Electromyographic analysis of fatigue and synergistic activation in striking and supporting legs during continuous side-kicks in elite Sanda athletes Authors: Qiang Wu, Yang Liu, Bin Zhao Published: 2025 ABSTRACT AbstractThis study aimed to investigate the functional roles and fatigue effects of the striking and supporting legs in consecutive stationary side-kick techniques among Sanda fighters. By analyzing muscle activation (root mean square, RMS), this study provides insights for targeted muscle training in Sanda. Surface electromyography (sEMG) was utilized to collect data from 10 Chinese elite-level Sanda athletes (height: 178.5 ± 4.3 cm, weight: 72.1 ± 5.6 kg, training experience: 7.8 ± 1.9 years; 4 international master athletes, 6 national master athletes) performing continuous static side-kick techniques. Vicon synchronization equipment was used to determine the start and end points of the movement. Thirty seconds of raw sEMG data were preprocessed and segmented into 1-second intervals. A sliding window approach with feature averaging was employed to calculate the root mean square (RMS) values for 50-millisecond windows, followed by the computation of mean RMS values per 1-second interval. The fatigue characteristics of major muscle groups in the lower limbs (kicking leg and supporting leg) were evaluated by analyzing the mean RMS values across different phases over time: the initial phase (1–5 s), the middle phase (initial: 1–5 s, middle: 15–19 s, final: 26–30 s). Regarding fatigue accumulation, elite Chinese Sanda fighters exhibited significant RMS reductions in the striking leg’s primary muscle groups (gastrocnemius, vastus medialis, and gluteus maximus) over time (change rates: -5.29% to -7.50%, p < 0.01), indicating pronounced fatigue. In contrast, the supporting leg’s major muscle groups showed stable RMS changes (change rates: -3.52% to -3.97%, p > 0.05), suggesting minimal fatigue accumulation during the task. For muscle activation, the RMS values of the striking leg’s primary muscles were significantly higher than those of the supporting leg (p < 0.01). The tibialis anterior in the supporting leg exhibited greater activation compared to the striking leg (p < 0.05), while hamstring activation showed no significant difference between legs (p > 0.05). Correlation analysis revealed that the striking leg’s force production is driven by synergistic activity among the vastus medialis, rectus femoris, vastus lateralis, and gluteus maximus (r = 0.74 ~ 0.78, p = 0.007 ~ 0.014), highlighting their critical role in hip and knee joint force generation. In contrast, the supporting leg relied on a strong correlation between the gastrocnemius and tibialis anterior (r = 0.81, p < 0.01) for ankle joint stability and a moderate correlation between the gluteus maximus and gluteus medius (r = 0.65, p < 0.05) for hip balance, underscoring distinct muscle coordination patterns between the two legs. In Chinese elite Sanda fighters performing continuous in-place side-kick movements, the striking leg’s power-generating muscle groups—primarily the quadriceps and gluteus maximus—exhibit significant fatigue. Meanwhile, the supporting leg maintains stability through the tibialis anterior and gastrocnemius. These findings highlight potential strength asymmetries that could signal fatigue onset or even injury risk, offering valuable insights for coaching strategies and rehabilitation protocols. Training recommendations include plyometric exercises to enhance the explosive power of the quadriceps and gluteus maximus, resistance training to improve the endurance of the striking leg, and isokinetic training to strengthen the stability of the tibialis anterior and gastrocnemius in the supporting leg, thereby further optimizing the training regimen.Keywords: Sanda, Side-kick, Electromyography, Muscle activation, Synergistic contraction, Fatigue effectSubject terms: Anatomy, Health care, Physiology Introduction The side-kick technique, unique to Chinese Sanda martial arts1–3, serves as a critical method for gaining a competitive advantage by targeting the opponent’s head during offensive maneuvers4 and interrupting or terminating continuous attacks during defensive actions5. To secure victory in competitions, athletes must sustain repetitive kicking actions over extended periods, requiring the ability to maintain muscle strength and power. However, prior studies have predominantly focused on single kicks6, with extensive research on punches1,7, whip kicks8,9, and roundhouse kicks10, while investigations into the fatigue characteristics of continuous kicking remain limited. Existing studies have explored the fatigue properties of continuous kicks. For instance, an analysis of lower-limb kinematics and muscle activation in six junior karate athletes performing 20 consecutive side kicks revealed significant declines in knee flexion, hip abduction, knee extension angular velocity, and biceps femoris activation with increasing repetitions11. Motion capture technology further quantified knee joint angles and hip joint moments, elucidating the synergistic working patterns of lower-limb muscles12–14. Existing literature has reported the 30-second continuous jump test (CJ30s) as a method for assessing anaerobic power in combat sport athletes15. Additionally, research on 10 consecutive soccer instep kicks reported significant reductions in ball velocity, shank and thigh angular velocities, and moments starting from the fifth kick, indicating that lower-limb fatigue substantially impacts biomechanical performance16. However, differences between soccer kicks and Sanda side kicks limit the direct applicability of soccer-related findings to martial arts, necessitating further research into the fatigue mechanisms of continuous side kicks in Sanda. Surface electromyography (sEMG) is a non-invasive, rapid, and convenient monitoring method widely reported in the field of sports science for assessing physical conditions17. By quantifying muscle activation levels and synergistic contractions, sEMG