0363-5465/103/3131-0241$02.

00/0
THE AMERICAN JOURNAL OF SPORTS MEDICINE, Vol. 31, No. 2
© 2003 American Orthopaedic Society for Sports Medicine

The Soleus Muscle Acts as an Agonist for

the Anterior Cruciate Ligament
An In Vitro Experimental Study
John J. Elias, PhD, Alfred F. Faust, MD, Yung-Hua Chu, MS, Edmund Y. Chao, PhD,
and Andrew J. Cosgarea,* MD

From the Department of Orthopaedic Surgery and the Orthopaedic Biomechanics Laboratory,
Johns Hopkins University, Baltimore, Maryland

Background: Although the quadriceps muscles are known antagonists for the anterior cruciate ligament and the hamstring
muscles are known agonists, the influence of the calf muscles on knee stability is not well understood.
Hypothesis: The soleus muscle acts as an anterior cruciate ligament agonist and the gastrocnemius muscle acts as an anterior
cruciate ligament antagonist.
Study Design: Controlled laboratory study.
Methods: Six cadaveric knees were tested with individual and combined activation of the gastrocnemius and soleus muscles
to determine the influence of simulated muscle contraction on tibiofemoral motion.
Results: At all flexion angles, applying the soleus muscle force tended to translate the tibia posteriorly, whereas applying the
gastrocnemius muscle force tended to translate the tibia anteriorly. Applying the soleus and gastrocnemius muscle forces
together also tended to translate the tibia anteriorly. The average anterior and posterior tibial translations were greatest at 50°
of flexion.
Conclusions: The soleus muscle is capable of acting as an agonist for the anterior cruciate ligament and the gastrocnemius
muscle can act as an antagonist.
Clinical Relevance: A better understanding of the agonistic behavior of the soleus muscle on the anterior cruciate ligament may
lead to the development of training and rehabilitation strategies that could reduce the incidence of injury and improve function
in both patients with anterior cruciate ligament deficiency and patients who have undergone anterior cruciate ligament
reconstruction.
© 2003 American Orthopaedic Society for Sports Medicine

The ACL is the primary passive restraint against patho-
logic anterior tibial translation. Muscles that dynamically
resist anterior tibial translation act as ACL agonists,
whereas muscles that dynamically produce anterior tibial
translation act as ACL antagonists. Because the ham-
string muscles attach on the proximal tibia and exert a
force with a posterior orientation when the knee is flexed,
they are considered the primary ACL agonists.16, 18, 22
Conversely, the quadriceps muscles are considered ACL
antagonists.5, 22
* Address correspondence and reprint requests to Andrew J. Cosgarea,
MD, Orthopaedic Surgery, Johns Hopkins Sports Medicine, 10753 Falls Road,
Suite 215, Baltimore, MD 21093.
No author or related institution has received any financial benefit from
research in this study. See “Acknowledgment” for funding information.
The calf muscles are active during activities that put the
ACL at risk of rupture, such as cutting movements and
landing from a jump.13 However, investigators have
mainly focused on the influence of these muscles on ankle
stability.4, 20, 23 The calf muscles, the soleus muscle, and
the medial and lateral heads of the gastrocnemius muscle,
in particular, plantar flex the foot and can decelerate the
rotation of the tibia around the ankle to modulate ankle
dorsiflexion when the foot is planted. Because the heads of
the gastrocnemius muscle originate on the distal femur
and insert on the posterior calcaneus, thereby crossing the
knee joint, their role in knee stabilization has been inves-
tigated. Theoretical21 and in vivo experimental9 studies
have indicated that the gastrocnemius muscle acts as an
antagonist for the ACL, although another in vitro experi-
mental study indicated that activation of the gastrocne-
mius muscle strains the PCL, rather than the ACL.5

241
242 Elias et al. American Journal of Sports Medicine

Because the soleus muscle does not cross the knee joint,
the possible influence of soleus muscle loading on ACL
strain has generally been ignored. Unlike the gastrocne-
mius muscle, the soleus muscle originates on the proximal
tibia. With the foot planted, the force exerted by the soleus
muscle that resists forward rotation of the tibia about the
ankle could theoretically limit anterior translation of the
proximal tibia with respect to the distal femur. To the best
of our knowledge, no other investigators have explored the
role of the soleus muscle in stabilizing the knee joint. For
the current study, we hypothesized that the soleus and
gastrocnemius muscles have opposite influences on tibial
translation. We believe that activation of the gastrocne-
mius muscle tends to translate the tibia anteriorly,
thereby loading the ACL, whereas activation of the soleus
muscle tends to translate the tibia posteriorly, thereby
unloading the ACL. In vitro kinematic testing was per-
formed to investigate the proposed hypothesis.

