Heel raises, also referred to as “calf raises,” are common exercises included in resistance training programs for increasing the size, strength, and power of the gastrocnemius and soleus muscles (11,24). Many different populations are concerned with training these muscles, such as sprinters (2) and jumpers for improving performance (5), bodybuilders for increasing muscular symmetry of the lower extremity (24), older adults for maintaining mobility (11), and patients recovering from Achilles tendinopathy (17,23). Both muscles function as ankle plantar flexors; however, differences in fiber type (1), architecture (13,16), and function (24) support the notion that they be considered independently (9). Although the soleus muscle may be targeted independent of knee position, the contribution of the gastrocenmius to ankle plantar flexion, being a biarticular muscle, is dependent upon both the knee and ankle joints (16,24). Thus, from an exercise training perspective, variations of heel-raise exercises involving the knee in a straight (soleus and gastrocnemius) and flexed (soleus) position should be incorporated to fully promote soleus and gastrocnemius function (24).
Similar to other multiheaded muscles, such as the hamstrings (8) and quadriceps (7), evidence has also suggested that functional differences exist between the medial (MG) and lateral (LG) heads of the gastrocnemius as a result of architectural characteristics (16,18). For example, Kawakami et al. examined the architectural features of the MG and LG at several different ankle and knee-flexion angles under both passive and active states (16). Although the LG was revealed to have longer fascicle lengths independent of ankle or knee position and activation state, the MG demonstrated more fibers within a certain volume secondary to shorter fascicle lengths and fascicle angles. As knee flexion increased, the MG became increasingly more disadvantaged regardless of ankle position. These results suggest functional differences in the force-producing capabilities between the MG and LG depending upon ankle and knee-joint position. From a resistance training perspective, understanding how different ankle and knee angular positions, and how other mechanical alterations, affect the MG and LG functions may provide evidence to support heel-raise variations as efficacious ways to bring about separate morphological and functional adaptations in the medial and lateral heads of the gastrocnemius.
Measuring muscle activation using electromyography (EMG) provides additional insight into functional differences between muscles. Specific to the MG and LG muscles, several studies have examined differences in activation between modes of exercise, such as cycling (3) and running (15), and levels of activation (submaximal to maximal) during plantar flexion with varying degrees of knee flexion (6,9,10). Using EMG in this manner provides objective rationale on which practitioners can base exercise-mode and intensity decisions (4).
Exercisers are often observed performing heel raises with the foot pointing in 3 different positions, inward (internally rotated [IR] leg), outward (externally rotated [ER] leg), and forward (neutral [NE] leg). The rationale for altering foot positions is based on the assumption that the different foot positions will prompt maximal activation of both the MG and LG across sets involving these variations. This notion is similar to variations of other exercises such as squats (20,22). Because research has not established whether foot position affects MG and LG muscle activity during heel-raise exercise, the purpose of this study was to compare MG and LG activation during the concentric (CP) and eccentric (EP) phases of the heel-raise exercise using neutral, IR, and ER foot positions.
Experimental Approach to the Problem
Heel raises are frequently used as gastrocnemius- and soleus-strengthening exercise often performed with 3 different foot positions, neutral, IR, and ER. The rationale for using different foot positions centers on attempting to maximize the activation of both the MG and LG during an exercise bout. This study was designed to investigate whether the 3 foot positions prompt different MG and LG activation during free-standing weighted (130-135% of body weight) heel raises. A repeated-measures counterbalanced design was used to answer the question: During heel raises, will varying foot position significantly change the levels of MG and LG activation?
Twenty physically active subjects (10 men, 10 women, 1.71 ± .07 m, 23.7 ± 3.1 years, and 72.75 ± 14.24 kg) voluntarily participated in this study. All subjects participated in some physical activity 3 times a week for at least 30 minutes, including a recent history of using heel-raise exercise. Additionally, all participants were void of previous-injury history prohibiting performance of free-standing heel-raise exercise or conditions that could confound MG and LG activation. Before participation, each subject received a verbal overview of the study's purpose followed by time to read, review, and sign an Institutional Review Board approved consent form.