provides a reliable tool for elucidating biomechanical mechanisms during movement18,19. Studies have utilized wearable devices to accurately evaluate the performance of leg techniques in Sanda athletes9. For example, surface EMG has been used to analyze differences in whip kick techniques among Sanda athletes of varying skill levels20. Further research on the five phases of the side-kick technique has highlighted the critical contributions of the rectus femoris and gluteus maximus during force generation21. Although existing studies have partially elucidated the fatigue patterns of continuous kicking11, significant research gaps remain. Current studies on sEMG fatigue characteristics of continuous kicking primarily focus on scenarios with kick durations ≤ 5 s (e.g., single kicks or 20 consecutive kicks), with insufficient attention to the more common scenario of muscle performance under fatigue during “continuous side kicks > 5 seconds” (e.g., 6–10 consecutive kicks). In fact, when side-kick durations exceed 8 s, significant changes in energy supply patterns and neural control mechanisms of lower-limb muscles may occur, resulting in sEMG fatigue characteristics fundamentally different from those of short-duration continuous kicks. Therefore, this study involved elite Chinese Sanda athletes performing 30 s of continuous in-place side-kick techniques, with analyses conducted at 5-second intervals to investigate muscle fatigue characteristics during the early, middle, and late stages of fatigue. Materials and methods Subjects Ten elite Chinese male Sanda fighters (age: 22.3 ± 2.1 years; height: 178.5 ± 4.3 cm; weight: 72.1 ± 5.6 kg; training experience: 7.8 ± 1.9 years; comprising 4 international-level masters and 6 national-level masters) were recruited as participants. All Sanda fighters had no history of lower limb injuries and provided informed consent prior to participation. This study was approved by the Ethics Committee of Chengdu Sport University(No. 2025 − 115). All methods were performed in accordance with the relevant guidelines and regulations (Table 1). Table 1List of information on elite-level Sanda fighters in China.numberNameSports levelTraining period (years)Best performanceS1Ma YiguInternational Elite12World University Games Sanda ChampionshipS2Lu HaitaoNational Elite10Second place in the National Wushu Sanda ChampionshipS3Zhou ShijieNational Elite13National Wushu Sanda Championship ChampionS4You XiongInternational Elite12World Sanda Championships ChampionS5He FengInternational Elite12Hangzhou 19th Asian Games Wushu Sanda ChampionS6Yu JingshengNational Elite13National Sanda ChampionshipS7Jiang HaidongInternational Elite17Hangzhou 19th Asian Games Wushu Sanda ChampionS8Feng JunjieNational Elite5World Youth Wushu Sanda Championships ChampionS9Zhan BohanNational Elite5Third place in China Sanda ChampionshipS10Wang YuhaoNational Elite6Third place in China Sanda Championship List of information on elite-level Sanda fighters in China. World University Games Sanda Championship Second place in the National Wushu Sanda Championship National Wushu Sanda Championship Champion World Sanda Championships Champion Hangzhou 19th Asian Games Wushu Sanda Champion National Sanda Championship Hangzhou 19th Asian Games Wushu Sanda Champion World Youth Wushu Sanda Championships Champion Third place in China Sanda Championship Third place in China Sanda Championship Experimental procedure Before the experiment, Sanda fighters were comprehensively briefed on the experimental procedures and precautions to ensure complete understanding. A 10-minute warm-up session, comprising jogging, stretching, and striking training, was conducted to activate the relevant muscle groups. Subsequently, surface electromyography electrodes were attached to specific muscle sites on the striking and supporting legs of the Sanda fighters. The placement of the electrodes followed the guidelines of the SENIAM (Surface Electromyography for the Non-Invasive Assessment of Muscles) protocol. Electrodes were placed by a certified professional in accordance with SENIAM guidelines specifically for the experimenters. The electrodes are gourd-shaped with a fixed inter-electrode distance. During the experiment, Sanda fighters performed continuous in-place side-kick movements for 30 s, striving to maintain standardized technique and consistent frequency (Fig. 1). The Vicon system was employed for synchronization to facilitate subsequent observation and verification of standardized movement patterns. By identifying the onset and offset of each movement, precise data points corresponding to the action period were extracted, ensuring accuracy and consistency in subsequent analyses. Schematic diagram of the continuous side-kick technique of Sanda fighters. “Experimenter Wu Qiang (Sanda World Champion and National Team Coach) and Zhao Bin (Sanda International Referee) supervised and provided guidance to ensure the quality of movements. As shown in Table 2, preliminary test results indicate that the overall homogeneity among the 10 elite Sanda athletes is high. With the exception of body weight, the coefficient of variation (CV) for all metrics is below 10%, demonstrating a high degree of consistency in the physical characteristics of this group. This uniformity provides a stable foundation for investigating the force exertion characteristics of consecutive side-kick techniques, effectively minimizing the impact of physical metric variations on electromyographic data. Table 2Homogeneity statistics of physical indicators.IndicatorUnitMean ± SDCV%RangeHomogeneity evaluationHeightcm174.91 ± 5.623.21%165.2 ~ 182.0High homogeneityWeightkg69.81 ± 8.3812.00%56.7 ~ 80.0Medium homogeneityShoulder widthmm78.3 ± 5.547.08%72 ~ 90High homogeneityElbow width (mean of left and right)mm71.57 ± 5.287.38%61.8 ~ 80.1High homogeneityWrist width(mean of left and right)mm55.01 ± 5.219.47%44.7 ~ 62.45High homogeneityLeg length(mean of left and right)cm88.6 ± 4.715.32%82.9 ~ 98.5High homogeneityKnee width (mean of left and