Muscular coactivation. The role of the antagonist musculature in maintaining knee stability

The objective of this study was to quantify the coactivation patterns of the knee flexor and extensor muscles as part of continued efforts to identify the role of the antagonist muscles in maintaining joint stability. The simultaneous EMG from the flexor and extensor muscles of the knee were recorded during maximal effort, slow isokinetic contractions (15 deg/sec) on the plane parallel to the ground to eliminate the effect of gravity. The processed EMG from the antagonist muscle was normalized with respect to its EMG as agonist at maximal effort for each joint angle. The plots of normalized antagonist EMG versus joint angle for each muscle group were shown to relate inversely to their moment arm variations over the joint range of motion. Additional calculations demonstrated that the antagonist exerts nearly constant opposing torque throughout joint range of motion. Comparison of data recorded from normal healthy subjects with that of high performance athletes with hypertrophied quadriceps demonstrated strong inhibitory effects on the hamstrings coactivations. Athletes who routinely exercise their hamstrings, however, had a coactivation response similar to that of normal subjects. We concluded that coactivation of the antagonist is necessary to aid the ligaments in maintaining joint stability, equalizing the articular surface pressure distribution, and regulating the joint's mechanical impedance. The reduced coactivation pattern of the unexercised antagonist to a hypertrophied muscle increases the risk of ligamentous damage, as well as demonstrates the adaptive properties of the antagonist muscle in response to exercise. It was also concluded that reduced risk of knee injuries in high performance athletes with muscular imbalance could result from complementary resistive exercise of the antagonist muscle.     

Co-activation of sprinter and distance runner muscles in isokinetic exercise

The purpose of this study was to investigate the extent of co-activation of quadriceps and hamstring musculature in sprinters and distance runners. Nine female intercollegiate track athletes performed maximal knee extensions and flexions on a modified orthotron isokinetic dynamometer at two speeds (100 degrees and 400 degrees X s-1). Simultaneous recordings of torque, joint position, and agonist/antagonist electromyographic activity from the quadriceps and hamstrings were computer-processed. The results revealed the hamstrings to be considerably more active during knee extension than the quadriceps during flexion. The integrated electromyographic activity of co-contracting hamstrings and quadriceps, throughout the joint range, averaged 33 and 6%, respectively, of the same muscle group during its agonist phase. Hamstring co-activation increased sharply during the last 25% of knee extension, generating 58% of the integrated electromyographic agonist activity. Co-activation of the sprinters' hamstrings was four times that of distance runners (57/14%), however, the faster speed of movement (400 degrees X s-1) increased hamstring coactivation of distance runners more acutely than sprinters in the final phase of extension. The data suggest that the hamstrings are used to a much greater extent than quadriceps for limb deceleration and that the distraction of antagonist muscle tension should be considered when analyzing agonist isokinetic torques. Furthermore, the relatively high co-activation of the hamstrings, particularly during the last 25% of extension, may induce hamstring soreness or strain in vulnerable subjects.     

Effect of knee joint angle on neuromuscular activation of the vastus intermedius muscle during isometric contraction

The purpose of this study was to compare the relationship between surface electromyography (EMG) and knee joint angle of the vastus intermedius muscle (VI) with the synergistic muscles in the quadriceps femoris (QF) muscle group. Fourteen healthy men performed maximal voluntary contractions during isometric knee extension at four knee joint angles from 90°, 115°, 140°, and 165° (180° being full extension). During the contractions, surface EMG was recorded at four muscle components of the QF muscle group: the VI, vastus lateralis (VL), vastus medialis (VM), and rectus femoris (RF) muscles. The root mean square of the surface EMG at each knee joint angle was calculated and normalized by that at a knee joint angle of 90° for individual muscles. The normalized RMS of the VI muscle was significantly lower than those of the VL and RF muscles at the knee joint angles of 115° and 165° and those of the VL, VM, and RF muscles at the knee joint angle of 140° (P<0.05). The present results suggest that the neuromuscular activation of the VI muscle is regulated in a manner different from the alteration of the knee joint angle compared with other muscle components of the QF muscle group.     