MATERIALS AND METHODS

Six knees harvested from six cadavers with ages ranging
from 67 to 77 years were used for kinematic testing. The
knees were free of deformities and showed no signs of
prior surgeries. Each knee was dissected to isolate the
quadriceps tendon, the semimembranosus tendon, and the
biceps femoris tendon. The gastrocnemius and soleus
muscles were completely separated from one another, in-
cluding distally where the tendinous contributions to the
Achilles tendon were carefully dissected apart. All other
muscles were removed, without disrupting the joint cap-
sule, patellar retinaculum, or patellofemoral ligaments.
Before testing, each knee was aligned by using a standard-
ized radiographic technique described previously17 to pro-
duce a neutral varus-valgus orientation and internal-ex-
ternal rotation at 0° of flexion. Electromagnetic sensors
(Flock of Birds, Ascension Technology, Burlington, Ver-
mont) were secured to the femur, tibia, and patella to
quantify the kinematics of the knee. The tibiofemoral ro-
tations were quantified with a standardized knee coordi-
nate system11 established during the alignment process.
Tibial translations were expressed with respect to a local
coordinate system fixed to the femur at the intersection of
the knee screw axis and a sagittal plane bisecting the
knee. The knee screw axis was established between 0° and
90° of flexion. Each translation and rotation was set to
zero at full extension.
The cadaveric knees were used to characterize the in-
fluence of simulated muscle activation of the soleus and
gastrocnemius muscles on the motion of the tibia. A knee
simulator6, 18, 19 was used to control the quadriceps and
hamstring muscle forces and hip loads applied to the knee
at various flexion angles (Fig. 1). The hip load was applied
through the main actuator of a modified Bionix 858 ma-
terials testing platform (MTS, Eden Prairie, Minnesota).
Two additional independent actuators applied the ham-
string and quadriceps muscle loads through a cable and
pulley system. A cryoclamp device secured the quadriceps
tendon to one actuator. Two other cryoclamps secured the
semimembranosus and biceps femoris tendons to a second
Figure 1. Each knee was positioned at 20°, 50° and 80° of
flexion with loads applied to simulate quadriceps (quads) and
hamstring (hams) muscle forces. The gastrocnemius (gas-
trocs) and soleus muscle were loaded individually and in
combination by using constant force springs. Each knee was
tested with the ACL intact and with the ACL cut and with and
without application of an anterior load.