right)mm110.67 ± 9.688.75%97.95 ~ 133.85High homogeneityAnkle width(mean of left and right)mm69.53 ± 5.718.21%61.6 ~ 79.45High homogeneity Homogeneity statistics of physical indicators. High homogeneity Medium homogeneity Shoulder width High homogeneity Elbow width (mean of left and right) High homogeneity Wrist width (mean of left and right) High homogeneity Leg length (mean of left and right) High homogeneity Knee width (mean of left and right) High homogeneity Ankle width (mean of left and right) High homogeneity Instruments and equipment Vicon Optical Motion Capture System: Twelve Vicon infrared high-speed cameras (Origin: United Kingdom, Sampling Frequency: 200 Hz) track 54 markers placed on joint points. By establishing a human biomechanical model, the system assists in analyzing the movement patterns of the Sanda side-kick technique during testing by repeatedly observing the accuracy of the movements, thereby improving the precision of the assessment (Fig. 1). Electromyography (EMG): A Noraxon wireless surface electromyography system (Brand: Noraxon, USA; Model: Ultium; 16 channels; sampling frequency: 2000 Hz) was utilized (Fig. 2). Noraxon wireless surface electromyography. This study monitored a total of 13 muscles across both limbs to investigate the biomechanical demands of side-kick kinematics. The striking leg involved seven muscles: gastrocnemius, tibialis anterior, biceps femoris (hamstrings), vastus medialis, rectus femoris, vastus lateralis, and gluteus maximus. The supporting leg included six muscles: gastrocnemius, tibialis anterior, biceps femoris (hamstrings), vastus medialis, gluteus maximus, and gluteus medius. Asymmetric Muscle Selection (7 vs. 6 Muscles): The selection of seven muscles for the striking leg and six for the supporting leg was informed by a preliminary experiment conducted prior to the main study. The preliminary results indicated that the root mean square (RMS) values of electromyography (EMG) signals for the supporting leg muscles exhibited smaller changes compared to those of the striking leg during side-kick execution. Given the limited number of EMG channels and the primary focus of this study on the striking leg’s biomechanical contributions, a greater number of muscles were prioritized for the striking leg to capture its more dynamic activation patterns. Alignment with Biomechanical Demands of Side-Kick Kinematics: The selected muscles align with the biomechanical requirements of the side-kick, which involves a complex interplay of lower limb movements including hip abduction, knee extension, and ankle stabilization. For the striking leg, the inclusion of vastus lateralis and rectus femoris captures the quadriceps’ role in powerful knee extension, while the gluteus maximus contributes to hip extension and external rotation. The gastrocnemius and tibialis anterior were selected to monitor ankle plantarflexion and dorsiflexion, respectively, which stabilize the foot during the kick. The biceps femoris was chosen for its involvement in hip extension and knee flexion, critical for initiating and controlling the kick’s trajectory. For the supporting leg, the gluteus medius was included to account for its role in maintaining pelvic stability and hip abduction, which are essential for balance during the dynamic motion of the side-kick. The shared muscles (gastrocnemius, tibialis anterior, biceps femoris, vastus medialis, and gluteus maximus) reflect their dual contributions to stability and force generation in both limbs, while the additional muscles in the striking leg address its greater kinematic complexity. Data processing and analysis Data processing Electromyographic Data Processing Method: Raw electromyographic (EMG) data were preprocessed using MATLAB software with a Butterworth bandpass filter (20–450 Hz), rectification, and smoothing. The preprocessed data were then segmented into 30 one-second intervals, with the average root mean square (RMS) value calculated for each segment. A 50-millisecond sliding window was applied to compute RMS values, from which the mean RMS value for each 1-second interval was extracted. Normalization was performed using the maximum 1-second mean RMS value within the 30-second duration as the reference. Finally, the 30 data points were divided into early, middle, and late phases, and the mean values for each phase were calculated for statistical analysis (Fig. 3). Raw electromyography processing. ① EMG extraction refers to the RMS index: (N data points, sampling interval Δ) signal Data [i]. EMG extraction refers to the RMS index: (N data points, sampling interval Δ) signal Data [i]. The RMS value is directly related to the energy of the sEMG signal, and is often used in practical applications to reflect the energy generated by the electromyography22. ② Data normalization, the data is listed as [0,1]. The data were normalized using formula 2, ensuring consistency in comparing RMS values across participants and time points. The specific formula is as follows Data normalization, the data is listed as [0,1]. The data were normalized using formula 2, ensuring consistency in comparing RMS values across participants and time points. The specific formula is as follows The normalized EMG RMS value, scaled to a range between 0 and 1. x the original EMG RMS value recorded from the muscle activity. The minimum EMG RMS value observed in the muscle activity data. The maximum EMG RMS value observed in the muscle activity data. ③ The change rate of EMG RMS values across different stages (Initial: 1–5 s; Middle: 15–19 s; Final: 26–30 s) is calculated based on the mean difference between adjacent time periods or the overall time period, divided by the initial mean, and then multiplied by 100% to express the result as a percentage. This method allows for a standardized assessment of muscle activation changes and fatigue progression over the 30-second task. The time intervals (1–5 s, 15–19 s, 26–30 s) were determined by evenly dividing the 30-second high-intensity side-kick task into three phases, representing