The Reverse Lunge: A Descriptive Electromyographic Study (2025)

2.1. Study Design

The data for this descriptive study were captured in single sessions in a biomechanics laboratory over the span of two months.

2.2. Informed Consent

This research was approved by the Grand Valley State University Institutional Review Board (14-026-H). All participants (n = 21) completed informed consent documentation on the day of testing.

2.3. Participants

A sample of convenience of self-reported healthy, athletically active male and female participants between the ages of 18 and 40 years were recruited. Inclusion criteria included the following: ability to perform a pain-free reverse lunge with appropriate technique for four repetitions and being physically active at least one day a week (equivalent to 30 min of moderate aerobic exercise). Exclusion criteria included a history of lower limb or spinal injury within the prior six months, a history of lower limb or spinal surgery, any musculoskeletal or neurological condition that could influence lower limb function or balance, a functional or anatomical leg-length discrepancy greater than one cm, participation in strenuous exercise in the 12 h prior to data collection, and a history of bleeding disorders.

2.4. Instrumentation

The MA-300, 16-channel EMG system (Motion Lab Systems, Baton Rouge, LA, USA) was used with MLS pre-amplifier and associated wired reusable electrode components for collection of surface EMG data. The EMG system had a high pass filter set at 10 Hz. Additional parameters included raw signals collected at 1200 Hz and anti-aliasing set at six to maintain a bandwidth between 0 and 500 Hz. MLS electrodes (32 × 18 mm each with two metal receiving units 12 mm in diameter) were used and placed in accordance with SENIAM guidelines on the RF, BF, GMed, and GMax on both LEs (Figure S1) [13].

2.5. Data Collection Procedures

Each participant was allowed to practice the reverse lunge with all verbal cues and parameters twice before completing the four recorded lunges. From the starting position, the participants stood with their feet hip-width apart, hands-on-hips, and lunged backwards with the lead limb towards a pre-marked bright line of tape (Figure 1). Four discrete reverse lunges were conducted on each side (i.e., four reverse lunges with the left LE as lead limb and four with the right LE as lead limb). The distance between the starting position and the tape was equal to the leg length of the participant (measured from the greater trochanter to the tip of the lateral malleolus). When stepping backwards to initiate the reverse lunge, the participant planted their foot on the tape, then lowered down so that their stationary limb reached approximately 90 degrees of knee flexion (Figure 1). The participant then extended the knees before bringing both feet back to the starting position. A vertical mirror was placed in front of the participant so that they were able to reverse lunge toward the pre-marked lines without having to look down or back (Figure 1). After each reverse lunge trial, the participant rested for 15 s (i.e., trial 1, 15 s rest, trial 2, 15 s rest, etc.). All participants completed each lunge wearing typical tennis shoes. A metronome of 65 bpm and verbal cueing were utilized to ensure standardization of the timing of the reverse lunge. Verbal cues consisted of the words “step, down, up, step” to cue the participant to the different phases of the reverse lunge. No data were collected regarding the participants’ dominant limb.

Prior to applying the electrodes, the skin was prepared by shaving (if necessary), abrading, and cleaning with alcohol to reduce skin impedance. EMG placement occurred in a private room and the participant was properly draped and covered to maintain modesty. Electrodes were placed on the muscles using Hypafix Cover Roll tape (Essity Medical Solutions, Charlotte, NC, USA) to secure them to the patient’s skin and to limit the amount of electrode movement artifact. The participants wore a vest with Velcro on the back to keep the EMG pack attached to the body during the lunge trials. Precise surface electrode placement was performed via SENIAM protocol, and real-time quality assurance of EMG data output was verified by a single member of the research team visualizing the activity on the computer screen as the patient voluntarily contracted and relaxed each muscle prior to data collection.

After electrode placement and confirmation, the participants warmed up on a stationary exercise bicycle at their preferred pace/resistance for five minutes. Before completing the reverse lunge practice and test trials, EMG data were collected to discern the maximum voluntary isometric contraction (MVIC) of each muscle to normalize the EMG data during the reverse lunge. The MVIC for the GMax was taken with the participant prone over a wedge with the hip extended to neutral (Figure 2). All other MVIC testing positions were consistent with standard manual muscle testing positions: GMed in the side-lying position with the participant’s hip abducted and slightly externally rotated; the BF with the participant in a prone position with the knee flexed to 70 degrees; and the RF with the participant seated with the knee flexed to 90 degrees [14]. Three trials consisting of a five-second hold with a two-second ramp up period were collected for MVIC of each muscle. The second trial was used to establish the MVIC.