actuator through a linkage that divided the actuator load
equally between the two tendons. The gastrocnemius and
soleus muscles were each loaded using a 180-N constant
force spring. The springs were secured to a frame fixed to
the materials testing platform. The gastrocnemius and
soleus muscles were wrapped around low friction rollers
also fixed to the platform. The rollers were positioned to
reproduce the anatomic orientation of the gastrocnemius
and soleus muscles with respect to the tibia. To load the
gastrocnemius and soleus muscles, we extended the
springs and secured them to the muscles by using No. 5
braided nonabsorbable suture sewn into the tendon with
an interlocking stitch. The spring force was independent
of the spring extension because of the constant force
spring design, ensuring that each muscle carried a load of
180 N.
Before testing at each prescribed flexion angle, each
knee was placed in its neutral position, based on the
anatomic coordinate system. A 100-N load was applied to
the quadriceps and hamstring muscles, and the knee was
flexed to the appropriate angle. At 20° and 50° of flexion,
a 180-N force was applied to the hamstring muscles while
a 540-N force was applied to the quadriceps muscles. At
80° of flexion, a 180-N force was applied to the hamstring
Vol. 31, No. 2, 2003
muscles while a 720-N force was applied to the quadriceps
muscles. The hip load was varied to hold the knee at the
prescribed flexion angle but was approximately equal to
120 N for all tests. The hamstring muscle force was set to
180 N on the basis of previous theoretical calculations
showing that the peak gastrocnemius muscle force is ap-
proximately equal to the peak hamstring muscle force for
multiple functional activities.12, 15 The quadriceps muscle
force at each flexion angle was based on a previously
measured ratio of the quadriceps muscle force to the ham-
string muscle force for the knee simulator during simu-
lated closed chain knee extension. With the knee at each
fixed flexion angle (20°, 50°, and 80°) and the quadriceps
and hamstring muscle forces applied, the gastrocnemius
and soleus muscles were loaded individually and in com-
bination to simulate activation of these muscles, while the
motion of the tibia in response to each load was quantified.
Each knee was initially tested in the intact state (test
case 1). The gastrocnemius and soleus muscle forces were
also applied after applying an additional 110-N force to
translate the tibia anteriorly (test case 2). The anterior
force was applied in the anterior-posterior plane, perpen-
dicular to the long axis of the tibia, by using a suspended
weight and a cable and pulley system. An S-hook on the
end of the cable was hooked to a plastic U-bolt attached to
a Kirschner wire that passed through the proximal tibia.
The whole set of tests was performed for each knee with
the ACL intact and after sectioning the ACL (no anterior
load, test case 3; anterior load, test case 4). The ACL was
sectioned through a 6-cm medial parapatellar approach.
The incision was closed with No. 2 braided nonabsorbable
suture in an interrupted figure-of-eight fashion. After
testing was complete, we performed a second arthrotomy
on each specimen to confirm that the ACL had been com-
pletely sectioned.
For each combination of test case and flexion angle,
statistical comparisons were performed to determine the
influence of application of the gastrocnemius and soleus
muscle loads on the position of the tibia. A Friedman test
for nonparametric comparison of multiple groups with
repeated measures was used to determine whether appli-
cation of the gastrocnemius or soleus muscle load, or both,
significantly (P ⬍ 0.05) influenced the position of the tibia.
A nonparametric form of the Student-Neuman-Keuls test
was used for post hoc comparisons. Because of missing
data, only five knees were considered in statistical com-
parisons for the ACL-intact knees with no anterior load at
50° of flexion and for the ACL-deficient knees with an
anterior load at 80° of flexion. The same statistical tests
were also used to determine whether applying the ante-
rior load or cutting the ACL influenced the initial position
of the tibia before loading of the gastrocnemius and soleus
muscles.
RESULTS
As expected, applying the anterior load and cutting the
ACL caused the tibia to translate anteriorly with respect
to the femur. At both 20° and 80° of flexion, with quadri-
ceps and hamstring muscle co-contraction force applied
The Soleus Muscle as ACL Agonist 243
but before the soleus and gastrocnemius muscles were
loaded, the tibia was significantly less anterior when the
ACL was intact and no anterior load was applied (test case
1) than for any of the three other test cases (Fig. 2). Also,
at these flexion angles, the tibia was significantly more
anterior when the ACL was cut and an anterior load was
applied (test case 4) than for any of the three other test
cases. At 50° of flexion, the variation between test cases
was not statistically significant (P ⫽ 0.13). The average
anterior shift from test case 1 to test case 4 was approxi-
mately 6 mm at 20° of flexion, 1.5 mm at 50° of flexion, and
less than 1 mm at 80° of flexion. The test case did not
significantly influence the original varus/valgus orienta-
tion, internal/external rotation, or medial/lateral shift of
the tibia at any flexion angle.
