Randomized Controlled Trial of Gastrocnemius Muscle Analysis Using Surface Electromyography and Ultrasound in Different Striking Patterns of Young Women’s Barefoot Running

1. Introduction

Running is an effective exercise that enhances cardiorespiratory function, reduces the risk of coronary heart disease and diabetes, aids in weight loss, and alleviates psychological stress; however, it is also associated with a high incidence of bone and muscle injuries, particularly affecting the knees, ankles, Achilles tendons, and gastrocnemius muscles (Blyton et al., 2023). Running patterns are broadly classified into forefoot striking (FFS) and rearfoot striking (RFS) (De Clercq et al., 1994). The key difference between forefoot and rearfoot striking lies in the striking moment (Hobara et al., 2012; Lieberman et al., 2010). Habitual barefoot runners prefer forefoot striking, while shod runners typically use rearfoot striking, transitioning from heel to metatarsal contact (Cheung & Davis, 2011).

Humans walked and ran barefoot for millions of years, predominantly using forefoot or whole-foot striking (Hobara et al., 2012), where the metatarsal bone contacts the ground first (Holowka & Lieberman, 2018). Forefoot striking involves a more anterior contact point, with the ankle joint flexing and dorsiflexing to minimize support and collision forces, allowing pressure to shift and providing a landing buffer for runners (Divert et al., 2005). Habitual shoe-wearing runners typically land on their rearfoot before transitioning to full-foot contact, facilitated by the coordinated action of the triceps surae and Achilles tendon on the ankle, making rearfoot landing more common (Cheung & Davis, 2011).

Forefoot striking requires stronger leg muscles due to the increased load on the Achilles tendon and plantar flexors compared to rearfoot striking (Lieberman, 2012). During the loading response stage, muscle activity prepares for striking and subsequent ground contact (Divert et al., 2005). To protect against repeated ground reaction force impacts, plantar flexors and lower limb muscles activate in advance. High-intensity, repetitive lower limb movements in running can lead to tissue tension limits and increased injury risk from ground reaction forces (Encarnación-Martínez et al., 2023). Ground reaction force differences between striking patterns show that forefoot striking results in smaller collision forces and ground reaction forces, reducing damage risk compared to rearfoot striking, which involves higher forces and loading speeds (Murphy et al., 2013). Rearfoot striking’s greater vertical ground reaction force, absorbed by the heel pad and shoe cushioning, may cause injuries due to higher impact peaks (Lieberman et al., 2010). Forefoot striking, with greater ankle compliance, lowers vertical loading rates and reduces lower limb injury risks (Tome et al., 2006). Although shoe insoles can cushion ground reaction forces, rearfoot striking still poses a higher risk of lower limb injuries due to greater ground collision forces and restricted knee and ankle joint movements, leading to increased ground reaction forces (Stearne et al., 2014).

The gastrocnemius muscle of barefoot runners activates earlier than that of shod runners, enhancing gastrocnemius strength and energy storage for running. This muscle promotes forward and upward movement during the period between rearfoot and toe-off stages and rests during the swing period (Beierle et al., 2019). Studies on electromyographic (EMG) characteristics have shown higher activation levels in the gastrocnemius medialis (GM) and gastrocnemius lateralis (GL) under forefoot striking (Yong et al., 2014). Barefoot running systematically strengthens lower limb muscles, promoting ankle joint strength and adjusting leg stiffness to minimize vertical ground reaction force impacts (Kim et al., 2021). Shod running delays EMG activity after rearfoot contact and increases pre-contact muscle activity, affecting running gait, muscle activation, and explosive force (Diebal et al., 2012). Barefoot running features higher stride frequency, smaller step size, greater plantar flexion range, and shorter ground contact time compared to shod running (Squadrone & Gallozzi, 2009). Forefoot striking results in shorter step length and faster stride frequency, reducing ground contact time and impact loads.

Energy absorption mechanisms differ between striking modes. Rearfoot striking relies on relaxed plantar fascia and surrounding tissue to absorb and transfer energy, tightening the fascia and pushing the foot forward via tibial external rotation and gastrocnemius activation (Vieira et al., 2013). Forefoot striking generates plantar flexion torque in the ankle joint, countering external dorsiflexion torque and storing elastic potential energy during the pre-swing period (Kulmala et al., 2013). Forefoot striking demands higher ankle strength and longitudinal arch tension, while rearfoot striking increases knee joint pressure and injury risk (Ness et al., 2008). EMG evaluates muscle activity and fatigue during movement, providing parameters for muscle strength, activation, and fatigue, aiding in athlete selection and training. Ultrasound measures muscle morphology accurately in both dynamic and static states, assessing thickness, pennation angle, fiber length, and cross-sectional area. It plays a crucial role in diagnosing muscle and nerve diseases, guiding rehabilitation, and evaluating muscle function. Ultrasound measures muscle morphological parameters like pinnation angle (PA), muscle thickness (MT), and fiber length, which are crucial for evaluating muscle contraction and function. PA reflects force transmission efficiency, with optimal angles enhancing muscle strength. Muscle thickness and fiber length are indicators of muscle content and explosive force, aiding in fatigue assessment and muscle function analysis (Franchi et al., 2018).