Subjects completed all study procedures during a single, 1-hour session. Before data collection, subjects were given ample time to become proficient with the free-standing heel raises using all 3 foot positions. Subjects completed 1 set of 12 repetitions using each foot position, with the set order counterbalanced between subjects. All heel raises began with the subjects adopting a comfortable stance width with their unshod (barefoot) forefeet elevated onto a 3.81-cm wooden block while holding 130-135% of their body mass that included a 16-kg Olympic weightlifting bar (Figure 1). This block height and load level were selected as a compromise between stimulating muscular activation without promoting excessive challenge to retaining postural equilibrium. The neutral stance position involved the subject maintaining both feet pointing anterior. During the IR position, subjects were instructed to point their toes inward by internally rotating their legs as far as possible (Figure 2), whereas for the ER position, subjects were instructed to point their toes outward by externally rotating their legs as far as possible (Figure 3). For all 3 foot positions, subjects were instructed to maintain the knee in full extension. The 12 repetitions under each foot position condition were self-initiated and completed within a 30-second period, with each repetition on an “up-one thousand, down-one thousand” cadence. Between sets, subjects placed the barbell on a squat rack and were given 3 minutes of rest.
Data Collection and Reduction
Using the Bagnoli-8 System (Delsys, Inc, Boston, MA), raw EMG data were sampled at 1,000 Hz, filtered (20-450 Hz), and amplified with a minimum common mode rejection ratio >84 dB and input impedance > 10 kΩ. The gain of each channel was adjusted (100; 1,000; or 10,000) to maximize signal resolution. Data were then analog to digital converted (ComputerBoards PCM16S/12, ComputerBoards, Inc., Middleboro, MA) and stored on a desktop computer using the Motion Monitor data acquisition software (Innovative Sports Training, Inc., Chicago, IL). Additionally, a digital video camera (Sony Handycam DCR-HC52), synchronized with the EMG data collection, captured all repetitions.
After practicing and becoming proficient with heel raises using the 3 foot positions, rectangular-shaped bipolar (1-cm interelectrode distance) (DelSys DE-2.1, Delsys, Inc, Boston, MA) 99.9% Ag electrodes were secured over the MG and LG muscles. Subjects lay prone on a plinth with feet overhanging the edge to keep their knees straight. Subjects were asked to plantarflex against resistance to determine the center of the MG and LG muscle bellies. A mark was made where electrodes would be placed according to the recommendations of SENIAM (12). To prepare the skin surface for the electrodes, hair was shaved, skin brushed with an abrasive cloth followed by an isopropyl alcohol wipe. A common reference electrode was placed on the superior-medial tibial crest. Prewrap and elastic tape were applied over the electrodes to provide strain relief for the electrode cables.
A 10-second quiet baseline was taken of the MG and LG activities with the subjects supine on the plinth. Six-second maximum voluntary isometric contractions (MVICs) were then collected during a standing unilateral isometric maximum plantar flexion contraction. With the ankle positioned midway between neutral and full plantar flexion, subjects were instructed to “contract the calf muscles as hard as possible.” Once baseline and MVIC data were collected, subjects began heel-raise exercises according to their assigned order while the EMG data were saved for offline analysis.
Using the video data collected, the frame numbers corresponding to initiation, midpoint (concentric-eccentric phase transition), and completion of 5 repetitions were recorded for analysis. Initiation began when heels first elevated off floor; midpoint was the end of the CP; and repetition completion defined as when the heels rested on the floor. All EMG data reduction procedures were conducted offline using MatLab (The Mathworks Inc., Natick, MA) -based routines. Before calculation of EMG variables, EMG data from the MVIC and heel raises were full-wave rectified and smoothed by low-pass filtering at 10 Hz using a zero-phase-lag Butterworth filter. For the MVIC data, the mean amplitude was calculated for the middle 5 seconds and used for amplitude normalizing the heel-raise data. To reduce the heel-raise data, the data for each selected repetition were separated into concentric and eccentric phases and interpolated to 100 points. Ensemble averages within each phase, using the dominant leg, were calculated across the 5 selected repetitions with the mean amplitude of the ensemble average for each phase used for statistical analysis.
Determining sample size was based on the assumption that 10% differences between muscles and foot positions would be relevant. Using pilot data to estimate variability, conducting a power analysis with the adoption of alpha of 0.05 revealed that 20 participants would provide a minimum power of 0.78 for all main effects and interactions.