When the soleus muscle force was applied, the tibia
translated posteriorly with respect to the femur. At 20° of
flexion (Fig. 3A), the measured tibial shifts caused by
application of soleus or gastrocnemius muscle forces, or
both, were not significant for any of the four test cases
(0.06 ⬍ P ⬍ 0.20). The posterior tibial shift that occurred
after application of a load through the soleus tendon was
statistically significant for all four test cases at 50° of
flexion (Fig. 3B). For the four test cases, the average
posterior shift of the tibia occurring in response to loading
the soleus muscle ranged from 0.24 ⫾ 0.06 mm to 0.36 ⫾
0.07 mm. At 80° of flexion (Fig. 3C), the posterior shift was
statistically significant only for test case 2 (P ⬍ 0.09 for
test case 3, P ⬍ 0.06 for test case 4).
When the gastrocnemius muscle force was applied, the
tibia translated anteriorly with respect to the femur. Al-
though no statistically significant changes were measured
at 20° of flexion, the anterior shift of the tibia after appli-
cation of a load through the gastrocnemius tendon was
Figure 2. The average anterior position of the tibia with
respect to the femur for the four test cases and the three
flexion angles before loading the soleus and gastrocnemius
muscles. The error bars represent the standard error of the
mean.
244 Elias et al.
Figure 3. The average (⫾standard error) tibial anterior shift
caused by applying a 180-N force to the soleus and gastroc-
nemius muscles individually and in combination at 20° (A),
50° (B), and 80° (C) of knee flexion. Each knee was tested
with the ACL intact and no anterior load (test case 1), with the
ACL intact and a 110-N anterior load (test case 2), with the
ACL cut and no anterior load (3), and with the ACL cut and a
110-N anterior load (4). A star indicates a statistically signif-
icant anterior or posterior shift (P ⬍ 0.05).
significant for all test cases except case 3 at 50° of flexion
(P ⬍ 0.06). At 50° of flexion, the largest anterior shift of
the tibia in response to a gastrocnemius muscle load was
0.5 ⫾ 0.1 mm, for test case 1. In general, the anterior shift
of the tibia in response to a gastrocnemius muscle load
was slightly smaller at 80° of flexion than at 50° of flexion,
but the shift was still statistically significant for all four
test cases at 80° of flexion.
American Journal of Sports Medicine
Applying the soleus and gastrocnemius muscle forces
together tended to translate the tibia anteriorly with re-
spect to the femur. The anterior shift of the tibia in re-
sponse to application of a combined gastrocnemius and
soleus muscle load was statistically significant for all test
cases except case 1 at 50° of flexion (P ⬍ 0.07) and was
statistically significant for all test cases at 80° of flexion.
Application of the calf muscle loads did not significantly
influence the varus/valgus orientation, the internal/exter-
nal rotation, or the medial/lateral shift of the tibia at any
flexion angle for any test case.
DISCUSSION
For this in vitro study, the soleus muscle acted as an ACL
agonist, whereas the gastrocnemius muscle acted as an
ACL antagonist. To our knowledge, no other study has
shown that the soleus muscle can act as an ACL agonist.