A heel switch was utilized to denote three moments during the lunge: movement initiation from the starting position, maximum (max) descent, and terminal foot plant. A single researcher placed the heel switch on a flat surface and manually tapped the heel switch three times during each lunge trial. The first tap occurred when the participant began the reverse lunge (i.e., began to move the lead limb backwards), the second tap occurred when the participant reached the lowest part of the reverse lunge (i.e., max descent) as discerned by visual observation and per standard voice commands, and the last tap occurred when the participant returned to the starting position (feet evenly placed hip-width apart). Each time the heel switch was tapped, a mark appeared on the EMG graph indicating the different phases of the lunge.

2.6. Data Reduction and Analysis

EMG data for both limbs were band-pass filtered (Butterworth high pass filter 10 Hz, low pass filter 300 Hz), rectified and smoothed (using root-mean squared [RMS] analysis), and normalized to MVIC using Motion Lab Systems EMG Analysis and Graphing software (https://motion-labs.myshopify.com/pages/software, accessed date: 4 April 2022) (Motion Lab Systems, Baton Rouge, LA, USA). RMS activity displays for each muscle in the lead and stationary limbs from which peak muscle activations were used to generate descriptive data. RMS is a sophisticated calculation that represents the mean power of the EMG signal. This is the preferred data analysis method for EMG smoothing and generally produces reliable results [15,16].

The second and third lunge trials were utilized for analysis due to the potential for novel learning and potential fatigue effects during the first and fourth trial. Upon preliminary visual analysis of outputs, no discernable differences were present between lunges performed with left and right lead limbs; thus, the lunges performed with the left limb as lead limb were randomly chosen for descriptive analyses. The reverse lunge was qualitatively analyzed in two parts, the lowering period and the rising period to discern where peak muscular activation occurred.

3.1. Sample Demographics

Twenty-one healthy participants aged 22–25 years (11 male and 10 female) participated (Table 3). The age distribution of included subjects was due to this being a sample of convenience that represents college-aged students. For the analysis of data for this project, we did not separate results by gender due to the sample size. Data from one participant were unable to be processed and were not included in analyses (one male). The majority of participants reported completing 15–30 min of exercise three to five times a week.

3.2. Mean Peak Muscle Activity

In the lead limb, the mean peak percent of MVIC muscle activity of the RF, BF, GMed, and GMax was 105.67%, 36.61%, 34.29%, and 34.88%, respectively (Table 4). For the stationary limb, the mean peak percent of MVIC muscle activity of the RF, BF, GMed, and GMax was 36.16%, 17.11%, 49.20%, and 36.61%, respectively (Table 4).

3.3. Mean Average Muscle Activity

The mean of the average percent of MVIC muscle activity over the entire reverse lunge movement was examined for each muscle and each limb. The resulting trends followed those seen in peak muscle activity in the lead and stationary limbs, but the values were reduced overall. The RF mean of the average percent of MVIC was 32.41% for the lead limb vs. 13.55% in the stationary limb. The BF mean of the average percent of MVIC in the lead limb was 10.72% vs. 7.04% in the stationary limb. The GMed mean of the average percent of MVIC in the stationary limb was 15.35% vs. 11.73% in the lead limb. The GMax mean of the average percent of MVIC in the stationary limb was 14.51% vs. 12.68% in the lead limb.