This study examines EMG data and ultrasound images of 20 young women with different running striking modes to explore muscle activity and morphology differences between forefoot and rearfoot striking in barefoot running. By comparing these data, the study aims to understand the gastrocnemius muscle’s role in running mechanics, informing training and injury prevention strategies. Despite extensive research on running mechanics, there remains a significant gap in understanding the biomechanical and muscular differences between forefoot and rearfoot striking, especially in barefoot running. Most studies focus on shod running, lacking data on how striking patterns affect muscle activity and morphology in barefoot conditions. This study addresses this by analyzing the gastrocnemius muscle’s role and activation in both striking styles. Utilizing EMG and ultrasound technologies, it provides insights into muscle activation and morphology differences, offering practical applications for enhancing training regimens, reducing injury risks, and improving athletic performance.

2. Methods

Twenty healthy females participated in this experiment, divided into two groups: ten habitual forefoot strikers (experimental group) and ten habitual rearfoot strikers (control group). Participants were recruited through local running clubs and online advertisements, ensuring a diverse sample of recreational runners who ran at least 10 km per week and had no history of deliberate barefoot running. They all had normal foot shapes, without flatfoot or high arches, and no history of lower limb injury within the past year. Detailed exercise histories were self-reported and verified through interviews. All participants provided informed consent, and ethical approval was from the local ethic committee (The ethics approval number for the experiment is UAHPEC20628). Basic information about participants is shown in Table 1.

Running strike patterns were recorded and classified using the Vicon motion capture system (Oxford Metrics Ltd., Oxford, United Kingdom). To classify the subjects into forefoot and rearfoot striking groups, we placed markers on their lower extremities and tracked their positions. Markers were placed on near the heel and toe. During a static standing pose, we established a baseline vertical difference between these markers. This difference was then measured at initial contact during running. A difference of 40 mm or less indicated a forefoot striker, while a difference greater than 70 mm indicated a rearfoot striker (Yong et al., 2014). The research proposal was developed in accordance with the CONSORT 2010 guidelines for randomized pilot and feasibility trials, following consultations with the Standard Protocol Items: Recommendations for the Reporting Trials (Eldridge et al., 2016).

A Q6-model portable ultrasonic diagnostic instrument (Q6, China) with a 5–10 MHz ultra-wideband probe was used for measurements. Before the experiment, subjects were briefed on the procedure, including gait and posture requirements. They were asked about any muscle discomfort, and if present, the measurements were postponed. Participants wore loose clothing to facilitate marking the gastrocnemius muscle position. Ultrasound measured the thickest part of the medial and lateral gastrocnemius muscles of the right calf to obtain muscle thickness and pennation angle. To minimize error and improve accuracy, the same operator conducted all measurements, and muscle thickness was manually marked. Leg length was determined by flexing the knee and ankle. The ultrasound probe, coated with a coupling agent, was placed below the popliteal fold of the right calf, as shown in Figure 1.

Figure 1

After measuring the ultrasound properties of the gastrocnemius muscle, the skin was cleaned with 75% medical alcohol to remove grease, sweat, and dead skin. Shaving was performed to ensure proper electrode adhesion and prevent data errors. Electrodes were placed in the middle of the thickest part of the muscle protuberance, aligned with the gastrocnemius fibers. Each subject’s ultrasound data was recorded before running to determine EMG electrode placement in subsequent experiments. The same operator placed all EMG electrodes to reduce measurement error. Medical tape was used to secure the electrodes and prevent displacement during high-speed running. Subjects performed stretching exercises to prepare their muscles before the experiment. Wireless EMG devices (Delsys, Boston, MA, USA) were attached to the marked positions after measurement. While subjects ran on a treadmill, EMG data from the medial and lateral gastrocnemius muscles were recorded. Sensors were placed along the longitudinal midline of the gastrocnemius muscle, avoiding the muscle edge and tendon to ensure accurate readings.