Separate 2-factor (muscle by foot position) repeated-measures analysis of variance (ANOVA) was used to statistically analyze each heel-raise phase (concentric, eccentric). Before conducting the ANOVAs, compliance with normality and sphericity assumptions was verified. Simple main effect post hoc tests, with a Bonferroni adjustment, between MG and LG at each foot position were used to examine the significant interactions. Statistical significance was considered at p ≤ 0.05.
All 20 subjects were able to successfully complete 12 repetitions under the 3 foot position conditions.
A significant muscle by foot position interaction (F[2,38] = 16.85, p < 0.001, partial η 2 = 0.470) was revealed for the concentric phase (Figure 4). There was no significant difference between MG and LG (p = 0.460) for the neutral position. During the IR position, significantly greater LG than MG (p = 0.003) activation occurred, whereas during the ER position, significantly greater MG than LG (p = 0.026) activation occurred. Additionally, LG activation during the IR position was significantly different than LG activation during the ER position (p = 0.014).
A significant muscle by foot position interaction (F[2,38] = 9.43, p < 0.001, partial η 2 = 0.332) was revealed for the eccentric phase (Figure 5). During the ER position, there was significantly greater MG than LG (p = 0.019) activation. There were no significant differences between MG and LG for the neutral (p = 0.108) and IR (p = 0.564) positions.
The current results support the notion that altering foot position during the heel-raise exercise will prompt varying degrees of MG and LG activation. Heel raises using a neutral stance elicited similar levels of MG and LG activation during both the concentric and eccentric phases. During both the concentric and eccentric phases, ER prompted significantly greater MG activity compared with the LG, whereas the concentric phase with IR provoked significantly greater LG activity compared with MG. Although this study cannot predict whether muscle-activation differences between foot positions will translate into greater adaptations, they do provide some initial objective evidence upon which practitioners can base gastrocnemius exercise-selection decisions.
Previous studies have investigated MG- and LG-activation differences during isometric (8,10) and isotonic (24) nonweight bearing and isometric weight-bearing (9) plantar flexion. During weight-bearing plantar flexion (9), the MG contributed a significantly greater percentage of the total EMG signal than the LG at the lowest resistance level (30% body weight). As the percentage of body weight increased, the MG-LG activation difference became significantly less to the point at which no significant difference existed during maximal isometric plantar flexion. Although the methodologies were quite different (i.e., the current investigation used a dynamic plantar flexion contraction, whereas the former used an isometric plantar flexion contraction), our results also yielded no significant difference between the MG and LG during the neutral foot position. This result was expected because the cross sectional area of the MG is twice the size of the LG (14). Collectively, the 2 studies would suggest that the MG and LG are equally activated in a standing neutral stance plantar flexed position during both isometric and dynamic contractions at intensities above body weight. Furthermore, the non-weight-bearing results of previous studies (6,10,24) also support no significant activation differences between the MG and LG during plantar flexion.
To obtain the IR and ER positions, we asked subjects to point their toes inward and outward as much as possible while completing the repetitions with the knee extended. Our directions prompted subjects to produce combinations of ankle, knee, and hip-joint rotations. We did not quantify how much rotation each subject selected or maintained across the ankle, knee, and hip joints throughout a set or the level of inversion and eversion achieved at the ankle. Normative data concerning active rotation limits between the ankle, knee, and hip joints (19,21) show that the hip demonstrates the greatest range of motion followed by the ankle and then the knee. Although the range of motion data is only provided for non-weight-bearing assessments, it is reasonable to assume that the rank order would remain the same (hip > ankle > knee). Subjectively, several subjects reported that initiating and maintaining the IR position was more difficult than the ER position. These subjective reports are consistent with the observation that active hip range of motion demonstrates approximately 10° greater for external rotation than internal rotation when the hip is 0° extension (25).