4.1. Muscle Activity Categorization Thresholds

Standardization of peak muscle activity into activation categories was essential to understand the utility of each muscle during the bodyweight reverse lunge, and to best allow clinicians to dose the exercise properly based on therapeutic goals. Reiman et al. designated 0–20% of MVIC as low-level muscle activation, 21–40% of MVIC as moderate-level activation, 41–60% of MVIC as high-level activation, and greater than 60% of MVIC as very high-level activation [17]. According to Reiman et al., EMG activity greater than 40% of MVIC is expected to produce strength gains [17]. Escamilla et al. used the same muscle categorization thresholds in their EMG study of core muscle activation during Swiss ball and traditional abdominal exercises [18]. They state that muscle activations greater than 60% of MVIC may be more effective in strengthening and muscle activations, while less than 20% may be more effective for enhancing endurance [18]. For this discussion, we adopted the suggestion offered by Reiman et al. where 40% of MVIC is considered the threshold for a potential strengthening stimulus and less than 40% of MVIC is considered an endurance stimulus [17]. The two muscles which achieved a strengthening stimulus in their mean peak percent of MVIC were the lead limb RF (105.67% of MVIC) and the stationary limb GMed (49.2% of MVIC). Of note, the mean peak percentage for the lead RF, which exceeded 100%, is likely due to not achieving a true maximum contraction during MVIC testing. The investigator that offered resistance to the RF may have been unable to elicit the fullest contraction of the muscle. The strengthening threshold was attained due to the forceful extensor torque required by the lead RF to rise from max descent and return to the starting position. The GMed acts as a pelvic stabilizer in a single limb stance (closed chain) [19]. As the stationary limb moves forward and the mass of the body is shifted onto a single limb, there is a high level of muscle activation required of the stationary limb GMed to maintain a level pelvis. The stationary limb RF (36.16%), lead and stationary limbs BF (36.61%, 17.11%), lead limb GMed (34.29%), and lead and stationary limbs GMax (34.88%, 36.61%) were in the moderate–level activation category and did not meet the strengthening threshold. This may be due to a lack of stimulus imposed upon these muscles as the reverse lunge in this study was only performed with bodyweight resistance and therefore these muscles were only recruited at the level needed to produce an endurance or stabilizing level stimulus. During data collection, participants were instructed to perform the lunge to the commands of “step, down, up, step”, utilizing both limbs during the knee extension phase. Due to the nature of this standardization, there may have been less weight acceptance onto the stationary limb during the rising half of the lunge, decreasing the required demands of these muscles. It is unclear what would have happened if the two limbs moved at their own pace or more forcefully since this was not investigated.

Regarding the lead limb muscles, there was a low demand placed on the lead limb BF during the hip extension phase (36.11%). As a primary knee flexor, this muscle only contracted enough to clear the foot during the swing phase to reach ground contact. Furthermore, since the reverse lunge movement is performed in the sagittal plane and the GMed functions mainly in the frontal plane, it was not required to work to a great degree throughout the lead limb motion (34.29%) during the reverse lunge.

A study by Distefano et al. found that the lead limb GMed mean of the average percent of MVIC muscle activity was 81% during side-lying hip abduction (a frontal plane movement) versus 48% and 42% for the transverse (side) and forward lunges (the closest sagittal plane movement to the reverse lunge) [6]. The findings related to the forward and transverse lunges by Distefano et al. are consistent with the findings of our study and allow us to infer that, just because a compound movement (such as a lunge) is performed in closed chain, it does not guarantee that greater muscle activation will be elicited. In fact, the results of Distefano et al. [6] and the current results of this study imply that an open chain movement performed in the muscle action’s primary plane may elicit a greater activation than a closed chain movement performed in a non-primary plane. This suggests that, when attempting to bias GMed activation using only bodyweight, the plane of movement may be more important than whether it is performed in an open or closed chain. This should be further studied with a direct comparison of muscle activation between reverse and lateral/side lunges.

The GMax is a large muscle built for power that requires moderate loads to induce high levels of muscle activation [20]. One study assessed a bodyweight reverse lunge and found moderate–level activation (21–40%, as defined by Reiman) from the lead limb GMax [17]. In standing, the lead limb GMax does not need to extend the hip completely against gravity requiring less activation than if the same movement was performed against gravity or a load [14]. An EMG study of the bodyweight forward lunge noted similar findings of peak GMax activation of 36% of MVIC [17]. Although this reference study did not determine during which phase of the forward lunge the peak GMax activation occurred, it could be inferred that the GMax is activated similarly in the forward and reverse lunge.

Muscle activation percentages in the range of 30–40% could likely be increased to peak muscle activations greater than 40% by adding an external load, such as dumbbells. It is inferred that a clinician could change the endurance nature of the bodyweight reverse lunge to a strengthening stimulus for several of the studied muscles with a large enough load increase.

