INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

THE PRIMARY FUNCTION OF THE quadriceps femoris (QF) muscle is to generate knee extensor torque and to stabilize the patella (29). As this muscle group plays a key role during ambulatory and functional activities, interest in the recruitment characteristics of its different components has escalated. One method that has been used to enhance the understanding of QF muscle function is surface electromyography (EMG), which has proven to be a valid and reliable tool of muscle recruitment (21, 36). Many anatomic and physiological factors that mediate QF recruitment behavior, such as contraction intensity, muscle fiber type, or motor unit activation pattern, can be manifested in the surface EMG signal. The QF muscle is unique in that the different components have been shown to display varying fiber-type compositions (20, 40) and divergent tension (10), while all receiving innervation from the femoral nerve (18). Differential recruitment of the superficial components of the QF muscle has also been recently demonstrated to be a function of contraction intensity (1, 31). Furthermore, potential gender differences that may exist in the surface EMG signal have yet to be conclusively established for this muscle group, despite demonstrated differences in fiber-type proportions (35).

The expression and quantification of the surface EMG signal in the frequency domain, via the median frequency (f_med), has displayed a positive relationship to muscle fiber conduction velocity (23, 25, 37) as a function of a larger fiber diameter (23). Defined as the frequency at which the EMG spectrum is divided into two parts of equal power (4), the f_med has been shown to increase with higher contraction intensities as a result of additional recruitment of larger diameter fibers (4, 6). As limited evidence regarding typical muscle fiber-type proportions of the vastus medialis (VM) and rectus femoris (RF) muscles exists (20, 40), compared with that for the vastus lateralis (VL) (35), the f_med may provide insight into specific EMG patterns for these different muscles. Typically, shifts to lower frequencies have been demonstrated during fatiguing contractions (2, 24, 25), yet a pattern of f_med changes at different contraction intensities for the different portions of the QF muscle has yet to be shown. The potential of a gender difference in f_med as a function of contraction intensity has been shown in select muscles (anconeus, tibialis anterior) by previous studies (4, 6). Such a finding was not evident, however, during maximal voluntary isometric contractions of the QF muscle, in which our laboratory demonstrated f_med of the superficial portions of the QF muscle that was not gender specific (32). The implications of varying recruitment patterns of the QF muscle between men and women, as a function of contraction intensity, may lend insight into gender-specific factors affecting physical activity and function, such as discrepant rates of athletic-related knee injury (14). Therefore, the purpose of this study was to examine the influence of gender and contraction intensity on the f_med statistic of the VM, VL, and RF muscles in healthy individuals.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Subjects. The present study incorporated a repeated-measures design. Subjects for this study consisted of 30 healthy male (n = 15, mean age = 26.5 ± 4.7 yr, mean height = 178.1 ± 6.1 cm, mean mass = 88.6 ± 16.1 kg) and female (n = 15, mean age = 24.9 ± 3.7 yr, mean height = 165.4 ± 5.6 cm, mean mass = 61.4 ± 8.4 kg) volunteers. All subjects were physically active but had not actively taken part in an intensive resistance training program (i.e., >2 days/wk) for the lower extremity for at least 6 mo before the study. Individuals with a history of cardiovascular disease, hypertension, or orthopedic pathology were excluded from participating in this study. All subjects provided written, informed consent as approved by the Institutional Review Board at Eastern Washington University.

Measurement of isometric torque. Before the measurement of isometric torque, all subjects completed a warm-up period that consisted of submaximal cycling for 3-5 min. Isometric torque was measured on the Biodex System II isokinetic dynamometer (Biodex Medical, Shirley, NY). Subjects sat in a comfortable, upright position on the Biodex accessory chair and were secured with the use of thigh, pelvic, and torso straps to minimize extraneous body movements. The lateral femoral epicondyle was used as the bony landmark for matching the axis of rotation of the knee joint with the axis of rotation of the dynamometer resistance adapter. Gravity correction was obtained by measuring the torque exerted on the dynamometer resistance adapter with the knee in a relaxed state at full extension. Values for isometric torque were automatically adjusted for gravity by the Biodex Advantage software program (version 3.2.6). During the assessment of isometric torque, subjects were required to fold their arms across their chest and were given verbal encouragement, as well as visual feedback from the Biodex computer monitor, in an attempt to achieve a maximal voluntary effort (16, 22, 27). All of the procedures and verbal encouragement were administered by the same investigator for all subjects. Calibration of the Biodex dynamometer was performed before every testing session according to the manufacturer's specifications.