With respect to the IR and ER heel raises, the effect of hip rotation, independent of ankle and knee rotation, would cause the line of force being projected through the ankle joint to shift laterally during IR and medially during ER. These shifts could partially explain the significantly greater LG activation during IR and the significantly greater MG activation during ER. Supporting this speculation is a report in which healthy subjects demonstrated statistically smaller vastus medialis oblique to vastus lateralis activation ratios during step ups and stepdowns with the leg in external rotation compared with neutral or internal-rotation positions (20). It is also very plausible that ankle and knee rotation occurred concurrently with hip rotation in our study. We speculate that the combined rotation of the ankle and the knee during the IR and ER foot positions altered several MG and LG architectural features such as line of action, angle of pennation, and fascicle lengths. Altering these architectural features may have influenced the MG and LG force-generating capabilities, which in turn could explain the changes in muscle activations observed.
Several factors related to our study design need to be considered with regard to the generalizability of our results. First, subject-inclusion criteria to this study consisted of moderately active college-aged men and women with experience of performing heel-raise exercises and no history of musculoskeletal injury or pathology that could have influenced MG and LG activation. Whether similar results would be obtained with other populations such as older adults or persons with gastrocnemius pathology will need to be further investigated. Secondly, although heel raises are commonly performed with shoes, in the absence of standardized shoes, the heel raises in the current study were performed unshod. All heel raises in the current study were also performed free standing with 130-135% body mass. Because the MG-LG-activation differences appear to be related to the level of activation during isometric plantar flexion (9), we can only conservatively generalize our results to standing heel raises performed with 130-135% body mass. Thirdly, our methods asked the subjects to internally and externally rotate their legs as far as they could. Because we did not standardize or quantify the leg rotation, we cannot define how much internal and external rotation is needed to elicit the changes in activation we identified. Finally, the heel raises in our study were performed free standing using a 3.81-cm block height. We can only speculate that the effect foot position has on MG and LG activation would be larger as a result of not having to rely on the MG and LG muscles, and secondary muscles such as posterior tibialis and peroneals, for balance. These latter 4 factors, unshod feet, external load magnitude, self-selected internal and external rotation, and free-standing exercise all represent recommendations for future research. Finally, including a 3-dimensional kinematic analysis in future heel-raise research is recommended to quantify the ankle, knee, and hip rotations accompanying the internal and external rotation variations, and provide an insight regarding the underlying sources of muscle-activation differences yielded in the current investigation.
Heel raises are often used by persons interested in increasing gastrocnemius muscle size, strength, and power. The results of this study provide some initial support to the common practice of using different foot positions during the heel-raise exercise in an attempt to promote maximal adaptations in the MG and LG. Specifically, during the free-standing heel-raise exercise with 130-135% body mass, it appears that using an ER foot position prompts MG activation, whereas using an IR foot position prompts LG activation. Whether these activation differences translate into greater MG and LG training adaptations, or whether the results extend to other variations of heel-raise exercise (i.e., machine, seated), loading conditions (>130-135% body mass), and block height (>3.81 cm) remains to be studied.
1. Alway, S, Macdougall, J, Sale, D, Sutton, J, and McComas, A. Functional and structural adaptations in skeletal muscle of trained athletes. J Appl Physiol
64: 1114-1120, 1988.
2. Bezodis, I, Kerwin, D, and Salo, A. Lower-limb mechanics during the support phase of maximum-velocity sprint running. Med Sci Sports Exerc
40: 707-715, 2008.
3. Dorel, S, Couturier, A, and Hug, F. Intra-session repeatability of lower limb muscles activation pattern during pedaling. J Electromyogr Kinesiol
18: 857-865, 2008.
4. Ebben, W, Feldmann, C, Dayne, A, Mitsche, D, Alexander, P, and Knetzer, K. Muscle activation during lower body resistance training
. Int Sports Med
30: 1-8, 2009.
5. Fatouros, I, Jamurtas, A, Leontsini, D, Taxildaris, K, Aggeloursis, N, Kostopoulos, N, and Buckenmeyer, P. Evaluation of plyometric exercise training, weight training and their combination on vertical jumping performance and leg strength. J Strength Cond Res
14: 470-476, 2000.
6. Fiebert, I, Correia, E, Roach, K, Carte, M, Cespedes, J, and Hemstreet, K. A comparison of EMG activity between the medial and lateral heads of the gastrocnemius muscle during isometric plantar flexion
contractions at various knee angles. Isokinet Exerc Sci
6: 71-77, 1996.