EMG power spectra of intercollegiate athletes and anterior cruciate ligament injury risk in females

PURPOSE: Females have a disproportionately high incidence of anterior cruciate ligament (ACL) injuries compared with males in analogous sports. Although the pathogenesis of this higher frequency has not been elucidated, gender differences in neuromuscular control of the knee may play an important role. This study evaluates EMG power spectra of the quadriceps and hamstring muscles during dynamic, fatiguing exercise to examine differences between male and female intercollegiate athletes. METHODS: Fifty-one collegiate basketball and soccer players (25 female, 26 male) were studied. Maximum voluntary contraction (MVC) was determined for knee flexion and extension. Three consecutive 2-min bouts of isokinetic knee flexion and extension exercise were performed at 40% MVC. EMG activity in the biceps femoris and vastus medialis obliquus was recorded using bipolar surface electrodes. RESULTS: MVC normalized to body weight was significantly greater in males than in females for the quadriceps (P< 0.01). Quadriceps coactivation ratios were significantly higher in females than in males during knee flexion exercises (P< 0.01). CONCLUSIONS: This study demonstrates differences in the EMG power spectra for females when compared with a matched group of males. Increased quadriceps coactivation in females may increase anterior tibial loads under dynamic conditions, thus placing the ACL at higher risk for injury in the female athlete.     

Quadriceps activation in closed and in open kinetic chain exercise

PURPOSE: For treatment of various knee disorders, muscles are trained in open or closed kinetic chain tasks. Coordination between the heads of the quadriceps muscle is important for stability and optimal joint loading for both the tibiofemoral and the patellofemoral joint. The aim of this study was to examine whether the quadriceps femoris muscles are activated differently in open versus closed kinetic chain tasks. METHODS: Ten healthy men and women (mean age 28.5 +/- 0.7) extended the knees isometrically in open and closed kinetic chain tasks in a reaction time paradigm using moderate force. Surface electromyography (EMG) recordings were made from four different parts of the quadriceps muscle. The onset and amplitude of EMG and force data were measured. RESULTS: In closed chain knee extension, the onset of EMG activity of the four different muscle portions of the quadriceps was more simultaneous than in the open chain. In open chain, rectus femoris (RF) had the earliest EMG onset while vastus medialis obliquus was activated last (7 +/- 13 ms after RF EMG onset) and with smaller amplitude (40 +/- 30% of maximal voluntary contraction (MVC)) than in closed chain (46 +/- 43% MVC). CONCLUSIONS: Exercise in closed kinetic chain promotes more balanced initial quadriceps activation than does exercise in open kinetic chain. This may be of importance in designing training programs aimed toward control of the patellofemoral joint.     

Fatigue differences between adults and prepubertal males

The purpose of the present study was to investigate the differences in neuromuscular activation of agonist and antagonist muscles between men and prepubertal boys during a maximal isokinetic fatigue test. Ten prepubertal boys (mean +/- SD age: 10.5 +/- 0.6 years) and fourteen adults (age: 24.3 +/- 2.5 years) executed 25 consecutive maximal isokinetic knee extensions at 60 deg . s (-1). Peak torque and the electromyogram (EMG) of the vastus lateralis, vastus medialis and biceps femoris muscles were recorded. During the fatigue protocol, the prepubertal boys were able to produce higher torque than the adults, when expressed as percent of their maximal value, indicating that adults were more fatigable. The agonist activity, especially for the vastus lateralis muscle, increased in both groups during the first 10 knee extensions, and then decreased more in adults. The antagonist activity of biceps femoris muscle in adults remained constant throughout the fatigue task, whereas the children showed, on average, an increased biceps femoris antagonistic activation, especially during the first 10 and last 5 knee extensions. These results suggest that adults are more fatigable than children during a maximal isokinetic fatigue protocol, probably due to an increased inhibition or reduced facilitation of their agonist drive.     

The effect of joint velocity on the contribution of the antagonist musculature to knee stiffness and laxity

The electromyographic (EMG) coactivation patterns of the knee flexors and extensors when acting as antagonists were studied as a function of limb velocity to assess their contribution to joint stiffness and laxity. Normalized antagonist coactivation patterns developed from surface EMG recordings from the hamstrings and quadriceps during maximal effort isokinetic extension and flexion, respectively, demonstrated characteristic variations as the joint velocity increased from 15 deg/sec up to 240 deg/sec. The two-tailed t-test (P less than 0.01) was performed on the data obtained from eight normal knees. The results indicate that both hamstrings and quadriceps demonstrate a significant increase (greater than 100%) in their antagonist coactivation pattern during the final 40 degrees of fast extension and flexion movements, respectively, as limb velocity increases. A minor decrease in antagonist activity of the hamstrings (24%) and quadriceps (8%) was evident during the initial phase of the extension and flexion movements, respectively, as joint velocity increased. We concluded that as limb velocity is increased, there is a substantial reflexive (unintentional) increase in the contribution of the antagonist musculature to joint stiffness and reduction of laxity. The results also suggest that strength training of the hamstrings (rather than quadriceps) should be considered as a modality for conservative treatment of ACL deficiencies, as well as an adjunct to surgical reconstruction. Such training can also reduce the risk of high performance athletes in a reflexive manner by increasing joint stiffness.     