the early, middle, and late stages of fatigue progression. This division aims to balance the evaluation of muscle activation and fatigue development during the task, with the early phase capturing initial muscle effort, the middle phase reflecting the onset of fatigue, and the late phase indicating advanced fatigue effects. The approach is based on the need to systematically assess the progressive changes in muscle performance over the task duration. the specific formula is as follows: M represents the mean. M2 represents the mean of the subsequent time period, while M1 represents the mean of the previous time period. The change rate of EMG RMS values across different stages (Initial: 1–5 s; Middle: 15–19 s; Final: 26–30 s) is calculated based on the mean difference between adjacent time periods or the overall time period, divided by the initial mean, and then multiplied by 100% to express the result as a percentage. This method allows for a standardized assessment of muscle activation changes and fatigue progression over the 30-second task. The time intervals (1–5 s, 15–19 s, 26–30 s) were determined by evenly dividing the 30-second high-intensity side-kick task into three phases, representing the early, middle, and late stages of fatigue progression. This division aims to balance the evaluation of muscle activation and fatigue development during the task, with the early phase capturing initial muscle effort, the middle phase reflecting the onset of fatigue, and the late phase indicating advanced fatigue effects. The approach is based on the need to systematically assess the progressive changes in muscle performance over the task duration. the specific formula is as follows: M represents the mean. M2 represents the mean of the subsequent time period, while M1 represents the mean of the previous time period. ④ In the field of training physiology, the muscle co-contraction ratio serves as a critical indicator of muscle coordination, with values ranging from 0 to 1. A ratio close to 0.5 suggests comparable activation levels between agonist and antagonist muscles, indicating a high degree of synergistic contraction that enhances movement stability. A ratio approaching 0 implies minimal antagonist muscle activation, with the movement primarily driven by the agonist muscle. Conversely, a ratio nearing 1 indicates excessive antagonist muscle activation, which may reduce movement efficiency. Depending on the movement pattern, the functional roles and divisions of labor between different legs vary accordingly. In the field of training physiology, the muscle co-contraction ratio serves as a critical indicator of muscle coordination, with values ranging from 0 to 1. A ratio close to 0.5 suggests comparable activation levels between agonist and antagonist muscles, indicating a high degree of synergistic contraction that enhances movement stability. A ratio approaching 0 implies minimal antagonist muscle activation, with the movement primarily driven by the agonist muscle. Conversely, a ratio nearing 1 indicates excessive antagonist muscle activation, which may reduce movement efficiency. Depending on the movement pattern, the functional roles and divisions of labor between different legs vary accordingly. Thigh synergistic contraction ratio: Striking leg: Vastus medialis (agonist), hamstrings (antagonist). Striking leg: Vastus medialis (agonist), hamstrings (antagonist). Support leg: Vastus medialis (antagonist), hamstrings (agonist). Support leg: Vastus medialis (antagonist), hamstrings (agonist). Calf co-contraction contraction ratio: Striking leg: Tibialis anterior (antagonist), Gastrocnemius (agonist). Striking leg: Tibialis anterior (antagonist), Gastrocnemius (agonist). Supporting leg, Gastrocnemius (antagonist), Tibialis anterior (agonist). Supporting leg, Gastrocnemius (antagonist), Tibialis anterior (agonist). Statistical analysis A time-series analysis was employed to evaluate changes in muscle activation levels (RMS) and co-contraction ratios over time. Paired-sample t-tests were used to compare differences in muscle activation levels between the striking leg and the supporting leg. Repeated-measures analysis of variance (ANOVA) was conducted to assess differences. Pearson correlation coefficients were calculated to analyze inter-muscle correlations, with a significance level set at p = 0.05. Results RMS differences of the main muscle groups of both lower limbs in Sanda side-kick Data are presented as mean ± standard deviation. As depicted in Fig. 4, the root mean square (RMS) values for the striking leg’s gastrocnemius (0.653 ± 0.182 vs. 0.451 ± 0.149, t = 4.25, p < 0.01), vastus medialis (0.597 ± 0.174 vs. 0.502 ± 0.166, t = 4.45, p < 0.01), rectus femoris (0.618 ± 0.183 vs. 0.508 ± 0.171, t = 4.23, p < 0.01), vastus lateralis (0.612 ± 0.179 vs. 0.519 ± 0.161, t = 4.47, p < 0.01), and gluteus maximus (0.583 ± 0.201 vs. 0.502 ± 0.166, t = 4.03, p < 0.01) were significantly higher than those of the supporting leg. These findings highlight the critical role of these muscles as primary force generators during striking movements. Conversely, the tibialis anterior in the supporting leg showed greater activation compared to the striking leg (0.553 ± 0.162 vs. 0.498 ± 0.168, t=-2.36, p = 0.032), emphasizing its essential contribution to maintaining stability. No notable difference was observed in hamstring activation between the two legs (p = 0.947). Comparison of muscle activation between striking leg and supporting leg. Correlation between the coordinated contraction of the main muscles of the lower limbs in Sanda side-kick The striking leg includes seven muscles: gastrocnemius, tibialis anterior, hamstrings, vastus medialis, rectus femoris, vastus lateralis, and gluteus maximus, whereas the supporting leg comprises six muscles: gastrocnemius, tibialis anterior, hamstrings, vastus medialis, gluteus maximus, and gluteus medius. In the striking leg, notable correlations were observed among the vastus medialis, rectus femoris, vastus lateralis, and gluteus maximus, with correlation coefficients ranging from 0.60 to 0.78. Specifically, the correlation coefficient between vastus medialis and rectus femoris is (r = 0.76, p = 0.010), between vastus medialis and vastus lateralis is (r = 0.74, p = 0.014), and between rectus femoris and vastus lateralis is (r = 0.78, p = 0.007). These results highlight the strong synergistic coordination among these muscles, which are essential for force production at the hip and knee joints during the striking movement.