Journal Logo

CLINICAL SCIENCES

Muscle Functional MRI to Evaluate Quadriceps Dysfunction in Patellofemoral Pain

PATTYN, ELS1; VERDONK, PETER2; STEYAERT, ADELHEID3; VAN TIGGELEN, DAMIEN1; WITVROUW, ERIK4

Author Information
Medicine & Science in Sports & Exercise: June 2013 - Volume 45 - Issue 6 - p 1023-1029
doi: 10.1249/MSS.0b013e318282672c
  • Free

Abstract

Patellofemoral pain (PFP) is one of the most common knee disorders affecting young and physically active adults as well as the general population (11,12,27). Despite the high incidence, the exact etiology of this disorder remains unclear, although evidence reveals a multifactorial origin. A deficiency in the physiology and function of the quadriceps muscle group has often been suggested to play an important role in the pathophysiology of PFP (21,28,34).

Patients with PFP seem to differ from healthy persons in several aspects of the main knee extensor muscles. For instance, a strength deficit of the quadriceps, presented as various patterns of muscular weakness, seems to be common in patients with PFP (4,9,33). Several authors also demonstrated a decreased quadriceps muscle mass in patients with PFP (3,17) and, in particular, atrophy of the vastus medialis obliquus (VMO) (16,26). Besides the strength deficit and atrophy, there is also evidence for a neuromuscular timing dysfunction of the quadriceps. In some patients with PFP, the vastus lateralis (VL) contracts earlier than the VMO, contributing to a laterally directed force on the patella and probably to abnormal patellar tracking (10,35). An imbalance in activity between VL and VMO or altered quadriceps activity has also been reported in persons with PFP (23,29,32). In literature and in clinical practice, restoration of quadriceps strength and function has been demonstrated to be imperative for a successful recovery from patellofemoral symptoms. Treatment generally includes exercise interventions with a view to strengthening and optimizing the use of the quadriceps and its components.

Quadriceps timing and amplitude measurements, usually performed by the means of EMG, provide an indication of the recruitment of the quadriceps muscle group during motor tasks. Muscle functional magnetic resonance imaging (mfMRI) is an innovative tool for assessing the extent of muscle activation following the performance of an exercise. The method relies on an acute increase of the transverse relaxation time (T2) of muscle water because of activity, resulting from underlying metabolic reactions (22,25). MfMRI quantifies shifts in T2 values on exercise, so-called T2 shift, which relate to the amount of work performed by the muscle. Compared to EMG, this technique has the potential to map spatial variations in activity within a muscle, which enables study of isolated activity of an entire muscle of interest, even of deep or underlying muscles, during a task (18,22,25). The recruitment patterns within several skeletal muscles like quadriceps (1,30), gastrocnemius (18), or lumbar muscles (14) have been extensively evaluated by fMRI in confined samples of healthy individuals.

Besides muscle function during specific exercise, T2 shift measurements are also a valuable tool to assess changes in activity patterns in patients with musculoskeletal disorders. Cagnie et al. (7) investigated the cervical flexor muscles activity in patients with whiplash-associated disorder. Another study evaluated the cervical extensor muscles in patients with mechanical neck pain (24).

Although physical therapy commonly intends to enhance and optimize the activation of the quadriceps muscle group, based on the assumption that the VMO and/or the quadriceps are less activated in patients with PFP, clear consensus is lacking regarding the actual recruitment pattern of the quadriceps in patients with PFP. Therefore, the aim of this study was to examine by mfMRI if PFP patients actually exhibit an altered activation of the quadriceps muscles that play a significant role in the dynamic balance of the patella.

METHODS

Participants.

Forty-six patients, diagnosed with PFP by an experienced orthopedic surgeon of the Ghent University Hospital, participated in this study. The patients were screened through physical examination and included if they exhibited two or more of the following criteria: tenderness on palpation of the posterior edge at the medial and/or lateral border of the patella, pain on direct compression of the patella against the femoral condyles with the knee in full extension, pain on direct compression of the patella against the femur during isometric quadriceps contraction with the knee in slight flexion, and pain on resisted knee extension (15). The screening process also included a functional assessment. The patients were enrolled in the study if they reported peripatellar or retropatellar pain provoked by at least two of the activities commonly associated with PFP (prolonged sitting with flexed knees, stair climbing, squatting, running, kneeling, and jumping) (2). The patients were required to have at least a 3-month history of PFP. The exclusion criteria for the patients consisted of 1) history or evidence of other knee disorders like patellar tendinopathy, ligament injury, bursitis, Osgood–Schlatter disease, meniscal injury, or osteoarthritis; 2) lower limb surgery or trauma within 1 yr; and 3) patellar instability. The patients were also screened for MRI contraindications such as pregnancy, claustrophobia, and implanted devices. At the intake of the study, the mean Kujala (20) score was 70.96, indicating reduced function. The mean duration of PFP symptoms was 17 months.