Electromyographic and kinematic analysis of cutting maneuvers. Implications for anterior cruciate ligament injury

The objective of this study was to qualitatively characterize quadriceps and hamstring muscle activation as well as to determine knee flexion angle during the eccentric motion of sidestep cutting, cross-cutting, stopping, and landing. Fifteen healthy collegiate and recreational athletes performed the four movements while knee angle and electromyographic activity (surface electrodes) of the vastus lateralis, vastus medialis obliquus, rectus femoris, biceps femoris, and medial hamstring (semimembranosus/semitendinosus) muscles were recorded. The results indicated that there is high-level quadriceps muscle activation beginning just before foot strike and peaking in mid-eccentric motion. In these maneuvers, the level of quadriceps muscle activation exceeded that seen in a maximum isometric contraction. Hamstring muscle activation was submaximal at and after foot strike. The maximum quadriceps muscle activation for all maneuvers was 161% maximum voluntary contraction, while minimum hamstring muscle activity was 14%. Foot strike occurred at an average of 22 degrees of knee flexion for all maneuvers. This low level of hamstring muscle activity and low angle of knee flexion at foot strike and during eccentric contraction, coupled with forces generated by the quadriceps muscles at the knee, could produce significant anterior displacement of the tibia, which may play a role in anterior cruciate ligament injury.     

Fatigue after eccentric quadriceps femoris work produces earlier gastrocnemius and delayed quadriceps femoris activation during crossover cutting among normal athletic women

Athletic women are at greater risk of anterior cruciate ligament (ACL) injury than men. Twenty, healthy, athletic women were evaluated for the effect of preferred stance limb isokinetic quadriceps femoris and hamstring fatigue from eccentric work compared with controls on the activation onset of vastus medialis, rectus femoris, vastus lateralis, the medial hamstrings, biceps femoris, and gastrocnemius muscles. Following 3 weeks of crossover cut training, subjects were tested for fatigue effects (5 subjects/week, 3 conditions, 1 condition/day, order effect controlled) on muscle activation onsets prior to crossover cut landing heelstrike (mixed model, ANOVA, P < 0.05). Fatigue from eccentric quadriceps femoris work produced delayed vastus medialis (P = 0.03), rectus femoris (P = 0.007), and vastus lateralis (P = 0.03) activation onsets compared with control, but did not differ compared to hamstring fatigue. Neither hamstring nor quadriceps femoris fatigue produced differences (P > 0.05) in medial hamstring or biceps femoris activation onsets compared to control. Quadriceps femoris fatigue from eccentric work produced earlier gastrocnemius activation onsets (P = 0.048) than control, but did not differ for hamstring fatigue. The gastrocnemius appears to provide synergistic and compensatory dynamic knee stabilization in closed kinetic chain function during quadriceps femoris fatigue. This finding in a normal group at high risk of ACL injury while performing a maneuver with a high ACL injury risk supports gastrocnemius inclusion in knee rehabilitation and conditioning programs and suggests the need for comparative evaluations of knee injured/reconstructed subjects.     

An evaluation of knee extensor and knee flexor torques and EMGs in patients with patellofemoral pain syndrome in comparison with matched controls

The relationship between concentric and eccentric isokinetic torques and EMGs of quadriceps and hamstring muscles in patients with unilateral patellofemoral pain was studied in 27 patients (13 males, 14 females). The patients and a group of controls matched for age, gender, and physical activity were tested on a Kin-Com dynamometer at 60⁰/s and 180⁰/s angular velocity. EMGs were recorded for eight of the patients and their matched controls. In addition, the reproducibility of isokinetic measurements made under the same conditions but on different occasions in patients with patellofemoral pain was evaluated. Twenty-one patients (11 males, 10 females) underwent testing of their quadriceps and hamstring muscles two or three times on a Kin-Com dynamometer. This was performed both concentrically and eccentrically in their painful leg while the patients evaluated their knee pain using Borg's pain scale. The data show that the patients had a significantly lower agonist as well as antagonist EMG activity during knee extension measurements in their painful leg compared with the controls. However, there were no differences in either agonist or antagonist EMG activities during knee flexion measurements between the patients and the controls. The quadriceps muscle torque was considerably weaker in the patients' painful leg compared with both their asymptomatic leg and with the controls. Peak torque for knee extension was reached at a mean of 66⁰ of knee flexion for both patients and controls. However, the patients showed a considerably wider range within which they produced their peak torque in their painful leg than in their asymptomatic and also in comparison with the controls. The hamstring muscle torque was weaker in both legs of the patients compared with the controls. There were no significant differences in knee pain according to Borg's pain scale between the isokinetic tests or within the tests. The reproducibility testing yielded no significant differences between the different testing occasions in the two muscle groups. In summary, patients with patellofemoral pain have a greatly reduced agonist and antagonist EMG activity during knee extension in their painful leg. They also have a considerably reduced quadriceps torque. Furthermore, isokinetic testing in patients with patellofemoral pain is a reproducible testing modality.     