(Table 3) (Fig. 5). Table 3Correlation matrix of the main muscle groups of the striking leg.MuscleGastrocnemiusTibialis anterior muscleHamstringsVastus medialisRectus femorisVastus lateralisGluteus maximusGastrocnemius10.420.350.480.460.470.51Tibialis Anterior Muscle0.421-0.250.310.290.30.28Hamstrings0.35-0.251-0.41-0.39-0.40.33Vastus Medialis0.480.31-0.4110.760.74*0.62*Rectus Femoris0.460.29-0.390.76*10.78*0.60Vastus Lateralis0.470.3-0.40.74*0.78*10.61Gluteus Maximus0.510.280.330.62*0.600.611In the supporting leg, a strong correlation was observed between the gastrocnemius and tibialis anterior (r = 0.81, p = 0.003), highlighting the critical role of ankle joint stability during support. A moderate correlation between the gluteus maximus and gluteus medius (r = 0.65, p = 0.038) indicates synergistic action of hip muscles in maintaining balance. Correlations among other muscles were relatively low (|r| ≤ 0.5), suggesting a more distinct functional differentiation among the supporting leg’s muscle groups (Table 4). Correlation matrix of the main muscle groups of the striking leg. In the supporting leg, a strong correlation was observed between the gastrocnemius and tibialis anterior (r = 0.81, p = 0.003), highlighting the critical role of ankle joint stability during support. A moderate correlation between the gluteus maximus and gluteus medius (r = 0.65, p = 0.038) indicates synergistic action of hip muscles in maintaining balance. Correlations among other muscles were relatively low (|r| ≤ 0.5), suggesting a more distinct functional differentiation among the supporting leg’s muscle groups (Table 4). Rend of muscle changes of the supporting leg and striking leg over time. Table 4Correlation matrix of main muscle groups of the supporting leg.MuscleGastrocnemiusTibialis anterior muscleHamstringsVastus medialisGluteus maximusGluteus mediusGastrocnemius10.81*0.380.340.410.45Tibialis anterior muscle0.81*10.290.320.390.43Hamstrings0.380.291-0.270.350.37Vastus medialis0.340.32-0.2710.360.40Gluteus maximus0.410.390.350.3610.65*Gluteus medius0.450.430.370.400.65*1 Correlation matrix of main muscle groups of the supporting leg. Tibialis anterior muscle Vastus medialis Gluteus maximus Gluteus medius Characteristics of fatigue effects of main lower limb muscle groups in Sanda side-kick Table 5, Fig. 6 illustrate that the RMS values of the striking leg’s muscles exhibited notable declines, indicative of significant fatigue. Specifically, the gastrocnemius decreased from 0.669 ± 0.318 to 0.620 ± 0.201 (-7.33%), vastus medialis from 0.612 ± 0.348 to 0.571 ± 0.281 (-6.7%), gluteus maximus from 0.600 ± 0.343 to 0.555 ± 0.286 (-7.5%), and tibialis anterior from 0.510 ± 0.305 to 0.483 ± 0.223 (-5.29%). Fatigue trends were evident by the middle phase. In contrast, the supporting leg displayed smaller RMS changes, with the gastrocnemius decreasing from 0.460 ± 0.234 to 0.443 ± 0.241 (-3.7%), tibialis anterior from 0.566 ± 0.252 to 0.544 ± 0.234 (-3.89%), and similar reductions in vastus medialis and gluteus maximus (-3.52% and − 3.97%, respectively), suggesting greater fatigue resistance. Regarding co-contraction ratios, the striking leg’s thigh ratio decreased from 0.561 ± 0.225 to 0.538 ± 0.174 (-4.1%), while the calf ratio slightly increased from 0.494 ± 0.231 to 0.505 ± 0.176 (+ 2.23%). In the supporting leg, the thigh ratio increased from 0.495 ± 0.265 to 0.510 ± 0.228 (+ 3.03%), whereas the calf ratio decreased from 0.415 ± 0.220 to 0.390 ± 0.202 (-6.02%), with trends approaching statistical significance. Table 5Trends in RMS values of the striking and supporting leg muscles.ProjectIndexLegsInitial phaseMiddle phaseFinal phaseRate of change 1Rate of change 2Overall rate of change(1 ~ 5s)(15 ~ 19s)(26 ~ 30s)(%)(%)(%)Muscle activationGastrocnemiusStriking Leg0.669 ± 0.3180.645 ± 0.2150.620 ± 0.201-3.59-3.88←-7.33←Supporting leg0.460 ± 0.2340.452 ± 0.2860.443 ± 0.241-1.74-1.99-3.7Tibialis anterior muscleStriking Leg0.510 ± 0.3050.497 ± 0.2390.483 ± 0.223-2.55-2.82-5.29Supporting leg0.566 ± 0.2520.556 ± 0.2140.544 ± 0.234-1.77-2.16-3.89Vastus medialisStriking Leg0.612 ± 0.2480.592 ± 0.2630.571 ± 0.281-3.27-3.55-6.7←Supporting leg0.512 ± 0.2800.504 ± 0.2300.494 ± 0.230-1.56-1.98-3.52gluteus maximusStriking Leg0.600 ± 0.3430.578 ± 0.3010.555 ± 0.286-3.67-3.98-7.50←Supporting leg0.479 ± 0.2760.470 ± 0.2380.460 ± 0.249-1.88-2.13-3.97Co-contractionThighStriking Leg0.561 ± 0.2250.550 ± 0.1610.538 ± 0.174-1.96-2.18-4.1Supporting leg0.495 ± 0.2650.502 ± 0.2230.510 ± 0.2281.411.593.03CalfStriking Leg0.494 ± 0.2310.499 ± 0.1700.505 ± 0.176+ 1.01+ 1.20+ 2.23Supporting leg0.415 ± 0.2200.402 ± 0.1900.390 ± 0.202-3.13-2.99-6.02 Trends in RMS values of the striking and supporting leg muscles. High and calf co-contraction trend. Figure 5 illustrates that during 30 s of continuous side-kick movements, the RMS values of the gastrocnemius in the striking leg exhibited a significant decline. A notable time effect was observed for the striking leg’s gastrocnemius RMS (F(2, 18) = 4.32, p = 0.029, partial η2 = 0.324), with significant differences between the initial vs. middle phases (p < 0.05) and initial vs. final phases (p < 0.05). The gluteus maximus in the striking leg also showed a significant time effect (F(2, 18) = 3.619, p = 0.048, partial η2 = 0.287), with a significant difference between the initial and final phases (p = 0.027). For the vastus medialis, a significant difference was found between the initial and final phases (F(2, 18) = 3.84, p < 0.05, partial η2 = 0.252). The declining trend