4.2. Peak Muscle Activity Throughout Reverse Lunge Movement

This study analyzed the reverse lunge in two halves: lowering and rising, each comprising ~50% of the lunge movement. Each half phase was then divided into subphases. Lead limb phases of the lowering half included toe off, backswing, ground contact, knee descent, and max descent (Table 1). The lead limb phases of the rising half are knee ascent, front swing, and terminal foot plant (return to bilateral stance) (Table 1). The stationary limb phases of the lowering half are weight acceptance, single limb stance, knee flexion, and max descent (Table 2). The stationary limb phases of the rising half are knee extension, forward weight shift into single limb stance, and terminal foot plant (return to bilateral stance) (Table 2). The phases of movement largely impacted the consistency of peak muscle activation during the reverse lunge. This makes sense when kinesiologic principles are considered. During the movement, each of the considered muscles may be activated in a concentric, eccentric, or isometric manner. Some muscles work as prime movers and others work as stabilizers according to their EMG output.

There is limited evidence in the literature analyzing where peak muscle activation occurs during lunges (forward, lateral, reverse). This study attempted to compare peak muscle activations between the rising and lowering halves of the reverse lunge by using a heel switch to mark the phases. When analyzing where peak muscle activity occurred, the data suggested that all four muscles of the stationary limb and the lead limb RF displayed their most consistent peak muscle activations in the rising half of the lunge. Using kinesiological principles, it may be likely that the peaks of the stationary limb and lead limb RF were consistently found in the rising phase due to the concentric nature/demands during the rising half. There is significant evidence suggesting that EMG activity was greater during concentric activation compared to eccentric activation [1,21,22]. More specifically, qualitative analysis led us to believe that the peak occurred during the knee ascent phase (lead limb) and knee extension phase (stationary limb). During the knee ascent and knee extension phases, the lead and stationary limb RF and stationary GMax were working concentrically against gravity to generate extensor torque to raise the center of mass and bring the body to an upright position. It is unclear why the stationary GMed and BF demonstrated peak muscle activation consistently in the rising half compared to the lowering half. Additionally, the peak muscle activity of the lead BF, GMed, and GMax did not consistently occur in either the rising or lowering half of the lunge. This is likely due to individual liberties and variability with the reverse lunge, especially during the backswing phase.

Lacerda et al. studied the effects of resistance duration during unilateral knee extension exercises on strength gains and muscle hypertrophy of the RF and vastus lateralis muscles [23]. They reported increased gains in MVIC when resistance duration was increased [23]. Since the peak percent of MVIC is a single point in time, there is limited resistance duration. Utilizing the peak percent of MVIC may be important in determining at which phase of the reverse lunge the peak occurred. The clinician may use that information to emphasize muscle activation during that phase (or a portion of that phase) by increasing the duration or load. While we did not quantitatively determine exactly when or during which phase peak muscle activity occurred, we qualitatively determined by visual analysis of the RMS analysis whether the peak occurred in the rising or lowering half. Thus, the phase where the peak occurred can be loosely inferred.

4.3. Clinical Implications

The reverse lunge movement is a widely and commonly used exercise in physical rehabilitation practices. Rationalizing the purpose for prescribed exercise is critical [24]. Since the reverse lunge can be used for many pathologies and conditions, it is important to consider a few examples. Clinically, it is important to understand a patient’s injury or condition to determine their ability to execute this movement safely. For example, a patient after a muscle strain injury (such as a rectus femoris or biceps femoris strain) may have unique LE considerations, and the clinician can intentionally choose a limb to act as the lead limb or stationary limb. When the involved limb acts as the stationary limb there will be increased strength demands of the RF and BF. Since there is increased muscular force required by the lead limb, clinicians should avoid using the involved limb as the lead limb until appropriate muscle healing, and neuromuscular control exists to perform the movement.

In another example, after hip arthroscopy or arthroplasty, a patient may present with impaired frontal plane pelvic stability and GMed weakness [25]. In addition to standard isolated rehabilitation exercises, the reverse lunge may be used to impact not only strength, but stability, balance, and motor control. This study’s results suggest that designating the involved limb as the stationary limb would best target strengthening for the involved side GMed, while using the involved limb as the lead limb may only be effective at enhancing endurance for the involved side GMed.