Measurement of f_med. The f_med was assessed through surface EMG for the VM, VL, and RF muscles. Preamplified bipolar circular surface electrodes (Ag/AgCl, 0.8 cm diameter) were placed on each muscle with a fixed interelectrode distance (center to center) of 2 cm. Before electrode placement, the skin area was shaved, cleaned with isopropyl alcohol, and abraded with coarse gauze to reduce skin impedance. Electrode placement for the VM was 20% of the distance from the medial joint line of the knee to the anterior superior iliac spine (44). This electrode was placed at an ~45° angle between the anatomic horizontal and sagittal planes to be oriented toward the longitudinal direction of the muscle fibers. Electrode placement for the VL was the midpoint between the head of the greater trochanter and the lateral femoral epicondyle (20), and electrode placement for the RF was 50% of the distance from the anterior superior iliac spine to the superior pole of the patella (44). The reference electrode was placed over the medial shaft of the tibia ~6-8 cm below the inferior pole of the patella. EMG activity was collected by a four-channel unit (Therapeutics Unlimited, Iowa City, IA) at a rate of 1,000 Hz for each muscle. The common mode rejection of the current system is 87 dB at 60 Hz with an input impedance of >25 M at direct current. The gain range used in this study was 10,000, and signals were band-pass filtered between 20 and 500 Hz. Raw EMG signals were digitized and stored on computer disks for subsequent analysis by the Acknowledge software program (version 3.2.6, Biopac Systems, Santa Barbara, CA). The signals collected within the first and last second of each 5-s isometric contraction were not used for analysis because of knee movement that may have occurred at the initiation and completion of the test. Therefore, a 3-s window of EMG signals was used for analysis. A power spectral analysis was performed on the 3-s window for each muscle. A fast Fourier transformation of 512 points (Hamming window processing) was performed on 11 consecutive 512-ms segments, overlapping each other by one-half their length (256 ms), over the middle of each contraction (4). The f_med was determined from each of the 11 overlapping windows. The f_med mean and SD of the 11 windows during each contraction were then calculated for each muscle. These two values were then used for statistical analyses. The integration algorithm utilized to calculate the area of the power frequency spectrum in the software is as follows
where f is frequency, f(n) represents the data values (mV/s), t is the change in time and represents the horizontal sampling interval (s) (MP100 Systems Guide, Biopac), and k is a variable.

Statistical analysis. A three-factor repeated-measures ANOVA (gender × muscle × intensity) was performed on the f_med mean and variability. When the overall F test revealed an intensity or muscle main effect or significant interaction, separate ANOVAs were performed across the repeating levels of contraction intensity, and the Bonferroni-Dunn inequality was invoked to ensure the familywise error rate (12). To specifically examine the effects of contraction intensity and gender on changes in the f_med, separate two-way ANOVAs (intensity × gender) were performed on each muscle to compare the values obtained at each intensity with 10% MVC. This was performed to examine significant changes across the range of contraction intensities. All tests of significance were carried out to ensure a familywise error rate of 0.05.

	RESULTS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Descriptive data (means ± SD) for the f_med means and variability for the VM, VL, and RF muscles at each intensity for men and women are presented in Tables 1 and 2, respectively.

Muscle differences. The results demonstrated an overall significant main effect for muscle (F_1,28 = 63.68, P < 0.001, ² = 0.70, 1   = 0.99) as f_med was significantly highest for the VL and lowest for the VM, with the RF f_med values in between the other two muscles (i.e., greater than VM and less than VL). The results for f_med variability across the 11 overlapping epochs demonstrated a significant intensity main effect (F_8,224 = 1.98, P = 0.05, ² = 0.07, 1   = 0.81) and a significant muscle main effect (F_2,56 = 30.74, P < 0.001, ² = 0.52, 1   = 0.99) and no gender main effect or interactions. The results indicate that f_med variability was highest for the VL muscle compared with the VM (F_1,28 = 35.78, P < 0.001, ² = 0.56, 1   = 0.99) and RF (F_2,56 = 68.66, P < 0.001, ² = 0.71, 1   = 0.99) muscles. After adjustment for the a priori familywise -level (0.05/3 comparisons = 0.02), these findings were found to be significant. The VM and RF muscles were not significantly different.