A previous in vivo study, in which ACL strain was meas-
ured while different muscles were activated, showed that
the gastrocnemius muscle acts as an ACL antagonist.9
The anterior translations produced by applying a gastroc-
nemius muscle force in this study would cause a smaller
increase in ACL strain than that seen in the in vivo
study.7, 8 This difference is most likely caused by differ-
ences in the loading conditions. The forces applied
through the quadriceps muscles, which are also ACL an-
tagonists, were larger for this in vitro study than for the
previous in vivo study. In addition, this in vitro study
modeled a simulated closed kinetic chain squat with co-
contraction of the quadriceps and hamstring muscles,
whereas the in vivo study was performed with the knee in
a nonweightbearing position. Joint loading from co-con-
traction and weightbearing increases joint stiffness and
decreases ACL strain.8
One unexpected finding of this study was that combined
loading of the soleus and gastrocnemius muscle produced
tibial translations similar to those seen after loading the
gastrocnemius muscle alone. The explanation of this phe-
nomenon is not clear from our data. We do believe, how-
ever, that this model probably oversimplifies the forces
acting on the lower leg. We applied a constant 180-N force
to both the gastrocnemius and soleus muscle at all flexion
angles. The literature suggests that even in full extension
the soleus muscle probably exerts a greater plantar flex-
ion force than the gastrocnemius muscle.2, 3, 10 With in-
creasing knee flexion, the length of the gastrocnemius
muscle, which crosses both the knee and ankle joints,
decreases as the distance from the origins on the distal
femur to the insertion on the posterior calcaneus de-
creases. In contrast, knee flexion does not change the
length of the soleus muscle, which crosses only the ankle
joint. Change in muscle length affects the capacity of the
muscle to generate contractile force. This phenomenon
results in a decreased ability of the gastrocnemius muscle
to generate force with increasing knee flexion angles in
comparison with the soleus muscle.2, 3, 10 Although it may
have been more accurate for us to use a soleus muscle
force that was greater than the gastrocnemius muscle
force, we chose to use equal forces in our simplified model
Vol. 31, No. 2, 2003
because the actual soleus to gastrocnemius muscle force
ratio has not been well defined, and we wanted to test our
hypothesis using a worst-case scenario. The possibility
exists that a greater soleus to gastrocnemius muscle force
ratio would have decreased anterior tibial translation or
even translated the tibia posteriorly. Further study of the
specific behavior of the calf muscles during sports activity
is clearly warranted.
The tibial translations caused by gastrocnemius and so-
leus muscle loading varied with the flexion angle. The ante-
rior translations were smallest at 20° of flexion. A possible
contributor to the decreased tibial anterior translation in
response to gastrocnemius muscle loading at 20° of flexion is
the original anterior position of the tibia with respect to the
femur. For these in vitro test conditions, the initial anterior
position of the tibia was greatest at 20° of flexion. Therefore,
tensioning the gastrocnemius muscle may not have applied
pressure to the proximal tibia, which may be the mechanism
by which the gastrocnemius muscle affects anterior tibial
translation. The largest tibial translations were seen at 50°
of flexion. The decrease in the tibial translations from 50° to
80° of flexion may be related to the increase in the quadri-
ceps muscle force. Additionally, the hamstring muscles be-
come relatively more important as the direction of force
becomes more posterior with increasing knee flexion. Both of
these factors act together to increase joint stiffness, which
may have limited the ability of the soleus and gastrocnemius
muscles to translate the tibia with respect to the femur. The
in vivo forces during sports activities or even during just
simple squatting are not well understood. Determination of
these forces would provide valuable information and lead to
a better understanding of dynamic stabilization.
Although four test cases with different combinations of
anterior load and ACL status were studied, the test case
had little influence on the relationship between calf mus-
cle loading and tibial translations. Sectioning the ACL
and applying an anterior load caused the tibia to be more
anterior before loading of the gastrocnemius and soleus
muscles, particularly at 20° of flexion. The anterior shift
caused by applying an anterior load and sectioning the
ACL decreased from 20° to 80° of flexion, most likely
because of the direction of the applied hamstring muscle
force. The hamstring muscle force becomes more posterior
as the flexion angle increases and is therefore better able
to resist tibial anterior translation.16, 18, 22 Even though
the test case had the greatest influence on the initial
position of the tibia at 20° of flexion, the test case did not
influence the statistical significance of the results at 20° of
flexion. The flexion angle was a more important factor in
determining the statistical significance of the tibial trans-
lations than the test case. The similarity of the results for
the four test cases indicates that if these results hold true
in vivo, the soleus muscle can potentially limit pathologic
tibial translation for an ACL-intact or ACL-deficient knee.
In vivo, an ACL injury is most likely to occur when
someone is landing from a jump or performing a quick stop
or a pivot.13 The ACL typically tears with the knee flexed,
as the quadriceps muscles fire to decelerate knee flexion.
