exercise, which, conversely, increased anterior tibial displacement. The findings of Yack et al.(22) are consistent with those of Henning et al.(6) that closed kinetic chain activities decrease anterior tibial displacement and ACL elongation.
Although these five studies demonstrated the positive effects of axial loading on knee joint stability, they did not examine the role of co-contraction of the hamstrings and quadriceps. This gap in the literature necessitates a study investigating all factors, including axial loading, joint geometry, anterior shear, and muscle performance, during closed kinetic chain activities such as the squat and lateral step-up in both ACL-deficient and normal knees.
Limitations. It has been suggested that EMG studies that used in-dwelling electrodes are superior to those which used surface electrodes, owing to the potential for cross talk between surface electrodes. We believe that our reliability results, as well as the unique muscle activity pattern of the rectus femoris as compared with the vastus medialis oblique and vastus lateralis, demonstrate that we have obtained a true representation of EMG activity during the squat. In addition, we believe there are advantages to surface electrodes as compared with indwelling electrodes, including the ability to obtain a broader sampling of motor units and the noninvasive nature of the technique. Further research comparing indwelling to surface EMG techniques is needed to support or refute our speculation.
It has been suggested that the use of open kinetic chain MVIC testing procedures at a single joint position is not biomechanically specific when used to generate normalized values for closed kinetic chain data when sampling EMG activity of multiple-function muscles. Nevertheless, we chose open chain techniques for MVIC testing instead of closed chain techniques because this has been the accepted method of testing to isolate a muscle during kinesiological studies. Furthermore, Schaub and Worrell(14), demonstrating the complexity of maximal closed chain testing, encountered extreme difficulty in stabilizing the trunk during a study involving a maximal isometric squat,. Therefore, to ensure acceptable reliability, we chose to remain with the accepted open chain MVIC testing procedures. Further research is needed to determine the potential for reliable closed chain MVIC testing positions and procedures.
During the 90° arc (Hold Phase), unloading forces cannot be accurately monitored visually. Although we did not attach a force plate to either the bench or bar to quantify any potential unloading forces during the Hold Phase, we wish to note that the EMG data during this phase was relatively consistent with the 60-90° arc for all muscles except the vastus lateralis. Therefore, we interpret this observation to indicate that unloading did not occur.
This study did not examine directly the potential mechanical efficiency of the six muscles, the hamstrings being of particular interest. During a 90° arc of motion, we would expect EMG activity levels to be affected by several factors related to mechanical efficiency, such as line of pull and length-tension relationships. Changes in hamstring EMG levels, however, were minimal (eccentric range ≈4-7% of MVIC, concentric range ≈7-12% of MVIC). Therefore, we question whether the hamstrings could elicit mechanically-efficient properties (i.e., demonstrate mechanical advantage) over such a large range of motion. To address this issue completely, however, would warrant an in-depth neurological and kinesiological investigation of the multidimensional ramifications of the muscles, which is beyond the scope of this study.
In conclusion, the findings of this study demonstrate that during a dynamic unloaded squat in normal subjects minimal hamstrings, gastrocnemius, and gluteus maximus EMG activity exists as compared with the highly active quadriceps in select arcs across the range of motion. In addition, we believe that the low levels of hamstring EMG activity demonstrated in this study reflect the low demands placed on the hamstring muscles to counter anterior shear forces acting at the proximal tibia. Finally, we speculate that joint compressive forces and joint geometry play very important roles in knee joint stability. Further research is needed, however, to support or refute the co-contraction hypothesis as well as to examine critically the role of other potential stabilizing factors in and around the knee joint.
1. Basmajian, J. V. Muscles Alive: Their Functions Revealed by Electromyography
, 4th Ed. Baltimore: Williams & Wilkins, 1989, p. 241.
2. Brask, B., R. H. Lueke, and G. L. Soderberg. Electromyographical analysis of selected muscles during the lateral step-up exercise. Phys. Ther.
3. Cook, T. M., C. L. Zimmerman, K. M. Lux, C. M. Neubrand, and T. D. Nicholson. EMG comparison of lateral step-up and stepping machine exercise. J. Orthop. Sports. Phys. Ther.
4. Decarlo, M., D. A. Porter, G. Gehlsen, and R. Bahamonde. Electromyographic and cinematographic analysis of the lower extremity during closed and open kinetic chain exercise. Isokin. Exerc. Sci.
5. Gryzlo, S. M., R. M. Patek, M. Pink, and J. Perry. Electromyographic analysis of knee
rehabilitation exercises. J. Orthop. Sports. Phys. Ther.
6. Henning, C. E., M. A. Lynch, and K. R. Glick. Anin vivo
strain gauge study of elongation of the anterior cruciate ligament. Am. J. Sports Med.
7. Hinkle, D. E., W. Wiersma, and S. G. Jurs. Applied Statistics for the Behavioral Sciences
, 2nd Ed. Boston: Houghton Mifflin, 1988, 418-422.
8. Hsieh, H. and P. S. Walker. Stabilizing mechanisms of the loaded and unloaded knee
joint. J. Bone Joint Surg. Am.
9. Lutz, G. E., R. A. Palmitier, K. N. An, and E. Y. Chao. Comparison of tibiofemoral joint forces during open-kinetic-chain and closed-kinetic-chain exercises. J. Bone Joint Surg. Am.
10. Markolf, K. L., W. L. Bargar, S. C. Shoemaker, and H. C. Amstutz. The role of joint load in knee
stability. J. Bone Joint Surg. Am.
11. More, R. C., B. T. Karras, R. Neiman, et al. Hamstrings: an anterior cruciate ligament protagonist an in vitro
study. Am. J. Sports Med.
12. Ohkoshi, Y., K. Yasuda, K. Kaneda, T. Wada, and M. Yamanaka. Biomechanical analysis of rehabilitation in the standing position.Am. J. Sports Med.
13. Palmitier, R. A., K-N An, S. G. Scott, and E. Y. S. Chao. Kinetic chain exercise in knee
rehabilitation. Sports Med.
14. Schaub P. A. and T. W. Worrell. EMG activity of six muscles and VMO:VL ratio determination during a maximal squat exercise.J. Sport. Rehab.
15. Shelbourne, D. K. and P. Nitz. Accelerated rehabilitation after anterior cruciate ligament rehabilitation. Am. J. Sports Med.
16. Shoemaker, S. C. and K. L. Markolf. Effects of joint load on the stiffness and laxity of ligament-deficient knees. J. Bone Joint Surg. Am.
17. Solomonow M., R. Baratta, B. H. Zhou, et al. The synergistic action of the anterior cruciate ligament and thigh muscles in maintaining joint stability. Am. J. Sports Med.
18. Steinkamp, L. A., M. F. Dillingham, M. D. Markel, J. A. Hill, and K. R. Kaufman. Biomechanical considerations in patellofemoral joint rehabilitation. Am. J. Sports Med.
19. Strout, P. E. and J. L. Fleiss. Intraclass correlations: Uses in assessing rater reliability. Psychol. Bull.
20. Wickiewicz, T. L., R. R. Roy, P. L. Powell, and V. R. Edgerton. Muscle architecture of the human lower limb. Clin. Orthop.
21. Worrell, T. W., E. A. Crisp, and C. N. Larosa. Comparison of EMG activity in the vastus medialis oblique, vastus lateralis, hamstrings, and gluteus maximus
during closed kinetic chain
exercises, submitted for publication.
22. Yack, H. J., C. E. Collins, and T. J. Whieldon. Comparison of closed and open kinetic chain exercise in the anterior cruciate ligament-deficient knee
. Am. J. Sports Med.