Mean f_med gender by intensity interaction: VL muscle. A significant intensity main effect (F_8,224 = 29.12, P < 0.001, ² = 0.51, 1   = 0.99), as well as a significant intensity by gender interaction, was demonstrated for the VL muscle (F_8,224 = 2.61, P = 0.01, ² = 0.09, 1   = 0.92). These results indicate that a significantly greater increase in f_med occurred from 10% MVC up to 90% MVC in the men compared with the women (Fig. 1). Repeated two-factor interactions were specifically detected between 10% MVC and every level of contraction intensity (20-90% MVC), with the exception of 30% MVC, as f_med increased significantly more in the men than women (² range = 0.16-0.27).

View larger version (18K): [in this window] [in a new window]   Fig. 1.   Median frequency (fmed) of the vastus lateralis muscle in men and women from 10 to 90% maximal voluntary contraction (MVC). Values are means ± SE. Significant gender by intensity interaction.

Mean f_med gender by intensity interaction: VM and RF muscles. The gender by intensity ANOVA for the VM muscle demonstrated a significant interaction (F_8,224 = 2.94, P = 0.004, ² = 0.09, 1   = 0.95) between contraction intensities of 10% MVC and the following: 20, 30, 40, 50, and 70% MVC (Fig. 2). After adjustment for the a priori -level (0.05/8 comparisons = 0.006), this interaction was found to be significant only between 10 and 50% MVC (F_1,28 = 9.69, P = 0.004, ² = 0.26, 1   = 0.85). These results display a significant decrease in f_med for the VM muscle in women between these two contraction intensities, compared with a minimal change in the men. A significant trial main effect was observed for the RF muscle (F_8,224 = 2.05, P = 0.04, ² = 0.07, 1   = 0.83), with differences noted between contraction intensities at 30 and 40% MVC as f_med increased (F_1,28 = 5.98, P = 0.02, ² = 0.18, 1   = 0.66) and between intensities at 70 and 80% MVC as f_med decreased (F_1,28 = 4.18, P = 0.05, ² = 0.13, 1   = 0.51). However, after the Bonferroni-Dunn adjustment for the preset -level (0.05/8 comparisons = 0.006), these specific interactions were not found to be statistically significant.

View larger version (17K): [in this window] [in a new window]   Fig. 2.   The fmed of the vastus medialis muscle in men and women. Values are means ± SE. * Statistically significant decrease in fmed in women compared with men between 10% MVC and 20, 30, 40, 50, and 70% MVC, P < 0.05.

	DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

The major findings of this study demonstrated that the superficial components of the QF muscle display distinctive differences in f_med mean and variability across the spectrum of submaximal contraction intensities. The most notable distinction occurred with the f_med mean of the VL muscle, as men displayed a significantly greater increase from 10 to 90% MVC (average 18.6% increase) than women (average 7.6% increase). Although the effect size of this interaction is considered to be small (² = 0.09), the effect of contraction intensity on f_med when pooled across all subjects was larger (² = 0.51). Significantly higher f_med variability was shown for the VL compared with the RF and VM, with these latter two muscles displaying no significant differences. The significantly greater change in VM f_med mean (i.e., decrease) in the women compared with the men is quantified as a small effect size (² = 0.09; Fig. 1) and should, therefore, be interpreted with caution. In other words, the inherent variability of f_med within a contraction (Table 2) may be used as the "yardstick" in lieu of statistically significant results. Within this context, the demonstrated significant effects for the RF and VM muscles approximate to a 5- to 8-Hz and a 5- to 6-Hz difference, respectively. These statistical differences in mean f_med fall within the range of expected f_med variability for any given contraction (Table 2). In regard to f_med variability, it was assumed in the present investigation that a systematic change (i.e., a decrease due to fatigue) across the 11 epochs was not present because of the short nature of the contraction. The possibility of a systematic change, however, was not specifically investigated in this study.