in RMS for the striking leg approached significance (p < 0.05), whereas the supporting leg showed no significant changes. No significant differences were observed in the tibialis anterior RMS for either leg, with the supporting leg remaining stable. The gluteus maximus RMS in the striking leg significantly decreased (p = 0.05), while the supporting leg showed no notable changes, highlighting the fatigue effect in the striking leg. Figure 6 illustrates the temporal changes in the co-contraction ratio (% RMS) of thigh and calf muscles for both the striking and supporting legs during a 30-second continuous side-kick training. For the striking leg, the co-contraction ratio decreased from 0.561 ± 0.225 in the initial phase to 0.538 ± 0.174 in the final phase, reflecting an overall reduction of 4.1%. This decline suggests diminished activation of the hamstring muscle group (antagonist), indicative of a slight fatigue-induced reduction in knee joint coordination. However, within-group ANOVA analysis (F = 0.957, p > 0.05) revealed no statistically significant differences. Conversely, the supporting leg exhibited an increase in co-contraction ratio from 0.495 ± 0.265 to 0.510 ± 0.228, corresponding to a 3.03% rise. This increase likely reflects enhanced activation of the quadriceps (antagonist) to support dynamic balance. Similarly, ANOVA analysis (F = 0.403, p > 0.05) indicated no significant time-dependent effect, underscoring the absence of a pronounced temporal influence on muscle co-contraction patterns. The co-contraction ratio (% RMS) of the lower leg muscles (tibialis anterior vs. gastrocnemius) during a 30-second continuous side-kick training revealed distinct patterns for the striking and supporting legs. For the striking leg, the co-contraction ratio increased marginally from 0.494 ± 0.231 in the initial phase to 0.505 ± 0.176 in the final phase, representing a 2.23% rise. This subtle increase suggests a slight enhancement in the coordinative role of the tibialis anterior, though the change was not statistically significant (F = 0.212, p > 0.05). In contrast, the supporting leg exhibited a decrease in co-contraction ratio from 0.415 ± 0.220 to 0.390 ± 0.202, a reduction of 6.02%. This decline indicates an increased reliance on the gastrocnemius for stability, with a diminished contribution from the tibialis anterior. Within-group ANOVA analysis (F = 2.78, p = 0.089) approached but did not achieve statistical significance (p > 0.05), suggesting a near-significant trend but no definitive time-dependent effect. Discussion This study aims to identify fatigue-prone and key stabilizing muscles by analyzing the unique side-kick technique in Sanda, providing a scientific basis for developing targeted strength, endurance, and balance training strategies to effectively mitigate fatigue-induced strength asymmetries and reduce potential injury risks. Characteristics of muscle fatigue in Sanda side-kick This study aims to elucidate the activation and force characteristics of primary muscle groups during repetitive in-place side-kick movements in fighters, providing practical guidance for training optimization. From a biomechanical perspective, the supporting leg of Sanda fighters remains firmly anchored to the ground during side kicks, primarily to maintain single-leg balance and provide stability and power for the striking leg. Energy is transmitted sequentially from the supporting ankle upward through each joint. The transmission progresses from the distal to proximal segments of the supporting leg, then from the proximal to distal segments of the striking leg, involving sequential braking and acceleration. During the striking leg’s execution of knee lift and hip extension, the supporting leg’s knee remains slightly flexed, with the hamstring muscle group stabilizing this flexion and the quadriceps assisting in extension to adjust posture. Regarding movement characteristics, during the striking phase of the striking leg, the knee joint extends rapidly, driven by the quadriceps (vastus muscles), while the hamstrings regulate flexion to prevent hyperextension. The root mean square (RMS) values of the gastrocnemius, quadriceps, and gluteus maximus in the striking leg were significantly higher than those in the supporting leg, indicating that the striking leg likely experiences fatigue earlier than the supporting leg, which provides power output for the striking leg in Sanda movements. The study revealed that during continuous side-kick execution, both legs exhibited a declining trend in RMS values for the gastrocnemius (-7.33%), quadriceps (-6.70%), gluteus maximus (-7.50%), and tibialis anterior (-5.29%). Time-series analysis within groups showed significant RMS reductions for the gastrocnemius, quadriceps, and gluteus maximus (p < 0.05). The side-kick technique comprises a series of fluid movements, including knee lift, hip rotation (abduction), and leg extension, with the gluteus maximus playing a near-continuous dual role in both extension and abduction. Notably, following leg extension, the increased inertial moment of the limb imposes an additional burden on the gluteus maximus to sustain the abduction torque, significantly elevating its workload. In contrast, the quadriceps is predominantly engaged in the knee lift and leg extension phases, with a more focused load distribution. Consequently, the gluteus maximus is likely to exhibit a markedly higher degree of fatigue. Electromyography (EMG) studies on fatigue suggest that RMS values significantly decrease during task execution17, confirming pronounced fatigue in the gluteus maximus and gastrocnemius of the striking leg. Fatigue was evident by the middle phase (15 ~ 19 s) and intensified in the final phase (26 ~ 30 s), indicating that these muscles are particularly susceptible to fatigue. The gluteus maximus plays a critical role in force transmission and sustaining continuous thigh output23. Furthermore, the striking leg’s RMS values showed a significant decline over time (-5.29% to -7.50%), with fatigue emerging