0363-5465/102/3030-0483$02.00/0
THE AMERICAN JOURNAL OF SPORTS MEDICINE, Vol. 30, No. 4
© 2002 American Orthopaedic Society for Sports Medicine
Motor Control of the Vastus Medialis
Oblique and Vastus Lateralis Muscles Is
Disrupted During Eccentric Contractions in
Subjects with Patellofemoral Pain
Tammy M. Owings,* MS, and Mark D. Grabiner,†‡ PhD
From *the Department of Biomedical Engineering, The Cleveland Clinic Foundation,
Cleveland, Ohio, and the †School of Kinesiology, University of Illinois at Chicago,
Chicago, Illinois
Background: Inappropriate control of the vastus medialis oblique and vastus lateralis muscles by the central nervous system
can contribute to maltracking of the patella.
Hypothesis: The activation timing and amplitude of the vastus medialis oblique and vastus lateralis muscles will be different
between normal subjects and patients with patellofemoral pain.
Study Design: Controlled laboratory study.
Methods: Subjects with patellofemoral pain and asymptomatic control subjects performed maximum voluntary knee extension
contractions initiated from a flexed and an extended position. The activation timing and amplitude of the vastus lateralis and
vastus medialis oblique muscles were quantified from the recorded electromyographic signals.
Results: There were no between-group differences in activation timing. The activation amplitude of the vastus medialis oblique
and vastus lateralis muscles of the patellofemoral pain subjects was altered to the greatest extent during eccentric contractions
and differed significantly from that of control subjects.
Conclusions: The activation amplitudes of the vastus medialis oblique and vastus lateralis muscles of subjects with patellofemoral pain are consistent with a laterally tracking patella during eccentric contractions.
Clinical Relevance: The findings suggest the clinical importance of determining whether altered activation patterns are sensitive
to rehabilitation, and, if so, if subjective reports of knee joint pain and function parallel changes in the activation patterns as a
result of rehabilitation.
© 2002 American Orthopaedic Society for Sports Medicine
can subsequently affect patellofemoral reaction forces and
pressures. On the basis of this premise, an important
functional role has been attributed to the vastus medialis
oblique muscle in the causation of patellofemoral tracking
disorders. Therefore, the vastus medialis oblique muscle is
often targeted in rehabilitation protocols. However, supportive quantitative evidence for both the functional importance of the vastus medialis oblique muscle and for the
effectiveness of rehabilitation that purportedly has a selective effect on it has not been conclusive.
The rationale for an important biomechanical role of the
vastus medialis oblique muscle in patellar tracking is
substantiated by its physiologic and biomechanical properties. Compared with the vastus lateralis muscle, the
vastus medialis oblique muscle has a smaller physiologic
cross-sectional area13 and thus a smaller estimated max-
Patellofemoral pain is the most prevalent knee disorder of
children and young adults; however, the causative mechanisms remain imprecisely defined.10, 14 Aberrant patellar
tracking is associated with some patellofemoral joint disorders. The position of the patella within the femoral
trochlea, controlled by passive and dynamic mechanisms,
influences patellofemoral joint reaction forces and pressures in cadaver specimens.7 Because the medially and
laterally directed muscular forces applied to the patella
influence the patellofemoral position in vivo, these forces
‡ Address correspondence and reprint requests to Mark D. Grabiner, PhD,
School of Kinesiology (M/C 194), University of Illinois at Chicago, Chicago, IL
60608-1516.
Funding was received from companies related to products used in this
study. See “Acknowledgments” for funding information.
483
484
Owings and Grabiner
imum contraction force. In addition, compared with the
vastus lateralis muscle, the vastus medialis oblique is
composed of a larger percentage of oxidative type I muscle
fibers.4 This difference in fiber type would be expected to
cause the vastus medialis oblique muscle to possess a
slower maximum contraction velocity than the vastus
lateralis muscle. Biomechanically, the pennation angle of
the vastus medialis oblique muscle is well suited to apply
a medially directed force to the patella. However, the
larger contraction force and higher contraction velocity of
the vastus lateralis muscle would be expected to dominate
the movement of the patella unless the activation patterns
of these muscles accounts for these differences.