A control group consisted of 30 individuals without any history of knee joint disorder, comparable for age, sex, body mass index, and level of physical activity (Table 1). The activity level was measured by the Baecke questionnaire (3).

TABLE 1
TABLE 1:
Demographic data for study participants.

Twenty-six of a total of 46 PFP patients reported bilateral knee pain. Only the most severely affected leg was used for the analysis. Twenty-nine dominant and 17 nondominant sides were tested in the patient group. The corresponding dominant or nondominant leg was randomly determined for the healthy control group. As a consequence, data of 46 injured legs and 30 corresponding healthy legs were considered for processing. All participants were thoroughly notified of the study procedures, and written informed consent was obtained. The experimental protocol has been approved by the local Ethics Committee of the Ghent University Hospital.

General experimental design.

A mfMRI measurement requires images that are obtained at rest (preexercise image) and images immediately after a specific exercise (postexercise image) (8). The first MRI scan was performed after 30 min of supine relaxation. The scan at rest was followed by a squat exercise in an adjacent room. The participants had to bend through their knees from an upright position until 90° flexion at the defined speed of a metronome (92 bpm). The range of motion of the squat exercise was controlled by using a bar adjusted to each participant’s pelvis height at the moment that the knee angle achieved 90° flexion. The exercise had to be carried out until exhaustion. The duration of the squat exercise was recorded. Immediately on completion of the exercise, the participants reentered the scanner for a second scan. The time interval between the completion of the exercise and the initiation of scanning was 100.4 ± 22.5 s, which corresponds to the durations reported in other studies (19,30).

mfMRI.

Images were obtained using a 3-T Siemens Trio Tim scanner (Siemens, Erlangen, Germany). The participants were placed in the magnetic bore in a comfortable and relaxed supine position with straight legs. Two body matrix coils on both legs, which restrained any limb movement, were combined with a spine coil underneath as a receiver coil combination. Two plastic tubes filled with water were fixed to the thigh and aligned with marks on the spine coil (26). They indicated where the images should be taken and allowed for similar positioning in the magnet bore over repeated scans. One axial slice was taken at the middle of the distance from the spina iliaca anterior superior to the superior border of the patella (Fig. 1). The second slice was located at 2 cm proximal to the superior border of the patella to measure the VMO (Fig. 1) (26). A CPMG (Carr Purcell Meiboom Gill) sequence was applied to measure the T2 relaxation times. The following sequence parameters were used: repetition time, 3000 ms; echo train of 16 equidistant echos ranging from 10.1 to 161.6 ms; field of view, 380 mm; voxel size, 2 × 2 × 5.0 mm; slice thickness, 5.0 mm; total acquisition time, 9 min 41 s. Imaging procedures were identical for the rest scan and the scan after exercise.

FIGURE 1
FIGURE 1:
Illustration of aT2 map at 2 cm proximal to the superior border of the patella (A) and at midthigh level (B). Underneath for both levels, a schematic illustration with the outline of the investigated muscles (C and D): vastus lateralis (VL), vastus medialis obliquus (VMO), biceps femoris (BF), semimembranosus (SM), sartorius (S), rectus femoris (RF), vastus medialis and intermedius (VMVI), and hamstrings and adductors (HM + AD).

Data management.

MR images were analyzed using ImageJ (Java-based version of the public domain NIH Image Software; Research Services Branch, National Institutes of Health). A T2 value (ms) per voxel was calculated out of 16 echos for all images. Subsequently, the regions of interest were manually outlined on the T2 maps for following muscles: VMO, VL, and vastus medialis and intermedius (VMVI). The VMVI were combined in a single measurement because they could not be fully distinguished from each other at this level. The inclusion of vascular and nerve bundles and visual fat was avoided. Finally, the mean T2 relaxation time (ms) was calculated for each region of interest. The reproducibility of T2 measurements using the T2 procedure has been reported previously (6,13).