in the middle phase (15 ~ 19 s) and worsening in the final phase. These findings align with Wang (2020), who reported high gastrocnemius activation in Sanda whip-kick techniques20, and Sun et al. (2024), who observed similar high-activation patterns in taekwondo whip-kick “air strikes”6. In contrast, the supporting leg exhibited smaller RMS changes (-3.89% to -3.97%), maintaining relative stability in the middle phase. During task execution, the supporting leg’s gastrocnemius, quadriceps, and hamstrings synergistically stabilize the knee and ankle joints, thereby supporting the body’s center of gravity. Analysis of the root mean square (RMS) differences in key muscle groups between the striking and supporting legs during a 30-second high-intensity leg extension task revealed that the gluteus maximus, quadriceps, and gastrocnemius of the striking leg exhibited sustained high-intensity activation. Compared to the supporting leg, the striking leg is more prone to fatigue. Electromyography (EMG) can effectively assess the fatigue levels of these muscles17,24. Therefore, in training practice, incorporating rapid stretch-shortening cycle training for the gluteus maximus and gastrocnemius to enhance endurance is recommended. This approach can improve the athletic performance of the striking leg in Sanda fighters. Characteristics of coordinated contraction of Sanda Side-kick Muscle co-contraction25, defined as the simultaneous activation of multiple muscles surrounding a joint to produce specific joint torques26, is central to the functionality of both legs. Previous research suggests that co-activation suppresses task-irrelevant muscle activity during skilled motor tasks, thereby enhancing movement efficiency27,28. In the context of side-kicks, co-contraction of the quadriceps and hamstrings in the striking leg optimizes knee joint control, balancing force output with joint protection. Similarly, in the supporting leg, co-activation of the tibialis anterior and gastrocnemius enhances ankle joint stability, particularly under the dynamic demands of the task. For example, during high-impact movements such as takeoff and landing, co-contraction around the knee joint mitigates impact forces and torques upon ground contact29, with the gastrocnemius playing a dominant role in the supporting leg to improve force production efficiency at the ankle joint30. The co-contraction ratio serves as an indicator of muscle co-activation, reflecting the relative dominance of agonist or antagonist muscles during temporal changes in activation. Physically, a ratio close to 0.5 indicates comparable activation levels between agonist and antagonist muscles, signifying a high degree of co-contraction that effectively enhances movement stability. In this study, the co-contraction ratios of the thigh and calf muscles in both the striking and supporting legs approximated 0.5, suggesting a high level of antagonist muscle coordination at the knee and ankle joints. Side-kick techniques were categorized into high, medium, and low kicks based on striking height, with this study focusing on continuous medium-height side kicks. Participants were required to perform the task as rapidly as possible while maintaining technical precision, balancing the demands of movement standardization and speed. First, standardization necessitated precise muscle contractions during high-intensity side-kick execution, emphasizing the role of agonist and antagonist muscles in maintaining joint stability. Second, movement efficiency required Sanda fighter to maximize speed within the constraints of standardization, highlighting the efficiency of agonist and antagonist muscle co-activation. The results showed a decline in the thigh co-contraction ratio of the striking leg (0.561 ± 0.225 to 0.538 ± 0.174, -4.1%), while the calf co-contraction ratio (tibialis anterior vs. gastrocnemius) increased (0.494 ± 0.231 to 0.505 ± 0.176, + 2.23%). These findings indicate that knee joint antagonist muscles predominated, with efficiency improving over time as the ratio approached 0.5, indicative of high-efficiency co-activation. At the ankle joint, the gastrocnemius initially dominated, but the tibialis anterior’s contribution to dorsiflexion adjustments increased over time, resulting in a ratio of 0.505, reflecting enhanced synergistic efficiency between agonist and antagonist muscles. These results align with Wang (2020), who highlighted the coordination demands of the ankle joint in Sanda leg techniques20. The co-contraction ratio of the quadriceps versus hamstrings in the supporting leg increased (0.495 ± 0.265 to 0.510 ± 0.228, + 3.03%, p = 0.053, η² = 0.28), indicating enhanced activation of the quadriceps (antagonist) to assist knee joint extension and maintain dynamic balance and stability. Conversely, the co-contraction ratio of the tibialis anterior versus gastrocnemius decreased (0.415 ± 0.220 to 0.390 ± 0.202, -6.02%, p = 0.089, F = 2.78), approaching statistical significance, suggesting that the gastrocnemius progressively dominated ankle joint stability, with a reduced contribution from the tibialis anterior. These findings reflect improved efficiency in the interaction between agonist and antagonist muscles at the knee joint, with the gastrocnemius increasingly assuming a dominant role at the ankle joint. This adaptation underscores the supporting leg’s enhanced efficiency in force production, particularly through the gastrocnemius, to meet the positional demands of the task30. Correlation analysis of the side-kick movement elucidates distinct functional roles and coordination patterns between the striking and supporting legs. During the knee-lift and hip-extension phases of the striking leg’s side-kick, rapid knee extension is primarily driven by the quadriceps, particularly the vastus medialis, with the hamstrings modulating knee flexion to prevent hyperextension. This results in a robust correlation in quadriceps activation (r = 0.62–0.76, p < 0.05), forming an effective hip-knee kinetic chain31, consistent