The large accelerations that occur and the large muscle
forces that are applied to the knee cannot be reproduced in
The Soleus Muscle as ACL Agonist 245
the situation of testing cadaveric tissue on a hydraulically
controlled knee simulator. In this study, the knees were
tested at individual knee flexion angles, with relatively
large quadriceps muscle loads with respect to the ham-
string, soleus, and gastrocnemius muscle loads. Theoreti-
cal modeling has shown that the ratio of the quadriceps to
hamstring muscle force can be greater than 2:1 during
normal gait1 and can exceed 8:1 during a seated knee
extension exercise,15 whereas the peak force exerted by
the hamstring and gastrocnemius muscles are simi-
lar.12, 15 For the current study, the ratio of the quadriceps
to hamstring muscle force ranged from 3:1 at 20° of flexion
to 4:1 at 80° of flexion, based on previous weightbearing
knee extension measurements made with the knee simu-
lator. The quadriceps muscle force was increased as the
static flexion angle increased to approximate in vivo load-
ing conditions. The magnitude of the hamstring muscle
force was equal to the magnitude of the gastrocnemius
and soleus muscle forces. Because the quadriceps muscle
force has an anterior component acting on the tibia, de-
creasing the ratio of the quadriceps muscle force to the
soleus muscle force may have increased the ability of the
soleus muscle to translate the tibia posteriorly. A large
quadriceps to soleus muscle force ratio was used for this
study, however, to provide a rigorous test of the proposed
hypothesis.
The results of this study suggest that the soleus muscle
can potentially act as an ACL agonist, protecting it during
dynamic activities. Even though the soleus muscle does
not cross the knee joint, soleus muscle loading can produce
a moment that rotates the tibia with respect to the ankle,
thereby shifting the proximal tibia posteriorly at the knee
joint. For static testing at individual flexion angles with
simulated quadriceps and hamstring muscle co-contrac-
tion, simulated activation of the soleus and gastrocnemius
muscles produced relatively small tibial shifts. During
dynamic activities, these muscles may have a greater in-
fluence on ACL strain, however. The muscle forces and
external forces applied to the tibia were dramatically
smaller than what would be expected in vivo for activities
such as landing or pivoting, because of the loading limi-
tations of the knee simulator and the cadaveric tissue. The
goal of the current study was to show the direction of tibial
translation in response to the application of gastrocne-
mius and soleus muscle forces, with the understanding
that the translation magnitude could increase for in vivo
load levels.
The muscle firing pattern and the loading rate of each
muscle are important parameters governing tibiofemoral
kinematics. The results of this study indicate that simul-
taneous activation of both calf muscles translates the tibia
anteriorly, but that individual activation of the soleus
muscle could decrease tibial anterior translation. Because
a previous study showed that the soleus muscle tends to
fire before the gastrocnemius muscle during cutting and
pivoting activities,20 we believe that early activation of the
soleus muscle during dynamic activities could be a mech-
anism acting to protect the ACL. This might be particu-
larly important in female athletes, who have been shown
to take longer to generate maximum hamstring muscle
246 Elias et al.
torque during isokinetic testing.14 Additionally, a better
understanding of the relative contributions of the gastroc-
nemius and soleus muscles at various flexion angles is
needed to more clearly characterize the influence of the
soleus muscle on tibial translation.
Although additional studies are needed to fully charac-
terize how the soleus muscle influences ACL strain during
dynamic activities, confirmation of ACL-agonistic behav-
ior could lead to a better understanding of the cause of
ACL injuries. A better understanding of the ACL-agonis-
tic behavior of the soleus muscle may also lead to devel-
opment of training methods that reduce the incidence of
injury and rehabilitation strategies that improve function
in patients with ACL deficiency and in those who have
undergone ACL reconstruction.
ACKNOWLEDGMENT
Funding for this work was provided by an Institutional
Research Grant from Johns Hopkins University.