Comparison of clinical and dynamic knee function in patients with anterior cruciate ligament deficiency

BACKGROUND: Whether passive measures of isokinetic muscle strength deficits and knee laxity are related to the dynamic function of the anterior cruciate ligament-deficient knee remains unclear. HYPOTHESES: Arthrometer measurements are not predictive of peak external knee flexion moment (net quadriceps muscle moment), isokinetic quadriceps muscle strength correlates with peak external knee flexion moment (net quadriceps muscle moment), and isokinetic hamstring muscle strength correlates with peak external knee extension moment (net flexor muscle moment). STUDY DESIGN: Cross-sectional study. METHODS: Gait analysis was used to assess dynamic function during walking, jogging, and stair climbing in 44 subjects with unilateral anterior cruciate ligament deficiency and 44 control subjects. Passive knee laxity and isokinetic quadriceps and hamstring muscle strength were also measured. RESULTS: Arthrometer measurements did not correlate with peak external flexion or extension moments in any of the activities tested or with isokinetic quadriceps or hamstring muscle strength. Test subjects also had a significantly reduced peak external flexion moment during all three jogging activities and stair climbing compared with the control subjects and this was correlated with significantly reduced quadriceps muscle strength. CONCLUSIONS: Absolute knee laxity difference did not correlate with dynamic knee function as assessed by gait analysis and should not be used as a sole predictor for the outcome of treatment. Patients with greater than normal strength in the anterior cruciate ligament-deficient limb performed low- and high-stress activities in a more normal fashion than those with normal or less-than-normal strength.     

Quadriceps EMG/force relationship in knee extension and leg press

PURPOSE: This study compared the relationship between surface electromyographic (EMG) activity and isometric force of m. quadriceps femoris (QF) in the single-joint knee extension (KE) and the multi-joint leg press (LP) exercises. METHODS: Nine healthy men performed unilateral actions at a knee angle of 90 degrees at 20, 40, 60, 80, and 100% of maximal voluntary contraction (MVC). EMG was measured from m. vastus lateralis (VL), m. vastus medialis (VM), m. rectus femoris (RF), and m. biceps femoris (BF). RESULTS: There were no differences in maximum EMG activity of individual muscles between KE and LP. The QF EMG/force relationship was nonlinear in each exercise modality. VL showed no deviation from linearity in neither exercise, whereas VM and RF did. BF activity increased linearly with increased loads. CONCLUSIONS: The EMG/force relationship of all quadricep muscles studied appears to be similar in isometric multi-joint LP and single-joint KE actions at a knee angle of 90 degrees. This would indicate the strategy of reciprocal force increment among muscles involved is comparable in the two models. Furthermore, these data suggest a nonuniform recruitment pattern among the three superficial QF muscles and surface EMG recordings from VL to be most reliable in predicting force output.     

Effects of aquatic resistance training on neuromuscular performance in healthy women

PURPOSE: The purpose of this study was to investigate the effects of a progressive 10-wk aquatic resistance training on neuromuscular performance and muscle mass of the knee extensors and flexors in healthy women. METHODS: Twenty-four healthy women (34.2 +/- 3.9 yr) were randomly assigned into aquatic exercise (N = 12) and control group (N = 12). Maximum knee extension and flexion torques were measured isometrically and at constant angular velocities of 60 degrees x s(-1) and 180 degrees x s(-1) (isokinetic) with simultaneous electromyography (EMG) recordings of the quadriceps and hamstrings. The lean muscle mass (LCSA) of the quadriceps and hamstring muscles was determined by computed tomography scanning. RESULTS: Significant interaction of group by time was observed in each of the measured parameters. The change in extension and flexion isometric/isokinetic torque varied between 8 and 13% and in EMGs between 10 and 27% in the exercise group. The change in the quadriceps LCSA of the exercise group was 4% and in hamstrings 5.5%. CONCLUSIONS: The results of the present study showed that 10 wk of progressive aquatic resistance training resulted in significant improvement in muscle torque of the knee extensors and flexors accompanied with proportional improvement in neural activation and with significant increase in the LCSA of the trained muscles. Aquatic training can be recommended for neuromuscular conditioning in healthy persons.     