Statistical analysis.

Descriptive statistics (mean ± SD) were obtained for the T2 relaxation time (ms) and the shift in T2 relaxation time (ms) because of activity. The T2 shift was defined as the difference between the T2 value after exercise and the T2 value at rest.

All the following analyses were performed with stratification by sex because unpaired Student’s t-tests showed a slight difference in T2 values between male and female. The differences in T2 relaxation times and the T2 shift between the PFP and the control groups were analyzed by means of unpaired Student’s t-test because the values were normally distributed.

The T2 shift difference between the individual muscles was analyzed and compared in both groups by using a two-way (patient or control and muscle) ANOVA with repeated measures.

Pearson correlation coefficients were used to evaluate correlations between T2 shifts, duration of squat exercise, and duration of symptoms (PFP group). All statistical analyses were performed using SPSS 20.0 for Windows (SPSS, Chicago, IL). The statistical significance level was set at P < 0.05.

RESULTS

The T2 values at rest and after exercise calculated for all muscles in the control and patient groups are shown in Table 2. T2 after exercise was significantly higher than T2 at rest for all muscles, for both groups and for both sexes. There were no significant differences in the T2 values at rest and the T2 shift values between patient and control groups, except for the T2 rest value of the VMVI of females (P = 0.007; Table 3).

TABLE 2
TABLE 2:
T2 values (ms) for quadriceps components in PFP subjects and healthy controls.
TABLE 3
TABLE 3:
T2 rest and T2 shift (ms) for quadriceps components in PFP subjects compared to healthy controls.

Mean T2 shifts plotted by muscle and group are shown in Figure 2. In the overall statistical model for T2 shift, there was a (borderline) significant main effect for muscle (male P = 0.053 and female P = 0.049) but not for group (male P = 0.709 and female P = 0.380). The muscle–group interaction effect was not significant (male P = 0.334 and female P = 0.598). The T2 shift of the VL was significantly smaller than the T2 shift of the VMVI in both study groups (male P < 0.001 and female P = 0.044), while in females, the T2 shift of the VMO was also significantly smaller than the T2 shift of the VMVI (P = 0.027).

FIGURE 2
FIGURE 2:
T2 shifts in milliseconds (mean ± SD) of the vastus medialis obliquus (VMO), vastus medialis and intermedius (VMVI), and the vastus lateralis (VL) for the group with patellofemoral pain (PFP) and the control group in (A) males and (B) females. *P < 0.05.

No significant correlation was found in the patient group between duration of symptoms and T2 shifts in all muscles (male r = 0.237–0.362 and female r = 0.219–0.388). For the male subjects, there were no significant correlations between the observed T2 shifts and duration of squat exercise (r = 0.079–0.143). For the female subjects, there were significant but moderate correlations between the T2 shifts and duration of squat exercise (r = 0.319–0.502).

DISCUSSION

The purpose of this study was to investigate with mfMRI if patients with PFP actually exhibit an altered recruitment pattern within the quadriceps muscle group. The main results indicated that there is no difference in activation pattern of the quadriceps components that play a significant role in the dynamic balance of the patella between patients with patellofemoral problems and healthy controls for both male and female subjects. This is striking as a deficiency in the function of the VMO, as well as the entire quadriceps, has often been suggested to play an important role in the pathophysiology of PFP (21,28,34). The results of this study, however, do not support the hypothesis of quadriceps dysfunction in PFP patients. It was demonstrated by means of the innovative technique of mfMRI that PFP patients exhibit equal activation of the quadriceps muscles during a functional and weight bearing activity.

There is no clear consensus in the literature regarding the recruitment pattern of the quadriceps in patients with PFP. The studies that investigated vasti muscles’ activity in PFP patients in comparison to healthy controls during weight bearing activities, applying EMG measurements, presented conflicting results. Powers et al. (29) demonstrated that subjects with PFP exhibited less activity of all the vasti muscles during level walking and ramp walking than did subjects without PFP. According to the authors, this decreased activity may be suggestive of a quadriceps femoris muscle avoidance pattern. Other EMG studies reported greater activity of the VM (5) and of the VMO and VL (23) compared to healthy controls during the loading response and single-leg stance intervals of stair descent (5) and the most demanding phases of the gait cycle, respectively (23). Mohr et al. (23) stated that these data reflected a generalized quadriceps muscle weakness in patients with PFP. One study found no differences between the control and the patient group for the peak VMO:VL EMG ratios while ascending and descending steps (31).