REFERENCES
1. Anderson FC, Pandy MG: Static and dynamic optimization solutions for
gait are practically equivalent. J Biomech 34: 153–161, 2001
2. Cresswell AG, Loscher WN, Thorstensson A: Influence of gastrocnemius
muscle length on triceps surae torque development and electromyo-
graphic activity in man. Exp Brain Res 105: 283–290, 1995
3. Delp SL, Zajac FE: Force- and moment-generating capacity of lower-
extremity muscles before and after tendon lengthening. Clin Orthop 284:
247–259, 1992
4. Duncan A, McDonagh MJ: Stretch reflex distinguished from pre-pro-
grammed muscle activations following landing impacts in man. J Physiol
526: 457– 468, 2000
5. Durselen L, Claes L, Kiefer H: The influence of muscle forces and external
loads on cruciate ligament strain. Am J Sports Med 23: 129 –136, 1995
6. Elias JJ, DesJardins JD, Faust AF, et al: Size and position of a single
condyle allograft influence knee kinematics. J Orthop Res 17: 540 –545,
1999
American Journal of Sports Medicine
7. Fleming BC, Beynnon BD, Nichols CE, et al: An in vivo comparison of
anterior tibial translation and strain in the anteromedial band of the ante-
rior cruciate ligament. J Biomech 26: 51–58, 1993
8. Fleming BC, Renstrom PA, Beynnon BD, et al: The effect of weightbearing
and external loading on anterior cruciate ligament strain. J Biomech 34:
163–170, 2001
9. Fleming BC, Renstrom PA, Ohlen G, et al: The gastrocnemius muscle is
an antagonist of the anterior cruciate ligament. J Orthop Res 19: 1178 –
1184, 2001
10. Fugl-Meyer AR, Sjostrom M, Wahlby L: Human plantar flexion strength
and structure. Acta Physiol Scand 107: 47–56, 1979
11. Grood ES, Suntay WJ: A joint coordinate system for the clinical descrip-
tion of three-dimensional motions: Application to the knee. J Biomech Eng
105: 136 –144, 1983
12. Heller MO, Bergmann G, Deuretzbacher G, et al: Musculo-skeletal loading
conditions at the hip during walking and stair climbing. J Biomech 34:
883– 893, 2001
13. Hewett TE, Lindenfeld TN, Riccobene JV, et al: The effect of neuromus-
cular training on the incidence of knee injury in female athletes. A pro-
spective study. Am J Sports Med 27: 699 –706, 1999
14. Huston LJ, Wojtys EM: Neuromuscular performance characteristics in
elite female athletes. Am J Sports Med 24: 427– 436, 1996
15. Li G, Kaufman KR, Chao EY, et al: Prediction of antagonistic muscle
forces using inverse dynamic optimization during flexion/extension of the
knee. J Biomech Eng 121: 316 –322, 1999
16. Li G, Rudy TW, Sakane M, et al: The importance of quadriceps and
hamstring muscle loading on knee kinematics and in-situ forces in the
ACL. J Biomech 32: 395– 400, 1999
17. MacWilliams BA, DesJardins JD, Wilson DR, et al: A repeatable alignment
method and local coordinate description for knee joint testing and kine-
matic measurement. J Biomech 31: 947–950, 1998
18. MacWilliams BA, Wilson DR, DesJardins JD, et al: Hamstrings cocontrac-
tion reduces internal rotation, anterior translation, and anterior cruciate
ligament load in weight-bearing flexion. J Orthop Res 17: 817– 822, 1999
19. Mizuno Y, Kumagai M, Mattessich SM, et al: Q-angle influences tibiofemo-
ral and patellofemoral kinematics. J Orthop Res 19: 834 – 840, 2001
20. Neptune RR, Wright IC, van den Bogert AJ: Muscle coordination and
function during cutting movements. Med Sci Sports Exerc 31: 294 –302,
1999
21. O’Connor JJ: Can muscle co-contraction protect knee ligaments after
injury or repair? J Bone Joint Surg 75B: 41– 48, 1993
22. Renstrom P, Arms SW, Stanwyck TS, et al: Strain within the anterior
cruciate ligament during hamstring and quadriceps activity. Am J Sports
Med 14: 83– 87, 1986
23. Santello M, McDonagh MJ: The control of timing and amplitude of EMG
activity in landing movements in humans. Exp Physiol 83: 857– 874, 1998