7. Fiebert, I, LeBlanc, W, McGuane, S, Schnoes, C, and Strickland, K. The relationship of electromyographic activity and force of the vastus medialis oblique and vastus lateralis muscles during maximal isometric knee extension contractions. Isokinet Exerc Sci
2: 116-123, 1992.
8. Fiebert, I, Pahl, C, Applegate, E, Spieholz, N, and Beernik, K. Medial-lateral hamstring electromyographic activity during maximum isometric knee flexion at different angles. Isokinet Exerc Sci
6: 157-162, 1996.
9. Fiebert, I, Spieholz, N, Applegate, B, Crabtree, F, Martin, L, and Parker, K. A comparison of iEMG activity between the medial and lateral heads of the gastrocnemius muscle during partial weight bearing plantarflexion contractions at varying loads. Isokinet Exerc Sci
8: 65-72, 2000.
10. Fiebert, I, Spieholz, N, Applegate, E, Carbone, M, Gonzales, G, and Gorack, W. Integrated EMG study of the medial and lateral heads of the gastrocnemius during isometric plantar flexion
with varying cuff weights loads. J Back Musculoskel Rehab
11: 19-26, 1998.
11. Flanagan, S, Song, J-E, Wang, M-Y, Greendale, G, Azen, S, and Salem, G. Biomechanics of the heel-raise exercise. J Aging Phys Activ
13: 160-171, 2005.
12. Freriks, B, Hermens, J, Disselhorst-Klug, C, and Rau, G. The recommendations for sensors and sensor placement procedures for surface electromyography. In: SENIAM 8: European Recommendations for Surface ElectroMyoGraphy
. Hermens, H, Maerletti, R, Hagg, GG, Stegeman, D, Blok, J, Rau, G, and Disselhorst-Klug, C, eds. Enschede, The Netherlands: Roessingh Research and Development, 1999.
13. Friedrich, J and Brand, R. Muscle fiber architecture in the human lower limb. J Biomech
23: 91-95, 1990.
14. Fukunaga, T, Roy, R, Shellock, F, Hodgson, J, Day, M, Lee, P, Kwong-Fu, H, and Edgerto, W. Physiological cross-sectional area of human leg muscles based on magnetic resonance imaging. J Orthop Res
10: 926-934, 1992.
15. Gazendam, M and Hof, A. Averaged EMG profiles in jogging and running at different speeds. Gait Posture
25: 604-614, 2007.
16. Kawakami, Y, Ichinose, Y, and Fukunaga, T. Architectural and functional features of human triceps surae muscles during contraction. J Appl Physiol
85: 398-404, 1998.
17. Kingma, J, Knikker, R, Wittink, H, and Takken, T. Eccentric overload training in patients with chronic Achilles tendinopathy: A systematic review. Br J Sports Med
41: e3, 2007.
18. Maganaris, C. Force-length characteristics of the in vivo human gastrocnemius muscle. Clin Anat
16: 215-223, 2003.
19. Magee, D. Orthopedic Physical Assessment
. Philadelphia, PA: Saunders, 2002.
20. Miller, JP, Sedory, D, and Croce, RV. Leg rotation and vastus medialis oblique/vastus lateralis electromyogram activity ratio during closed chain kinetic exercises prescribed for patellofemoral pain. J Athl Train
32: 216-220, 1997.
21. Neumann, D. Kinesiology of the Musculoskeletal System
. St. Louis, MI: Mosby, Inc, 2002.
22. O'Sullivan, SP and Popelas, CA. Activation of vastus medialis obliquus among individuals with patellofemoral pain syndrome. J Strength Cond Res
19: 302-304, 2005.
23. Shalabi, A, Krostoffersen-Wilberg, M, Scensson, L, Aspelin, P, and Movin, T. Eccentric training of the gastrocnemius-soleus complex in chronic Achilles tendinopathy results in decreased tendon volume and intratendinous signal as evaluated by MRI. Am J Sports Med
32: 1286-1296, 2004.
24. Signorile, J, Applegate, B, Duque, M, Cole, N, and Zink, A. Selective recruitment of the triceps surae muscles with changes in knee angle. J Strength Cond Res
16: 433-439, 2002.
25. Simoneau, G, Hoenig, K, Lepley, J, and Papanek, P. Influence of hip position and gender on active hip internal and external rotation. J Orthop Sports Phys Ther
28: 158-164, 1998.