In our study, the patients’ quadriceps muscles were not less or more active during the squat exercise compared to the healthy controls. This may indicate that the participants with PFP did not reduce quadriceps activity by applying compensatory strategies to decrease patellofemoral compression.

The results of this study also indicated that there was a significant difference in T2 shift between the mutual muscles. For both males and females, the T2 shift of the VL was significantly smaller than the T2 shift of the VMVI. This means that, during the squat exercise, the VMVI was more active than the VL. In addition, the VMVI was also more recruited in relation to the VMO in females. However, these differences were equal in the patient group as well as in the healthy control group. This may reflect the similar relative contribution of the quadriceps muscles in PFP patients and control subjects during a functional activity.

The most novel findings of this study are that the relative contribution of the quadriceps muscles to a functional activity seems not to be modified in patients with PFP. The study results may yield important consequences for the current guidelines to therapy. There is general agreement that restoration of quadriceps strength and function is a must for a successful recovery from patellofemoral symptoms. Exercise interventions are aimed at strengthening and reactivating the quadriceps and, in particular, the VMO. The patients are taught how to use the VMO more during functional tasks. The underlying thought for this approach is that the VMO is less activated in functional tasks because of weakness or a neuromuscular deficit in relation to the VL contributing to patellar maltracking and knee pain. However, the present study indicated that there is no point to teach patients to activate the VMO more during functional exercises because patients do not use the VMO less than healthy individuals do. As such, the results do not support the “myth” of contracting the VMO more during functional activities, which is an important message for physical therapists, physicians, and sport coaches. However, some studies have demonstrated that patients with PFP exhibit atrophy of the quadriceps (9,17) and, in particular, atrophy of the VMO (16,26). As a consequence, general quadriceps strengthening as well as specific strengthening exercises for the VMO aiming at muscle volume gain seems imperative if atrophy is present. Because of the eccentric and the neuromuscular problems in PFP, functional eccentric training to increase neuromuscular control of all related muscles seems essential for all patients with PFP.

In contrast with other studies investigating muscle recruitment patterns in PFP patients in comparison to healthy controls, we used mfMRI instead of EMG. The most notable feature of mfMRI is the potential to map spatial variations in activity within a muscle, which enables study of isolated activity of an entire muscle of interest, even of deep, adjacent, or overlying muscles, during a task (25). EMG is not well suited for simultaneous measurements from many muscles, some of which may lie deep under the skin (22). It is also difficult to detect with surface EMG muscle activity over large regions or in regions deep within the muscle or to detect the activity of an entire muscle of interest (18). In addition, mfMRI overcomes the limitations of EMG by the elimination of signal crosstalk and the avoidance of signal issues attributed to impedance from subcutaneous tissue and electrode type and placement. The mfMRI technique allowed in the current study to evaluate the relative recruitment of different muscles during a functional exercise until exhaustion, which reflected a prolonged activity of daily life. In contrast, previous studies recorded the mean intensity of EMG activity or peak EMG ratios of different muscles during two to five trials of a functional movement.

The present results must be viewed within the limitations of the study. In accordance with other studies, there is considerable variability in the activity-dependent T2 response across subjects. Individuals may differ significantly in their baseline metabolic capacity. Therefore, a larger increase in postexercise T2 observed in one subject cannot always be interpreted as more effective recruitment relative to another subject with a lower T2 shift. Accordingly, the use of T2 mapping to compare the activation strategies and intensity of recruitment between subjects remains controversial.

Our study reported higher T2 values compared with other studies. Different imaging sequences will be probably the main reason for this discrepancy. The current T2 values are of equal size as the values found in studies evaluating cervical and lumbar muscles with the same apparatus and imaging sequences (7,13).