A plausible activation strategy of the central nervous
system that could diminish the tendency of the vastus
lateralis muscle to dominate the movement of the patella
would be to activate the vastus medialis oblique muscle
earlier than or to a greater relative extent than the vastus
lateralis, or both earlier and to a greater extent. We designed a protocol that could test this hypothesis by exploiting the dependency of the potential for mediolateral patellar movement on the knee joint angle. With the knee
flexed to 90°, the mediolateral movement of the patella is
limited, in part by the geometry of the intercondylar
groove. In contrast, with the knee extended, the extent to
which the patella can move medially and laterally is increased because the patella then resides in a more shallow
aspect of the intercondylar groove. The purpose of the
study was to determine whether the activation timing and
activation amplitude of the vastus medialis oblique and
vastus lateralis muscles are related to the extent to which
the patella can move medially and laterally within the
intercondylar groove. In addition, we determined whether
the activation timing and the activation amplitude of the
vastus medialis oblique and vastus lateralis muscles are
altered by the presence of patellofemoral pain. Two experimental questions were addressed in the study. First, is
the vastus medialis oblique muscle activated earlier or to
a greater extent than the vastus lateralis muscle to accommodate knee joint angles at which the potential for
mediolateral patellar movement is increased? Second,
does the activation timing and the amplitude of the vastus
medialis and vastus lateralis muscles of subjects with
patellofemoral pain differ from those of asymptomatic
subjects?
MATERIALS AND METHODS
Subjects volunteered to participate in this study and were
placed into an experimental group (N ⫽ 20) or a control
group (N ⫽ 14) based on the presence of symptoms of
patellofemoral pain in either one or both knees. The experimental group consisted of 12 women with a mean age
of 33.7 ⫾ 6.9 years, height of 1.65 ⫾ 0.06 meters, and
weight of 670.8 ⫾ 156.5 N, and 8 men with a mean age of
29.1 ⫾ 10.7 years, height of 1.77 ⫾ 0.08 meters, and
weight of 884.8 ⫾ 220.3 N. When subjects had bilateral
knee pain, the knee with the greatest subjectively reported pain was tested. The control group was composed of
4 women with a mean age of 22.3 ⫾ 1.6 years, height of
American Journal of Sports Medicine
1.63 ⫾ 0.07 meters, and weight of 667.5 ⫾ 146.7 N, and 10
men with a mean age of 24.5 ⫾ 2.3 years, height of 1.81 ⫾
0.06 meters, and weight of 752.1 ⫾ 124.2 N. They were
healthy, recreationally active, and reported no history of
knee injury to the right leg.
The protocol prescribed maximum voluntary concentric
and eccentric knee extension contractions. However, because of the potential for discomfort of the subjects with
patellofemoral pain, especially during eccentric contractions, these subjects performed fewer repetitions and at a
slower isokinetic velocity than did the control subjects.
The control subjects performed three concentric and three
eccentric contractions at 60 deg/sec with their right knee.
The patellofemoral pain subjects performed two concentric and two eccentric contractions at 15 deg/sec with the
painful knee. The order of concentric and eccentric contractions was randomized.
The knee extension contractions were performed on a
Kin-Com isokinetic dynamometer (Chattanooga Corporation, Chattanooga, Tennessee). Subjects were seated upright in the dynamometer with their waist and distal
thigh stabilized with straps. The hip-trunk angle was
fixed at approximately 100°. All contractions were performed through a 60° range of motion. Concentric and
eccentric contractions were initiated from a knee flexion
angle of 80° and 20°, respectively (full extension, 0°).
Bipolar silver-silver chloride surface electrodes (recording diameter, 8 mm; center-to-center distance, 20 mm)
were placed over the vastus lateralis, vastus medialis
oblique, and biceps femoris muscles of the tested leg. Electrode site preparation, which included shaving, light abrasion, and cleaning with isopropyl alcohol, generally reduced interelectrode resistance to less than 5 and 10 k⍀
for men and women, respectively. The long axis of the
electrodes was positioned over each muscle in the assumed direction of the underlying muscle fibers. A reference electrode was placed over the fibular head of the
tested leg.
The detected EMG signals were amplified using a Noraxon telemetered EMG system (gain, 1000; CMRR, 130
dB; bandpass, 15 to 500 Hz; Noraxon USA, Inc., Scottsdale, Arizona). The transmitted signals were input to an
analog-to-digital circuit (Data Translation, 2801A, Marlboro, Massachusetts), digitized at 1 kHz, and stored. The
same analog-to-digital circuit was used to digitize the
force, speed, and position signals from the Kin-Com
dynamometer.