Muscle fatigue during repetitive voluntary contractions: A comparison between children with cerebral palsy,typically developing children and young healthy adults

AimTo combine peak torque and EMG analyses to investigate the hypothesis that 1) children with cerebral palsy (CP) have lower muscle fatigability than typically developing children (TD) and whether 2) muscle fatigue correlates with muscle strength.MethodsSeven CP children, eight TD children and ten young healthy adults (YHA) performed an all-out fatigue test of 35 maximal concentric knee extension and flexion contractions on an isokinetic dynamometer. Angular velocity was set at 60°/s. Peak torque (PT) was determined per repetition and either normalized to bodyweight or maximum voluntary torque. Surface-EMG of quadriceps and hamstring muscles was measured to obtain changes in median frequency (EMG-mf) and smooth rectified EMG amplitude per contraction.ResultsDecline in PT differed between all groups for extensors and flexors, where YHA showed the largest decline and CP children the smallest decline over the course of the test. YHA showed a larger decline in EMG-mf of all quadriceps and hamstrings than TD and CP children, while TD children showed a larger decline in EMG-mf of m.rectus femoris and m.vastus lateralis than CP children.InterpretationResults confirm that children with CP show lower fatigability than TD children and that the lower fatigability coincides with lower maximal muscle strength.     

Muscle activation and torque development during maximal unilateral and bilateral isokinetic knee extensions

BACKGROUND: The aim of this study was to investigate whether or not a bilateral strength deficit occurs during bilateral (BL) velocity controlled dynamic knee extensions and if the neural control of the knee extensors and flexors is altered during homologous muscle BL efforts. METHODS: Twenty-eight healthy and habitually active subjects, 13 female and 15 male, performed maximal unilateral (UL) and BL isokinetic leg extensions at a velocity of 60 degrees.s-1 through a 90 degrees range of motion of the knee joint (90 to 180 degrees). Knee extension torque and electromyographic activity (EMG) of the quadriceps and hamstrings muscles were recorded. RESULTS: The mean knee extensor torque produced in the BL condition (168 +/- 52 Nm) was 17% less than the sum of the two UL conditions (Sigma=202 +/- 56 Nm). During BL conditions, quadriceps EMG activity was less in both legs (left, 8.2 +/- 7.4% less and right, 13.9 +/- 9.1% less, respectively). There were no significant differences between BL and UL efforts for either left or right hamstrings activity. Eighteen subjects, who when asked to perform a maximal knee extension simultaneously activated their contralateral hamstrings, had significantly higher bilateral deficits (21%) compared to those who exhibited little or no contralateral hamstrings EMG activity (14%). CONCLUSIONS: The main findings of the study were that a bilateral strength deficit occurred when simultaneously maximally activating the homologous knee extensor muscles. This deficit was in all likelihood due to a less than maximal efferent drive to the quadriceps muscles. Hamstrings EMG activity was not greater during the BL knee extensions, which supports the notion that antagonistic muscle activity was not primarily responsible for the observed bilateral deficit.     

Effects of eccentric versus concentric training on thigh muscle strength and EMG

The purpose of this study was to compare pure eccentric and concentric strength training regarding possible specific effects of muscle action type on neuromuscular parameters, such as a decreased inhibition during maximal voluntary eccentric actions. Two groups of young healthy adult men performed 10 weeks of either eccentric or concentric unilateral isokinetic knee extensor training at 90 degrees.s(-1), 4 sets of 10 maximal efforts, 3 days a week. Knee extensor torque and surface EMG from the quadriceps and hamstring muscle groups were collected and quantified in a window between 30 and 70 degrees knee angle (range of motion 90-5 degrees ) during maximal voluntary eccentric and concentric knee extensor actions at 30, 90, and 270 degrees.s(-1). Changes in strength of the trained legs revealed more signs of specificity related to velocity and contraction type after eccentric than concentric training. No major training effects were present in eccentric to concentric ratios of agonist EMG or in relative antagonist (hamstring) activation. Thus, for the trained leg, the muscle action type and speed specific changes in maximal voluntary eccentric strength could not be related to any effects on neural mechanisms, such as a selective increase in muscle activation during eccentric actions. Interestingly, with both types of training there were specific cross-education effects, that is, action type and velocity specific increases in strength occurred in the contralateral, untrained, leg, accompanied by a specific increase in eccentric to concentric EMG ratio after eccentric training.     