Effects of contraction intensity. To reiterate the function of the QF muscle, its role as a generator of knee extensor torque is critical for normal activities of daily living. Moreover, the relative activation of its different components has been shown to alter patellofemoral mechanics significantly (10) and may ultimately impact the presence of various knee pathologies (13, 33). The examination of an EMG frequency-domain statistic (f_med) to describe intermuscular differences within the QF may provide additional evidence toward the understanding of this muscle's characteristics under voluntary conditions. This notion is deduced from experimental evidence regarding 1) the presence of a greater number and higher density of Na⁺ channels within the sarcolemma of fast-twitch muscle fibers, compared with slow-twitch muscle fibers (26); 2) the generation of higher frequency action potentials, and hence EMG signals, from fast-twitch muscle fibers (23); 3) the significant relation between the EMG f_med and muscle fiber conduction velocity (23, 37); and 4) the progressive recruitment of fast-twitch muscle fibers at higher contraction intensities (7, 9, 17). Although it is well known that the surface EMG signal, as measured in the present study, provides limited evidence of specific motor unit recruitment or rate-coding changes, the observed f_med statistic can be considered to be, in part, a reflection of muscle fiber-type activation (11, 23). The scientific literature, however, has not yielded absolutely consistent findings regarding f_med changes as a function of contraction intensity. Petrofsky and Lind (30) demonstrated no relationship between contraction intensity and mean power frequency (MPF) of the forearm muscles. Similarly, Hagberg and Ericson (15) found that the MPF of the elbow flexors increased only at low-contraction intensities and became independent of contraction intensities >30% MVC. Subsequently, however, Broman et al. (5) demonstrated a significant increase in f_med in the tibialis anterior muscle in eight healthy men during isometric contractions ranging from 10 to 100% MVC. It was not until the work of Moritani and Muro (28) that the raw data (i.e., frequency statistics) were presented to illustrate the magnitude of the contraction intensity effect. During a linearly increasing contraction (0-80% MVC) of the biceps brachii muscle, a significant increase in the MPF was observed in the 12 male subjects (MPF increase range = 89 ± 4.4 to 123 ± 7.8 Hz). It was also shown that the increase in MPF occurred concomitantly with greater motor unit spike amplitudes and firing frequencies, as detected by intramuscular recordings. To support the findings of Moritani and Muro regarding the tibialis anterior muscle, Cioni et al. (6) observed significant increases in f_med of this muscle in both healthy men (n = 15) and women (n = 15) across isometric contraction intensities ranging from 10 to 100% MVC. Based on the data presented by Cioni et al., it appeared that f_med increased from ~85 Hz at 10% MVC to 125 Hz at 100% MVC. Thus far, these observations suggest that sensitivity of f_med to changes in contraction intensity may largely be dictated by the muscle in question. For instance, Bilodeau et al. (4) found statistically significant increases in f_med in the anconeus and triceps brachii muscle isometric contractions that ranged from 10 to 90% MVC in 10% increments (n = 13 men, n = 16 women). However, the graphical representation of these data suggest only a moderate rise in f_med (i.e., ~10-15 Hz), with the exception of the anconeus muscle in the male subjects (~35 Hz). These discrepant findings within the scientific literature appear to be present in the present investigation. Whereas only small changes in mean f_med were observed across the contraction intensities for the VM and RF muscles, larger and consistent increases were only noted for the VL muscle. It can be reasonably speculated that contraction intensity differences for the VM and RF are attributable to the inherent nature of the EMG as a "quasi-random" signal (32) and to the observed variability within the present study (Table 2). As very limited data suggest that the VM muscle contains a relatively higher proportion of slow-twitch muscle fibers (20, 40), research with larger sample sizes demonstrates a wide variation in muscle fiber types of the VL muscle (36). As a result, increasing contraction intensities that progressively recruit fast-twitch fibers may give rise to higher frequency EMG signals if a high degree of muscle fiber-type heterogeneity were present.