Two dependent variables, representing activation timing and activation amplitude, were extracted from the
digitized EMG data. The time at which the vastus lateralis and vastus medialis oblique muscles became active was
determined with use of a custom interactive graphics and
analysis package. The onset and offset times were provided by the software package after the investigator visually positioned a vertical cursor at the location. The muscles were deemed active when the EMG signal surpassed
the resting values. Because the contractions were vigorously performed, the onset of activation was unambiguous. The relative timing of the vastus lateralis and the
vastus medialis oblique muscles, expressed in millisec-
Vol. 30, No. 4, 2002
onds, was determined by subtracting the onset time of the
vastus medialis oblique from that of the vastus lateralis.
Thus, a negative value indicates that activation of the
vastus lateralis muscle preceded activation of the vastus
medialis oblique muscle.
The activation signals of the vastus lateralis and vastus
medialis oblique muscles were full-wave rectified and integrated for 100 ms after activation onset to derive a
measure of activation amplitude. This period of time generally preceded the onset of the dynamometer motion that
was set, in software, as a threshold of knee extension force
required to trigger motion of the Kin-Com dynamometer.
Because the analysis window preceded motion of the dynamometer, the EMG signal during this period was not
influenced by afferent feedback that signals changes in
muscle length and muscle shortening velocity. Furthermore, the initial muscle force generated to trigger motion
of the Kin-Com dynamometer is always isometric, regardless of whether the contraction that followed triggering of
the dynamometer motion was concentric or eccentric.3 The
integrated values for each muscle were expressed as a
percentage of the averaged value from the contraction
initiated with the knee flexed.
A two-by-two (knee joint angle by group) analysis of
variance with repeated measures for the knee joint angle
was used to analyze activation timing for the two knee
joint angles from which the contractions were initiated.
Post hoc single-sample t-tests were used to compare the
activation timing to the value zero, effectively testing the
null hypothesis that between-muscle differences were not
significant. Because the activation amplitude during the
contraction initiated with the knee extended was normalized to the value derived during the contraction initiated
with the knee flexed, single-sample t-tests were used to
compare knee joint angles to the value 1. A two-by-two
(muscle by group) repeated-measures analysis of variance
was used to compare the activation amplitude of the vastus lateralis and the vastus medialis oblique muscles.
Statistical analysis was performed with SPSS Version 7.0
(SPSS, Chicago, Illinois), and a P value of 0.05 was accepted as reflecting statistical significance.
Motor Control in Subjects with Patellofemoral Pain
485
TABLE 1
Activation Timing (in milliseconds) in the Control Subjects and
Subjects with Patellofemoral Pain as a Function of Knee Joint
Angle from Which the Contraction was Initiateda
Group
Flexed
Extended
Control
Patellofemoral pain
5.5 ⫾ 11.0
⫺1.4 ⫾ 16.7
4.5 ⫾ 10.9
⫺6.7 ⫾ 18.9
a
Activation timing is expressed as a difference (vastus lateralis muscle ⫺ vastus medialis oblique muscle) for which a negative
value indicates that the activation of the vastus lateralis muscle
preceded that of the vastus medialis oblique muscle.
ralis muscle by 5 ⫾ 9 ms (pooled across initial knee joint
angles). In contrast, in the patellofemoral pain subjects,
activation of the vastus medialis oblique muscle preceded
that of the vastus lateralis muscle by 4 ⫾ 15 ms. However,
although the analysis of variance revealed significant between-group differences (P ⫽ 0.05) (Table 1), the subsequent post hoc analysis revealed that the activation timing values of the control and patellofemoral pain groups
were not significantly different from zero (P ⫽ 0.06 and
0.23, respectively).
The vastus medialis oblique muscle was not activated to
a greater extent than the vastus lateralis muscle when
contractions were initiated with the knee extended,
which, compared with the flexed knee position, increased
RESULTS
The vastus medialis oblique muscle was not activated
earlier relative to the vastus lateralis muscle during contractions initiated with the knee extended, a position that,
compared with the flexed-knee position, increases the extent to which the patella can move medially and laterally
within the intercondylar groove. The effect of the knee
joint angle at which contractions were initiated on the
activation timing of the vastus medialis oblique and vastus lateralis muscles was not significant (P ⫽ 0.30) (Table
1). In addition, the knee joint angle by group interaction
term was not significant (P ⫽ 0.48).