Compared to the forward lunge, the reverse lunge has been demonstrated to invoke decreased patellofemoral (PF) joint compressive forces due to the smaller angle of knee flexion on the lead limb [7]. The results of the current study revealed that the RF, which is a primary knee extensor and eccentric controller of knee flexion, was activated to a low-level (stability/endurance) stimulus in the stationary limb and a high-level (strengthening) stimulus in the lead limb. If a patient is experiencing pain at the PF joint, designating the involved limb as the stationary limb while performing the reverse lunge (compared to the involved limb as the lead limb) may decrease the force required by the muscles. This would result in decreased contact stresses at the PF joint and allow for stability and endurance training. However, in the absence of PF pain, the RF is activated to a greater extent in the lead limb compared to the stationary limb, and it is more likely to produce strength gains.

It is difficult to make direct comparisons with the existing literature due to the limited studies available on the reverse lunge EMG. However, we can make loose comparisons of the overall activation of key LE musculature between similar movements. A study by Stastny et al. investigating the activation of muscles during the split squat showed similar peak muscle activation for the BF (19%), RF (38–40%), and GMed (25%) of the forward leg in the split squat, which correlates to the stationary limb in the reverse lunge [26]. The GMed activation of the stationary limb during the reverse lunge (49.2%) was greater than that found in the study by Stastny et al. and was likely due to the demands placed on the muscle in single limb stance [26]. Distefano et al. [6] found the GMed muscle activation of the moving limb during the bodyweight lateral lunge to be 39%. Although the lateral lunge is a frontal plane movement, the GMed activation was similar to the current results in the reverse lunge lead limb (34.29%).

When comparing the reverse lunge to similar loaded movements, one study by Muyor et al. assessed a weighted barbell forward lunge and found that of all LE muscles assessed (GMed, GMax, BF, RF, vastus lateralis, and vastus medialis), the RF was consistently the most active [1]. Loose inferences can be made that, in both loaded and unloaded lunges, the RF would be the most activated. Additionally, the LE muscles assessed had generally greater muscle activation in the concentric phase compared to the eccentric phase. This is consistent with the findings in our study.

4.4. Limitations

During the completion of this study, several limitations were identified. First, the sample size is small and only includes healthy 22–25-year-olds, which limits the generalizability of the observed results to larger populations including different age groups. Second, typical of surface EMG, artifact and impedance may have negatively affected the ability to collect quality EMG data of key musculature intermittently, especially the GMax. It was essential to limit this as much as possible. This was conducted through precise EMG electrode placement via SENIAM protocol and real-time quality assurance of the EMG data during collection by a member of the research team. However, even with controls, there was potential signal interference with the EMG electrodes during data collection including crosstalk between other adjacent muscles, electrical wire disturbances, and EMG electrical wire port damage. We believe these factors primarily impacted on the validity of the data collected regarding the GMax. The raw data of the GMax indicated more muscle activity than expected with excessive EMG noise at baseline. Despite this, we believe the EMG data displayed a reasonably accurate representation of muscle activity in the lead and stationary limb. Another limitation was that we did not measure skin impedance and its relation in dampening the EMG signal. Utilization of the heel switch introduced potential human error. However, using a single researcher to manually hit the switch at key moments in the reverse lunge allowed for a systematic approach for sequencing the reverse lunge phases and adding context to the EMG data. Lastly, this study only analyzed the left limb acting as the lead limb despite also collecting data for the right limb acting as the lead limb.

Since this study is a preliminary descriptive study, further research is needed for exploration of the reverse lunge. This study worked with primarily healthy, physically active individuals. Future research would be required to examine the differences in muscle activation across other populations, including those who are deconditioned, injured, or otherwise impaired. With this data, future research can analyze the differences between lunge types, i.e., forward, side, and reverse within participants. Since the current study only analyzed the left limb as the lead limb, future studies could analyze both limbs as lead limbs to decipher interlimb differences. Finally, this study only assessed the bodyweight lunge, and future research could examine other variables, including the effects of changes in arm position, trunk position, and addition of weight or resistance.

Supplementary Materials