The experiment utilized a 16-channel wireless EMG measurement system developed by Delsys. Subjects began by walking barefoot on a treadmill at an initial speed of 4 km/h, increasing by 0.5 km/h every 30 seconds until reaching 7.5 ± 0.5 km/h in the fifth minute. After one minute of stable running, EMG data were recorded for one minute. The collected EMG and ultrasound imaging data were analyzed. EMG signals were collected and processed using Emgworks Acquisition software (Delsys, Boston, MA, USA) at a frequency of 1000 Hz. The root mean square (RMS) amplitude and median frequency of the EMG signals from the medial and lateral gastrocnemius muscles were analyzed. Gait cycle periods were defined based on average myoelectric amplitude: pre-swing (0–15% and 30–50% of the gait cycle), swing (30–80% of the gait cycle), late swing/pre-activation (80–100% of the gait cycle), and middle and late swing (70–80% of the gait cycle). The 10-second running data segments were processed using a Butterworth filter for high-pass (30 Hz) and low-pass (10 Hz) filtering. Full-wave rectification was applied, and data were standardized by dividing by the maximum value. The root mean square amplitude was calculated for different gait stages. Statistical analysis compared muscle activation and strength between groups. Median frequency values were calculated with a moving window length of 0.07 seconds, averaged over three measurements. To account for individual differences, EMG eigenvalues were standardized using the peak muscle value during exercise, allowing for comparable EMG values across subjects. Ultrasound imaging data were manually processed. Superficial and deep fascia positions were marked, and muscle thickness and pennation angles were measured using ImageJ software 1.48 (National Institute of Health, USA) for analysis and comparison.

Ultrasound imaging and surface electromyography data from habitual forefoot and rearfoot striking groups were analyzed to assess significant differences between the groups. Statistical analyses were conducted using independent samples t-tests following the Shapiro-Wilk test for normality. If the data were not normally distributed, the Mann-Whitney U test was applied. Results are presented as mean ± standard deviation. All preprocessed data were analyzed using SPSS 22.0 software (SPSS Inc., Chicago, IL, USA). The significance levels for the statistical tests were set at P < 0.05 (denoted by *), P < 0.01 (denoted by **), and P < 0.001 (denoted by ***).

3. Results

The analysis of the GL muscle root mean square (RMS) across different phases between the forefoot striking (FFS) and rearfoot striking (RFS) groups indicates that most data follow a normal distribution, except for the 70–80% phase in the RFS group, which was close to non-normality. Independent t-tests for the normally distributed phases revealed significant differences between the FFS and RFS groups in the 30–50%, 80–90%, and 90–100% phases, while no significant difference was found in the 0–15% phase. For the 70–80% phase, where normality was in question, the Mann-Whitney U test confirmed a significant difference between the groups. These results suggest notable variations in GL muscle behavior between the FFS and RFS groups, particularly in the later phases of movement. Similarly, the analysis of the GM muscle across different phases revealed that most phases followed a normal distribution, except for the 30–50% phase in the RFS group, which showed borderline non-normality. T-tests for the normally distributed phases indicated significant differences in the 70–80%, 80–90%, and 90–100% phases, while the 0–15% phase showed no significant difference. The Mann-Whitney U test for the 30–50% phase confirmed a significant difference, suggesting notable variations in GM muscle activity between the FFS and RFS groups, particularly in the mid and later phases of movement. Changes in the RMS amplitude value of the gastrocnemius muscle throughout the gait cycle are shown in Table 2.

In the pre-swing stage (30–50% of the gait cycle), the RMS amplitudes of both the medial and lateral heads of the gastrocnemius muscle are significantly higher in the forefoot striking group compared to the rearfoot striking group, with differences marked as statistically significant. During the 70–80% phase, the medial head in the forefoot striking group also shows a significantly higher RMS amplitude, and this trend continues in the late swing/pre-activation stages (80–90% and 90–100%), where both the medial and lateral heads demonstrate substantial increases in the forefoot striking group. Across these phases, the overall RMS amplitude for the forefoot striking group remains consistently higher than that of the rearfoot striking group, reflecting significant variations in gastrocnemius muscle activity between the two groups, particularly in the later stages of the gait cycle. The RMS changes over the 100% gait cycle are shown in Figure 2.

Figure 2

Table 3 presents the median frequency values under different striking modes. After Shapiro-Wilk normality tests, the data for GM and GL in both the FFS and RFS groups appear to follow a normal distribution. For the GM, the median frequency value in the forefoot striking group is 51.27 ± 17.09 Hz, while it is 47.00 ± 16.14 Hz in the rearfoot striking group, with no significant difference between the groups. For the GL, the median frequency is 55.82 ± 12.94 Hz in the forefoot striking group and 52.12 ± 17.58 Hz in the rearfoot striking group, showing no significant difference. Although the differences are not statistically significant, the median frequency values for the GM and GL are slightly higher in the forefoot striking group compared to the rearfoot striking group, as depicted in Figure 3.