The activation timing of the vastus medialis oblique and
vastus lateralis muscles was not affected in the subjects
with patellofemoral pain, compared with control subjects.
Among the control group subjects, activation of the vastus
medialis oblique muscle followed that of the vastus late-
Figure 1. The normalized activation amplitude of the vastus
lateralis muscle was significantly greater than that of the
vastus medialis oblique muscle during maximum voluntary
knee extensions initiated with the knee in extension in control
subjects (N ⫽ 14) and patellofemoral pain subjects (N ⫽ 20).
The initial activation amplitude of the vastus medialis oblique
and vastus lateralis muscles was significantly greater in the
patellofemoral pain subjects compared with the control subjects. The activation levels are normalized to the activation
values observed during maximum voluntary knee extension
initiated with the knee flexed and that are represented by the
value 1 on the vertical axis.
486
Owings and Grabiner
the extent to which the patella can move medially and
laterally within the intercondylar groove (Fig. 1). In the
control subjects, the activation amplitude of the vastus
medialis oblique muscle during the contractions initiated
with the knee extended was significantly decreased compared with the contractions initiated with the knee flexed
(P ⫽ 0.02). However, the decreased activation amplitude
of the vastus lateralis muscle was not significant (P ⫽
0.07). In the patellofemoral pain subjects, activation amplitude of the vastus lateralis muscle during the contractions initiated with the knee extended increased compared
with the contractions initiated with the knee flexed (P ⫽
0.01). However, the increased activation amplitude of the
vastus medialis oblique muscle was not significant (P ⫽
0.18). For both groups of subjects the activation amplitude
of the vastus lateralis muscle was significantly larger
than that of the vastus medialis oblique muscle during the
contraction initiated with the knee extended (P ⫽ 0.02).
The muscle-by-group interaction term was not significant
(P ⫽ 0.31).
Subjects with patellofemoral pain had activation amplitude profiles that were functionally different from those of
the asymptomatic control subjects. For the contractions
initiated with the knee extended, the activation amplitudes of the vastus medialis oblique and vastus lateralis
muscles of the patellofemoral pain subjects were significantly greater than those of the control subjects (P ⬍ 0.01)
(Fig. 1).
DISCUSSION
The protocol used in the current study exploited the dependency on knee joint angle of the extent to which the
patella can normally move medially and laterally within
the intercondylar groove. We found that the activation
timing of the vastus medialis oblique and vastus lateralis
muscles of the subjects with patellofemoral pain was not
significantly different from that of control subjects. In
addition, we found that the activation amplitude patterns
for the subjects with patellofemoral pain were significantly different from those of control subjects.
We originally reasoned that the central nervous system
could adjust for the smaller contractile force and contractile velocity of the vastus medialis muscle by activating it
earlier than the vastus lateralis muscle. We also reasoned
that such an activation strategy would be evident to a
greater extent when contractions were initiated with the
knee joint extended, a position that increases the extent to
which the patella can normally move medially and laterally within the intercondylar groove. Computer simulation has suggested that activation of the vastus medialis
oblique muscle 5 ms earlier than the vastus lateralis muscle can reduce both the laterally directed motion of the
patella and the initial peak patellofemoral contact force.8
The results implied that causing the vastus lateralis muscle to be activated 5 ms before the vastus medialis oblique
muscle would increase lateralization of the patella and the
patellofemoral contact force in excess of that observed
when the muscles were activated simultaneously. The
results show that, for the open kinetic chain task used in
American Journal of Sports Medicine
the present study, these two muscles are activated simultaneously in both asymptomatic control subjects and subjects with patellofemoral joint pain.