Effects of QF muscle differences. Optimal function of the knee joint is highly dependent on the coordinated action within the QF complex. It is becoming ever more apparent that the components of the QF exhibit different cross-sectional areas (43), fiber angles (29, 33), muscle fiber types (20, 40), contributions to knee extensor torque (10, 29, 34), and activation characteristics (1, 31). The examination of the EMG signals from the present experimental procedure in the time domain (i.e., full-wave rectified, integrated, and normalized to the MVCs) demonstrated nonparallel increases in activation of these muscles at higher contraction intensities (31), which were recently supported by the findings of Alkner et al. (1). The overall higher f_med values for the VL, compared with the VM and RF muscles, concur with the findings recently published during the performance of the three MVCs of the present subjects (32). These results are, however, contrary to the findings of Weir et al. (42) during a sustained 50% MVC, Ebenbichler et al. (8) during three successive MVCs, and Gerdle et al. (11) during 25 and 75% MVCs, in which intermuscle differences were not present. It is noteworthy to point out that f_med values reported in previous studies (3, 8, 41) are typically lower (i.e., 68-105.7 Hz) than those in the present investigation. The methodology of the present investigation incorporated a 2-cm interelectrode distance, which is narrower than that used by others (3, 41), that may account for the observed higher values. This is due to the fact that a larger interelectrode spacing acts as a low-pass filter, thereby attenuating higher frequency signals and biasing the f_med toward a lower value. It may also be argued that a better representation of f_med was obtained in the present investigation as a result of the 11 overlapping epochs. The variability of f_med within a single contraction illustrates a relative fluctuation of power spectral density statistics, even for brief, submaximal, isometric, constant-force contractions. Such variability should be taken into consideration in cases in which these statistics are used to objectify the effects of various interventions, such as training or disease, on neuromuscular function. The f_med values that are comparable with the present investigation have been shown for the anconeus muscle (160-180 Hz) and the tibialis anterior muscle (75-125 Hz) (4, 6). The differences among the superficial QF portions in the present study may be reflective of existing variations in muscle fiber types among these muscles.

Gender differences. The examination of a gender difference in muscle characteristics has partially been fueled by recent data demonstrating the higher incidence of knee injuries in female athletes than in their male counterparts (14). The results from the present study demonstrated that the increase in VL f_med from contraction intensities of 10-90% MVC was significantly greater in men than women. These results appear to be supported by the findings of Cioni et al. (6), who observed that women exhibited a "slower" increase in f_med than do men. Similar findings were also presented by Bilodeau et al. (4), namely, that men displayed a significantly greater increase in f_med of the anconeus muscle across the submaximal intensity spectrum than did women. As female VL muscle has been typically shown to exhibit a greater proportion and smaller type I muscle fibers than male (35), the potential range of f_med values may be limited. It is critical to highlight that the characteristics of the surface EMG signal are a manifestation of a progressive increase in type I to type II muscle fiber recruitment as contraction intensity increases. Because the f_med statistic is commonly used to quantify muscle fatigue, for instance, consideration of a contraction intensity effect should be made evident in such an evaluation. Unfortunately, very little data are apparent in the scientific literature regarding f_med of the superficial QF muscle across the range of submaximal contraction intensities in men and women. It should also be pointed out that such differences in EMG f_med between men and women are likely affected by the low-pass-filtering effect of subcutaneous tissue. This relationship has been demonstrated by Bilodeau et al. (4) during submaximal contraction of the triceps brachii and anconeus muscles. Significantly higher levels of skinfold thickness in women were correlated with lower EMG frequency measures. Although the present study is limited by the fact that skinfold thickness was not obtained, such an effect may have been manifested in the VL muscle only, as a gender-specific pattern regarding f_med was not clearly evident for the other muscles. However, the findings from the present study suggest that previously demonstrated differences in muscle fiber-type profiles of the VL muscle between men and women (35) are reflected in the f_med statistic as contraction intensity increases from low (10% MVC) to near-maximal levels. This presumption, coupled with the established effect of skinfold thickness on attenuation of high-frequency components, provides ample rationale for the VL muscle to be a unique muscle for EMG investigations between men and women.

Summary. Distinct differences in QF muscle structure and function are clearly apparent, as outlined. The results of the present study demonstrated that this fact is also reflected in the surface EMG signal. It is speculated that the sensitivity of f_med to reflect progressive muscle fiber recruitment is dependent on the profile of the muscle. Muscles that contain a relatively greater proportion of type I fibers may not demonstrate increases in f_med with greater forces, despite the additional recruitment of these fibers. As a result, the ability to document progressive muscle fiber recruitment with the f_med statistic may be limiting in muscles with significant muscle fiber-type homogeneity. It is apparent, however, that this measure is useful and may be applicable to many different situations when muscle activation characteristics of the VL muscle are examined.