Figure 3

Except for the GM thickness, all variables from the Shapiro-Wilk test have p-values greater than 0.05, indicating normal distribution. For the non-normally distributed GM thickness data, the Mann-Whitney U test yielded a U statistic of 33.5 and a p-value of 0.0856. Table 4 and Figure 3 show the differences in muscle thickness, pennation angle, and muscle fiber length of the gastrocnemius muscle between the forefoot and rearfoot striking groups. In terms of muscle thickness, a significant difference is observed in the GM, where the thickness is greater in the rearfoot striking group compared to the forefoot striking group. The GM thickness in the rearfoot striking group is 1.89 cm, significantly larger than the 1.63 cm measured in the forefoot striking group. Although there is no significant difference in GL thickness, the GL in the rearfoot striking group is slightly thicker (1.49 cm) than in the forefoot striking group (1.27 cm). In terms of muscle fiber length, no significant difference is observed between the groups, though the muscle fiber length in the forefoot striking group (4.96 cm) is shorter than in the rearfoot striking group (5.36 cm). The muscle fiber length of the GL in the rearfoot striking group is 6.24 cm, significantly greater than the GM length in the forefoot striking group (4.61 cm). No significant difference in pennation angle is found between the two groups.

4. Discussion

Muscle activation precedes muscle force generation, impacting force generation during the loading response stage (Dietz et al., 2001). In this experiment, during the early gait cycle (pre-swing period), the myoelectric activity of the gastrocnemius muscle shows little difference between the forefoot striking group and the rearfoot striking group, likely due to the delay between muscle activation and force generation. After the pre-swing period, the RMS of the GM and GL in the forefoot striking group are greater than those in the rearfoot striking group, indicating higher muscle requirements for the gastrocnemius in the forefoot striking group. During the suspension stage of running, the EMG amplitude of both striking patterns significantly increases, suggesting prominent gastrocnemius muscle activity. This increase may be due to the calf force when the lower limb moves upward from the ground. However, the RMS amplitude of EMG on GM and GL decreases significantly at the swing stage of running, approaching levels of muscle inactivity, consistent with previous findings that gastrocnemius muscle force plays a key role in walking (Beierle et al., 2019).

At the pre-activation and pre-striking stages, the RMS amplitude in both groups increases slightly. Studies on the EMG characteristics of different lower limb muscles indicate that GM and GL activation increases at the end of the swing period in the forefoot striking group (Yong et al., 2014).This activation before landing during ankle plantar flexion increases gastrocnemius muscle stiffness to prepare for ground collision force, thus reducing joint and ligament damage from the load and collision force. Increased leg stiffness may lead to a larger ground reaction force and reduced lower limb activity range, potentially causing injuries such as knee osteoarthritis and stress fractures (Altai et al., 2024).

The range of plantar flexion decreases during the suspension period, significantly reducing gastrocnemius muscle amplitude. Studies show that the RMS amplitude value of lower limb muscles more than doubles at the loading response stage compared to the pre-swing period, indicating a significant increase in muscle activation (Dietz et al., 2001). This experiment’s data corroborate this, with RMS values of GM and GL before landing significantly higher than in the pre-swing period in both groups, showing that calf muscles are fully activated before landing to reduce the impact of ground reaction force.

The gastrocnemius muscle plays a crucial role in preparing for foot-ground contact at the late swing stage, specifically during the pre-activation period (80–90% and 90–100% of a gait cycle). The main goal of muscle activity is to prepare for landing and subsequent ground contact. Pre-activation enhances muscle activity, affecting and regulating leg stiffness to prepare for ground contact (Chumanov et al., 2012). In this experiment, the root mean square amplitude in the forefoot striking group is significantly higher than in the rearfoot striking group, indicating better gastrocnemius muscle activation in the forefoot striking group. Higher activation levels may imply stronger muscle stiffness, better buffering the impact of ground reaction force.

At the mid-swing stage (70–80% of a gait cycle), increased calf muscle activity may play a key role in promoting landing position change, reducing knee joint torque and energy absorption (Heiderscheit et al., 2011). This experiment shows that the RMS amplitude of GM and GL in the forefoot striking group is higher than in the rearfoot striking group, suggesting that increased gastrocnemius muscle activity in the forefoot striking group reduces knee joint torque, aiding landing position adjustment in running. At the early stage of foot-ground contact, lower limb extensor muscles significantly increase their activation due to the large impact of ground reaction force. Muscle stiffness may relieve this impact, causing EMG delay due to pre-activation and central nervous system judgment (Campanini et al., 2020). Studies find the peak ground collision force often appears within 40 milliseconds after ground contact, too short for full muscle excitation (Honert et al., 2022). Thus, increasing pre-activation levels is crucial for improving muscle force and stiffness.