with findings on kinetic chain coordination by De Swart et al. (2022)23. The synergistic activity among the vastus medialis, rectus femoris, vastus lateralis, and gluteus maximus underscores their pivotal role in generating force at the hip and knee joints, enabling powerful and coordinated striking movements. In contrast, the supporting leg exhibits a strong correlation between the tibialis anterior and gastrocnemius (r = 0.81, p = 0.003), reflecting co-activation of agonist and antagonist muscles to ensure ankle joint stability. Additionally, a moderate correlation between the gluteus maximus and gluteus medius (r = 0.65, p = 0.038) highlights their synergistic contribution to hip stability, facilitating balance during dynamic movements. These findings emphasize the functional differentiation between the two legs: the striking leg prioritizes force production through synergistic power-generating muscles, while the supporting leg focuses on stability through coordinated stabilizing muscles. However, co-contraction involves a trade-off. While it enhances joint stability and movement precision, excessive antagonist activity, particularly during skill acquisition, may reduce movement efficiency32. For instance, in boxing techniques such as jabs and hooks, co-contraction of upper arm muscles facilitates coordinated shoulder, elbow, and wrist movements, improving striking accuracy28. Conversely, during the learning of new motor skills, heightened antagonist activity may hinder efficiency as the neuromuscular system adapts to task demands. These observations suggest that optimizing co-contraction patterns is critical for balancing stability and efficiency in Sanda side-kick performance. These findings have significant implications for Sanda training programs. The striking leg’s reliance on the quadriceps and gluteus maximus for force production indicates a need for targeted endurance and explosive strength training, particularly for the gastrocnemius, vastus medialis, and gluteus maximus. Concurrently, the supporting leg’s role in maintaining stability necessitates exercises that strengthen the tibialis anterior and gastrocnemius, alongside balance training to enhance hip and ankle joint coordination. By addressing these distinct muscle coordination patterns and fatigue profiles, tailored training regimens can optimize side-kick performance, improving both power output and movement stability. Training recommendations and limitations Training recommendations include endurance training for the striking leg, such as high-intensity interval training, to delay fatigue in the gastrocnemius, quadriceps, and gluteus maximus33. For the supporting leg, exercises targeting ankle joint stability and knee joint coordination, such as single-leg balance drills, are advised. Rapid stretch-shortening cycle training for the gluteus maximus and quadriceps can be implemented using plyometric training like box jumps34. Integrated drills, such as simulated match-specific leg technique combinations, can optimize the functionality of both legs, enhancing performance in high-intensity competitive scenarios. This study is constrained by several limitations. The small sample size (n = 10) and the analysis of only a 30-second movement duration do not fully capture the dynamics of longer competitive scenarios. Additionally, the no-target design, which reduces muscle impact forces, may underestimate fatigue in the striking leg compared to real-world conditions involving reaction forces. The exclusive participation of male subjects further limits the generalizability of the findings. Moreover, the focus solely on elite athletes restricts the applicability of the results to broader populations. To address these limitations, future research should expand the sample size, incorporate kinematic and dynamic analyses to examine prolonged fatigue effects and movement efficiency over extended durations, and compare elite and non-elite athletes to enhance the relevance of findings for preventive measures. Conclusion In Chinese elite Sanda fighters performing continuous in-place side-kick movements, the striking leg’s power-generating muscle groups—primarily the quadriceps and gluteus maximus—exhibit significant fatigue. Meanwhile, the supporting leg maintains stability through the tibialis anterior and gastrocnemius. These findings highlight potential strength asymmetries that could signal fatigue onset or even injury risk, offering valuable insights for coaching strategies and rehabilitation protocols. Training recommendations include plyometric exercises to enhance the explosive power of the quadriceps and gluteus maximus, resistance training to improve the endurance of the striking leg, and isokinetic training to strengthen the stability of the tibialis anterior and gastrocnemius in the supporting leg, thereby further optimizing the training regimen. Author contributions Qiang Wu and Yang Liu contributed equally to this work, sharing first authorship. They were involved in study design, data collection, and manuscript drafting. Yang Liu performed the statistical analysis and contributed to the interpretation of results. Bin Zhao, the corresponding author, supervised the study, revised the manuscript critically, and ensured the integrity of the work. All authors read and approved the final manuscript. Funding This study was funded by the National-sponsored Social Sciences Funding Program (23BTY051) and the 2025 Approved Research Project of the Sichuan Provincial Key Chinese Culture Research Institute – Chengdu Sport University Chinese Guoshu Academy (GS2025A01). Data availability The raw data supporting the findings of this study are available from the corresponding author. Competing interests The authors declare no competing interests. Footnotes Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Qiang Wu and Yang Liu contributed equally to this work. References Associated Data Data Availability Statement The raw data supporting the findings of this study are available from the corresponding author.