Because females are more likely to develop PFP, it might be valuable to investigate potential differences in T2 values between the different phases of the menstrual cycle. Another interesting issue would be the inclusion of the hamstring and adductor muscles into the T2 analysis to gain insight into the general recruitment pattern of the squat of both patient and control group. Further detailed fundamental research will be needed to elucidate the quadriceps activation pattern of patients with PFP to clarify its contribution to the etiology of PFP and to estimate the implications for the treatment.

In conclusion, this study is the first to examine the recruitment pattern of the muscles that play a significant role in the dynamic balance of the patella in PFP patients by means of mfMRI. Patients with PFP did not exhibit an altered recruitment pattern within the quadriceps muscle compared to healthy controls. The vasti muscles were not equally activated during the squat exercise, but this difference in T2 shift was the same for the patients as well as the control group. Further research is needed to clarify the muscle recruitment pattern of the quadriceps and its clinical implications for PFP.

No outside funding or grants have been received that assisted the study.

No conflicts of interests are reported.

The results of the present study do not constitute endorsement by the American College of Sports Medicine.

REFERENCES

1. Akima H, Takahashi H, Kuno SY, Katsuta S. Coactivation pattern in human quadriceps during isokinetic knee–extension by muscle functional MRI. Eur J Appl Physiol. 2004; 91 (1): 7–14.
2. Arroll B, Ellis-Pegler E, Edwards A, Sutcliffe G. Patellofemoral pain syndrome. A critical review of the clinical trials on nonoperative therapy. Am J Sports Med. 1997; 25 (2): 207–12.
3. Baecke JA, Burema J, Frijters JE. A short questionnaire for the measurement of habitual physical activity in epidemiological studies. Am J Clin Nutr. 1982; 36 (5): 936–42.
4. Bennett JG, Stauber WT. Evaluation and treatment of anterior knee pain using eccentric exercise. Med Sci Sports Exerc. 1986; 18 (5): 526–30.
5. Bolgla LA, Malone TR, Umberger BR, Uhl TL. Comparison of hip and knee strength and neuromuscular activity in subjects with and without patellofemoral pain syndrome. Int J Sports Phys Ther. 2011; 6 (4): 285–96.
6. Cagnie B, Dickx N, Peeters I, et al. The use of functional MRI to evaluate cervical flexor activity during different cervical flexion exercises. J Appl Physiol. 2008; 104 (1): 230–5.
7. Cagnie B, Dolphens M, Peeters I, Achten E, Cambier D, Danneels L. Use of muscle functional magnetic resonance imaging to compare cervical flexor activity between patients with whiplash-associated disorders and people who are healthy. Phys Ther. 2010; 90 (8): 1157–64.
8. Cagnie B, Elliott JM, O’Leary S, D’hooge R, Dickx N, Danneels LA. Muscle functional MRI as an imaging tool to evaluate muscle activity. J Orthop Sports Phys Ther. 2011; 41 (11): 896–903.
9. Callaghan MJ, Oldham JA. Quadriceps atrophy: to what extent does it exist in patellofemoral pain syndrome? Br J Sports Med. 2004; 38 (3): 295–9.
10. Cowan SM, Bennell KL, Hodges PW, Crossley KM, McConnell J. Delayed onset of electromyographic activity of vastus medialis obliquus relative to vastus lateralis in subjects with patellofemoral pain syndrome. Arch Phys Med Rehabil. 2001; 82 (2): 183–9.
11. DeHaven KE, Lintner DM. Athletic injuries: comparison by age, sport, and gender. Am J Sports Med. 1986; 14 (3): 218–24.
12. Devereaux MD, Lachmann SM. Patello-femoral arthralgia in athletes attending a sports injury clinic. Br J Sports Med. 1984; 18 (1): 18–21.
13. Dickx N, Cagnie B, Achten E, Vandemaele P, Parlevliet T, Danneels L. Changes in lumbar muscle activity because of induced muscle pain evaluated by muscle functional magnetic resonance imaging. Spine (Phila Pa 1976). 2008; 33 (26): E983–9.