Pre-activation activates muscle joints, tendon spindles, and other receptors, transforming elastic energy into forward driving force, improving muscle conversion and utilization rates (Seiberl et al., 2021). Increased muscle fiber length leads to increased muscle mass, though it does not increase the physiological cross-sectional area (Kawakami et al., 1998). Longer muscle fibers contain more sarcomeres, affecting force generation speed and benefiting sprint performance with strong explosive force. As muscle shortening speed increases with decreased force, longer muscle fibers may output greater force at the same shortening rate, improving sprint performance (Wickiewicz et al., 1983).

Muscle fiber growth produces greater maximum speed and force, enhancing sprint speed. Sprinters have longer muscle fibers and smaller pinnation angles than ordinary people, contributing to their explosive muscle force (Abe et al., 2001). Elite sprinters exhibit significantly longer muscle fibers than elite long-distance runners or untrained controls, likely due to more intensive sprint practice (Abe et al., 2000). In this experiment, the muscle fiber length of GL in the forefoot striking group is significantly smaller than in the rearfoot striking group, indicating better explosive force in the rearfoot striking group. GL is the longest muscle in the triceps of the leg, with structural characteristics determining the difference in maximum force and shortening speed between GM and GL (Kawakami et al., 2006; Wickiewicz et al., 1983). Despite similar muscle fiber types, GL’s greater fiber bundle length provides greater shortening potential (Monte et al., 2023). GM’s shorter muscle fibers and larger pinnation angle accommodate more fibers in a certain volume, affecting force generation. GL’s larger muscle thickness and smaller pinnation angle result from greater muscle fiber length and cross-sectional area (Geremia et al., 2019).

This experiment shows that while there is no significant difference between the groups, GL fiber length in the rearfoot striking group is significantly larger, suggesting stronger muscle explosive force in the rearfoot striking group. Muscle activation and explosive force are key factors in running, with sufficient activation affecting lower limb muscle stiffness. Increased muscle stiffness resists ground reaction force and reduces joint load. Studies show significant differences in muscle thickness and fiber length between dominant and non-dominant legs in some athletes, driven by increased muscle mass after training. Increased muscle mass leads to increased muscle thickness and fiber length (Kruse et al., 2021; Xu et al., 2024). In this experiment, GM muscle thickness in the rearfoot striking group is significantly greater than in the forefoot striking group, indicating larger muscle mass and greater muscle content used in running. Therefore, it can be inferred that the rearfoot striking group’s explosive force is better than that of the forefoot striking group, with longer gastrocnemius muscle fiber and greater muscle thickness.

One of the limitations of this study is that inferring greater muscle explosive power based on muscle thickness and fiber length alone may not be entirely sufficient. Although there is a well-established correlation between these anatomical characteristics and explosive power, they are not direct measures of it. Therefore, future research should incorporate additional biomechanical indicators, such as direct measurements of muscle explosive power through tools like force plate testing, to more comprehensively understand the relationship between muscle structure and function across different running styles.

5. Conclusion

At the loading response stage, no significant difference was found in gastrocnemius RMS amplitude between the forefoot striking (FFS) and rearfoot striking (RFS) groups, likely due to a delay between muscle activation and force generation. However, from the mid-swing to late swing stages, the FFS group exhibited consistently higher RMS amplitudes, indicating increased muscle activation and stiffness in response to ground reaction forces, which could contribute to greater fatigue susceptibility. In contrast, the RFS group showed greater muscle thickness and fiber length, suggesting superior muscle explosive force. While these anatomical characteristics are correlated with explosive power, they do not directly measure it. Thus, future research should incorporate direct biomechanical assessments, such as force plate testing, to better understand how muscle structure and function relate to explosive power in different running styles.

Ethics and Consent

The study was conducted in accordance with the Declaration of Helsinki and approved by the Ethical Committee of University of Auckland (UAHPEC20628). Legal guardians of all participants provided informed written consent to participate in the study prior to the study.

Funding Information

This work was supported by the China Scholarship Council (CSC) scheme for Yuwei Liu.

Competing Interests

Justin Fernandez is an Editorial Board Member for [Physical Activity and health] and was not involved in the editorial review or the decision to publish this article.