A dynamic model of quadriceps and hamstrings function

The mechanical effect of hamstrings and quadriceps contractions on hip and knee joint motion was investigated using a dynamic model of the musculoskeletal system. The model consisted of 13 anatomically linked segments. The geometry of bones, joints, and muscle attachments was derived from magnetic resonance imaging of a healthy adult. The knee joint was represented by a crossing bars linkage to simulate cruciate ligament function, and muscles were represented by spring actuators. The effects of hamstring and quadriceps contractions, in various combinations, were tested on different configurations of hip and knee joint position in the absence of gravity.In the standing posture, with the foot free to move and the pelvis fixed in space, the effect of semimembranosus (SM) contraction was hip and knee flexion. If the foot was fixed to the ground, SM contraction produced hip extension and knee flexion. The addition of quadriceps contraction reduced or abolished the knee flexion and enhanced hip extension. In all other simulations, SM alone produced knee flexion and hip extension and the combination of SM with vastus (VA) and rectus femoris (RF) contractions resulted in knee extension and enhanced hip extension. Our findings suggest that co-contraction of quadriceps and hamstrings may be a strategy to increase the hip extension function of the hamstrings.     

Relationship of spasticity to knee angular velocity and motion during gait in cerebral palsy

This study investigated the effects of spasticity in the hamstrings and quadriceps muscles on gait parameters including temporal spatial measures, knee position, excursion and angular velocity in 25 children with spastic diplegic cerebral palsy (CP) as compared to 17 age-matched peers. While subjects were instructed to relax, an isokinetic device alternately flexed and extended the left knee at one of the three constant velocities 30 degrees/s, 60 degrees/s and 120 degrees/s, while surface electromyography (EMG) electrodes over the biceps femoris and the rectus femoris recorded muscle activity. Patients then participated in 3D gait analysis at a self-selected speed. Results showed that, those with CP who exhibited heightened stretch responses (spasticity) in both muscles, had significantly slower knee angular velocities during the swing phase of gait as compared to those with and without CP who did not exhibit stretch responses at the joint and the tested speeds. The measured amount (torque) of the resistance to passive flexion or extension was not related to gait parameters in subjects with CP; however, the rate of change in resistance torque per unit angle change (stiffness) at the fastest test speed of 120 degrees/s showed weak to moderate relationships with knee angular velocity and motion during gait. For the subset of seven patients with CP who subsequently underwent a selective dorsal rhizotomy, knee angular extension and flexion velocity increased post-operatively, suggesting some degree of causality between spasticity and movement speed.     

Central and peripheral fatigue of the knee extensor muscles induced by electromyostimulation

The main purpose of this study was to characterise neuromuscular fatigue induced by 30 contractions of the knee extensor muscles evoked by electromyostimulation (EMS). Twelve healthy subjects were tested before and after a typical EMS session (frequency: 75 Hz, on-off ratio: 6.25 s on-20 s off) used for quadriceps femoris muscle strengthening. Surface electromyographic (EMG) activity and torque obtained during maximal voluntary and electrically evoked contractions were analysed to distinguish peripheral from central fatigue. Maximal voluntary torque of the knee extensor muscles decreased approximately 20 % (p < 0.001) following EMS. In the same way, peak torque associated to single (p < 0.05) and paired (p < 0.001) stimuli as well as M-wave amplitude (p < 0.05) significantly decreased as a result of EMS. The raw EMG activity of both vastus lateralis and rectus femoris muscle recorded during maximal voluntary isometric contraction significantly decreased after the session (-17.3 and -14.5 %, respectively) whereas no changes were observed when EMG signals were normalised to respective M-wave amplitudes. Similarly, voluntary activation estimated by using the twitch interpolation technique was unchanged following EMS. In conclusion, a typical session of EMS of the knee extensor muscles mainly induced neuromuscular propagation failure while excitation-contraction coupling and neural mechanisms were not significantly affected. It is recommended to interpret surface EMG data together with the corresponding M wave, at least for the knee extensor muscles, in order to distinguish peripheral from central causes of fatigue.