14. Dickx N, D’hooge R, Cagnie B, Deschepper E, Verstraete K, Danneels L. Magnetic resonance imaging and electromyography to measure lumbar back muscle activity. Spine (Phila Pa 1976). 2010; 35 (17): E836–42.
15. Insall J. Current Concepts Review: patellar pain. J Bone Joint Surg Am. 1982; 64 (1): 147–52.
16. Jan MH, Lin DH, Lin JJ, Lin CH, Cheng CK, Lin YF. Differences in sonographic characteristics of the vastus medialis obliquus between patients with patellofemoral pain syndrome and healthy adults. Am J Sports Med. 2009; 37 (9): 1743–9.
17. Kaya D, Citaker S, Kerimoglu U, et al. Women with patellofemoral pain syndrome have quadriceps femoris volume and strength deficiency. Knee Surg Sports Traumatol Arthrosc. 2011; 19 (2): 242–7.
18. Kinugasa R, Kawakami Y, Fukunaga T. Quantitative assessment of skeletal muscle activation using muscle functional MRI. Magn Reson Imaging. 2006; 24 (5): 639–44.
19. Kinugasa R, Yoshida K, Horii A. The effects of ice application over the vastus medialis on the activity of quadriceps femoris assessed by muscle function magnetic resonance imaging. J Sports Med Phys Fitness. 2005; 45 (3): 360–4.
20. Kujala UM, Jaakkola LH, Koskinen SK, Taimela S, Hurme M, Nelimarkka O. Scoring of patellofemoral disorders. Arthroscopy. 1993; 9 (2): 159–63.
21. Lin F, Wilson NA, Makhsous M, et al. In vivo patellar tracking induced by individual quadriceps components in individuals with patellofemoral pain. J Biomech. 2010; 43 (2): 235–41.
22. Meyer RA, Prior BM. Functional magnetic resonance imaging of muscle. Exerc Sport Sci Rev. 2000; 28 (2): 89–92.
23. Mohr KJ, Kvitne RS, Pink MM, Fideler B, Perry J. Electromyography of the quadriceps in patellofemoral pain with patellar subluxation. Clin Orthop Relat Res. 2003; (415): 261–71.
24. O’Leary S, Cagnie B, Reeve A, Jull G, Elliott JM. Is there altered activity of the extensor muscles in chronic mechanical neck pain? A functional magnetic resonance imaging study. Arch Phys Med Rehabil. 2011; 92 (6): 929–34.
25. Patten C, Meyer RA, Fleckenstein JL. T2 mapping of muscle. Semin Musculoskelet Radiol. 2003; 7 (4): 297–305.
26. Pattyn E, Verdonk P, Steyaert A, et al. Vastus medialis obliquus atrophy: does it exist in patellofemoral pain syndrome? Am J Sports Med. 2011; 39 (7): 1450–5.
27. Powers CM. Rehabilitation of patellofemoral joint disorders: a critical review. J Orthop Sports Phys Ther. 1998; 28 (5): 345–54.
28. Powers CM. Patellar kinematics, part I: the influence of vastus muscle activity in subjects with and without patellofemoral pain. Phys Ther. 2000; 80 (10): 956–64.
29. 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. 1996; 76 (9): 946–55.
30. Prior BM, Jayaraman RC, Reid RW, et al. Biarticular and monoarticular muscle activation and injury in human quadriceps muscle. Eur J Appl Physiol. 2001; 85 (1–2): 185–90.
31. Sheehy P, Burdett RG, Irrgang JJ, VanSwearingen J. An electromyographic study of vastus medialis oblique and vastus lateralis activity while ascending and descending steps. J Orthop Sports Phys Ther. 1998; 27 (6): 423–9.
32. Souza DR, Gross MT. Comparison of vastus medialis obliquus: vastus lateralis muscle integrated electromyographic ratios between healthy subjects and patients with patellofemoral pain. Phys Ther. 1991; 71 (4): 310–6.
33. Werner S. An evaluation of knee extensor and knee flexor torques and EMGs in patients with patellofemoral pain syndrome in comparison with matched controls. Knee Surg Sports Traumatol Arthrosc. 1995; 3 (2): 89–94.
34. Witvrouw E, Lysens R, Bellemans J, Cambier D, Vanderstraeten G. Intrinsic risk factors for the development of anterior knee pain in an athletic population. A two-year prospective study. Am J Sports Med. 2000; 28 (4): 480–9.
35. Witvrouw E, Sneyers C, Lysens R, Victor J, Bellemans J. 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. 1996; 24 (3): 160–5.
Keywords:

PATELLOFEMORAL; MAGNETIC RESONANCE IMAGING; QUADRICEPS; ACTIVATION PATTERN

©2013The American College of Sports Medicine