The statistically significant between-group differences
with respect to activation amplitude are biomechanically
meaningful. First, for the control subjects, the vastus medialis and vastus lateralis muscle activation levels were
generally diminished during contractions initiated with
the knee joint extended compared with those when the
knee joint was flexed, reflecting the differences in the way
the central nervous system activates motor units during
concentric and eccentric contractions.3 In contrast, the
activation levels of patellofemoral pain subjects were generally increased. In particular, compared with the activation level during the contraction initiated with the knee
flexed, the increase in vastus lateralis muscle activation
level, but not vastus medialis, was significant for the
patellofemoral pain subjects. The functional importance of
these differences is that, after the onset of motion of the
dynamometer, the contraction initiated with the knee extended was eccentric. The significant increase in vastus
lateralis muscle activation and the associated increase in
contraction force were not paralleled by an increase in
vastus medialis muscle activation. This finding raises the
testable hypothesis that the subsequent eccentric contraction conditions are accompanied by increased lateralization of the patella and increased patellofemoral pressures
and forces. The experimental approach used in the
present study in conjunction with a computational model
will be required to truly test this hypothesis in vivo.
Consistent with our findings, the latency between the
vastus medialis oblique and vastus lateralis muscle activation during voluntary concentric contractions by control
subjects and patellofemoral pain subjects has previously
been reported as not significantly different.5 However, the
reflex activation of the vastus medialis oblique muscle
occurs significantly later than that of the vastus lateralis
in subjects with patellofemoral pain.12, 15 This contrast
suggests that the activation timing during reflex and voluntary contractions may not be related in a functionally
intuitive manner.
The differences we observed in activation amplitude
between the control subjects and those with patellofemoral pain are difficult to directly compare with the findings
of others. The difficulty arises because our method of EMG
analysis is different from those used by other authors. The
analyses of EMG data reported by others during dynamic
muscle contractions have not generally been independent
of the effects of muscle lengthening and shortening, for
example, during level gait, stair climbing, and ramp walking.9, 11 Our analysis of the initial 100 ms of activation, a
period during which the contraction is essentially isometric, precludes influence by afferent activity related to muscle length changes. However, we can make one betweenstudy comparison. In asymptomatic control subjects in the
present study, the normalized activation amplitude of the
vastus medialis oblique muscle was, on average, about
14% smaller than that of the vastus lateralis muscle. This
between-muscle difference is similar to the 11% difference
previously reported for uninjured subjects.2 Thus, by our
Vol. 30, No. 4, 2002
methods, the vastus lateralis muscle of asymptomatic control subjects consistently has a larger initial activation
amplitude than does the vastus medialis muscle.
In other studies in which isolated knee extension was
performed on a dynamometer, analysis of the EMG data is
often reported as an activation ratio of vastus medialis
oblique:vastus lateralis muscle.1, 6 In the present study,
the ratios for the control and patellofemoral pain subjects
were 0.82 and 0.77, respectively. These values are quite
different from the ratios reported by Laprade et al.,6
Sheehy et al.,11 and Boucher et al.,1 which ranged from 1.3
to 2.0 and 1.1 to 2.25 for control subjects and patellofemoral pain subjects, respectively. These differences reflect, in
part, the very different contraction conditions during
which the EMG data were collected and the analysis window used to quantify the signal. However, we believe that
the use of activation ratios, which presume a linear force:
EMG relationship for each muscle, can be problematic if
the relationships demonstrate nonlinearity.
We recognize that a number of limitations were present
in our experimental design. First, the subjects in the
patellofemoral pain group were a convenience sample,
recruited using criteria of activity-related patellofemoral
pain but without a history of knee surgery. The subjects’
symptoms of patellofemoral pain varied both in intensity
and duration, and thus, in some respects, were heterogeneous. Nevertheless, as a group, these subjects were associated with a pattern of activation amplitude that was
different from that of control subjects. Another potential
limitation of the study relates to the different isokinetic
speeds used to test the control subjects (60 deg/sec) and
the patellofemoral pain subjects (15 deg/sec). The slow
speed was selected for the patellofemoral pain subjects in
an effort to reduce any pain or discomfort that they may
have experienced or anticipated. However, because all of
our measurements were acquired before the onset of dynamometer motion and because we were not interested in
the muscle strength differences between the groups of
subjects, the speed and strength disparities were not considered to have been experimental confounds. In retrospect, it would have been valuable to make bilateral measurements, especially on the patellofemoral pain subjects.
The absence of bilateral measurements is not actually a
limitation of the study. However, bilateral measurements
would have allowed discrimination of a localized unilateral outcome associated with the pathologic condition
from that of a general influence.
The activation levels of the vastus medialis and vastus
lateralis muscles of the subjects with patellofemoral pain
were consistent with a laterally tracking patella. The unexpectedly large activation amplitudes of both muscles in
the patellofemoral pain subjects compared with the control subjects are notable. This finding provides evidence
that patellofemoral pain is associated with a disruption in
Motor Control in Subjects with Patellofemoral Pain
487
the control of the knee extensor muscles during the initial
activation of eccentric contractions. Although it is not
possible to ascribe a causal effect of the altered activation
timing and amplitude to the presence of patellofemoral
pain, these findings raise follow-up questions. Are the
altered activation patterns sensitive to rehabilitation? Do
subjective reports of knee joint pain and function parallel
changes in the activation patterns as a result of
rehabilitation?
ACKNOWLEDGMENTS
The authors gratefully acknowledge funding for the study
from the Chattanooga Corporation, Chattanooga, Tennessee, and Bauerfiend USA, Corporation, Atlanta, Georgia.
The authors also thank Michele E. George and Thomas M.
Lundin for their assistance in the data collection and
analysis.
REFERENCES
1. Boucher JP, King MA, Lefebvre R, et al: Quadriceps femoris muscle
activity in patellofemoral pain syndrome. Am J Sports Med 20: 527–532,
1992
2. Grabiner MD: Maximum rate of force development is increased by antagonist conditioning contraction. J Appl Physiol 77: 807– 811, 1994
3. Grabiner MD, Owings TM: Activation differences between concentric and
eccentric maximum voluntary contractions are evident prior to movement
onset. Exp Brain Res, in press, 2002
4. Johnson MA, Polgar J, Weightman D, et al: Data on the distribution of fibre
types in thirty-six human muscles. An autopsy study. J Neurol Sci 18:
111–129, 1973
5. Karst GM, Willett GM: Onset timing of electromyographic activity in the
vastus medialis oblique and vastus lateralis muscles in subjects with and
without patellofemoral pain syndrome. Phys Ther 75: 813– 823, 1995
6. Laprade J, Culham E, Brouwer B: Comparison of five isometric exercises
in the recruitment of the vastus medialis oblique in persons with and
without patellofemoral pain syndrome. J Orthop Sports Phys Ther 27:
197–204, 1998
7. Lieb FJ, Perry J: Quadriceps function. An anatomical and mechanical
study using amputated limbs. J Bone Joint Surg 50A: 1535–1548, 1968
8. Neptune RR, Wright IC, van den Bogert AJ: The influence of orthotic
devices and vastus medialis strength and timing on patellofemoral loads
during running. Clin Biomech 15: 611– 618, 2000
9. Powers CM, Landel R, Perry J: Timing and intensity of vastus muscle
activity during functional activities in subjects with and without patellofemoral pain. Phys Ther 76: 946 –967, 1996
10. Sanchis-Alfonso V, Roselló-Sastre E, Monteagudo-Castro C, et al: Quantitative analysis of nerve changes in the lateral retinaculum in patients with
isolated symptomatic patellofemoral malalignment: A preliminary study.
Am J Sports Med 26: 703–709, 1998
11. Sheehy P, Burdett RG, Irrgang JJ, et al: An electromyographic study of
vastus medialis oblique and vastus lateralis activity while ascending and
descending steps. J Orthop Sports Phys Ther 27: 423– 429, 1998
12. Voight ML, Wieder DL: Comparative reflex response times of vastus
medialis obliquus and vastus lateralis in normal subjects and subjects with
extensor mechanism dysfunction: An electromyographic study.
Am J Sports Med 19: 131–137, 1991
13. Wickiewicz TL, Roy RR, Powell PL, et al: Muscle architecture of the
human lower limb. Clin Orthop 179: 275–283, 1983
14. Witvrouw E, Lysens R, Bellemans J, et al: Open versus closed kinetic
chain exercises for patellofemoral pain: A prospective, randomized study.
Am J Sports Med 28: 687– 694, 2000
15. Witvrouw E, Sneyers C, Lysens R, et al: Reflex response times of vastus
medialis oblique and vastus lateralis in normal subjects and in subjects
with patellofemoral pain syndrome. J Orthop Sports Phys Ther 24: 160 –
165, 1996