For the study, 20 healthy young adults (11 women, 9 men) were recruited. Due to the novelty of directly comparing EQI contractions between different intensities and including a different muscle group, determining an appropriate sample size through conventional power analysis was challenging. The sample size was similar to previous studies with a similar design which focused on knee muscles (Oranchuk et al., 2021; Ličen et al., 2024). The participants were students recruited through social media posts and word of mouth. The inclusion criteria for the study were the absence of musculoskeletal injuries in the last 3 months, any ankle injuries in the past year, and at least a moderate experience with resistance training (minimal of average 2 sessions per week for the past 6 months). Exclusion criteria included pregnancy, the presence of chronic non-communicable diseases, and prior experience with EQI-based training. Participants were required to sign an informed consent form to participate in the research. The study was approved by the Commission of the Republic of Slovenia for Medical Ethics (No. 0120-690/2017/8).

The measurements took place between 19 February 2024 and 23 March 2024. Each subject completed the protocol in 1 day. This study employed a cross-sectional, within-participant design in which each participant’s two legs were utilized to examine the effects of moderate (75% MVIC) and high (90% MVIC) load EQI muscle actions The assignment of intensity to the left and right legs, as well as the sequence of measurements, was quasi-randomized using a Latin square design. The protocol involved participants first undertaking plantar flexion MVIC assessments, followed by three EQI muscle contraction with 2-min intervals for rest. Immediately after the final EQI set, another MVIC measurement was conducted to evaluate the fatiguing effects of the EQI protocol. Following a 5-min rest period, the entire procedure was replicated with the alternate leg. The full measurement protocol was developed jointly by the authors, while the measurements were carried out entirely jointly by LK and ŽK. All assessments were conducted in a consistent environment, specifically, in an air-conditioned room with the temperature maintained between 22°C and 23°C.

MVIC assessment was performed on an isokinetic dynamometer (HumacNorm, Computer Sports Medicine Inc., Massachusetts, United States). The dynamometer backrest inclination was set to 80° in all participants (Figure 1A). Although there was no additional lumbar inclination, we ensured that the lumbar region of all participants was supported by pads and was slightly curved to maintain the natural arch of the lower back. Other dynamometer position settings were tailored to each individual to ensure correct participant positioning. The participant’s leg was in line with the dynamometer’s foot support (Figure 1B), with the knee padded and slightly flexed. The axis of rotation of the ankle was at the same level as the axis of the dynamometer foot support. The participants were additionally fixated with straps across the shoulders, pelvis, and distal part of the active limb femur. These fixations were refined through pilot testing and were done to minimize the activation of gluteal and quadriceps muscles.

Upon fixating the participant, RoM was determined by asking the participant to place the ankle joint in maximal plantar and dorsal flexion, respectively (Figure 2). Before starting the MVIC measurements, a correction for the effect of gravity was performed in the neutral position. After three warm-up trials for familiarization (first repetition with 50%, second with 75%, and third with 90% of maximum effort), participants performed three 5-s repetitions of MVIC in the neutral ankle position (i.e., 0°), interspersed with 30-s rest intervals. Participants had real-time visual feedback available (time-torque curve) and were verbally encouraged throughout each repetition. After the EQI protocol for each leg (see next section), the procedure was immediately repeated to assess the fatiguing effects.

The dynamometer was set to the isotonic mode, which allows the resistance to be consistently applied in the same direction (i.e., dorsal flexion in the case of the present study). The average torque was 131.9 ± 29.1 Nm for the 75% MVIC load and 161.3 ± 31.7 Nm for the 90% MVIC load, respectively. Each EQI set consisted of one repetition which started in maximal dorsal flexion. First, participants performed a concentric burst against the direction of resistance (i.e., performing a plantarflexion movement), followed by holding this position and maximally resisting the eccentric contraction by maintaining maximal voluntary plantarflexion torque. Once the dynamometer overcame the participant, the joint moved towards a more dorsiflexed position; consequently, plantar flexors were stretched, which resulted in greater plantarflexion torque capacity. Thus, the participant could maintain the new position for additional time despite prior fatigue. This process, characterized by alternating slow eccentric contractions and isometric holds, continued until maximal dorsal flexion RoM was achieved. The available RoM was the same as determined in MVIC contraction measurements for each participant. While the end RoM was the same as the maximal dorsal flexion RoM, the starting RoM was typically lower than the maximal plantar flexion RoM, as the participants could not push the dynamometer to full RoM against the resistance. Note that in previous studies, EQI contraction typically begun by moving the joint to a position near maximal shortening of the target muscle with examiner’s help. In our study, we operationalized this as the position achieved through maximal concentric effort under load, rather than anatomical end-range, to ensure tolerability and safety. Participants were verbally encouraged throughout the duration of the EQI muscle contraction. The rest period between three same-intensity repetitions was 2 min, and the rest period between testing the other intensity was 5 min. After that, the whole protocol, including a warm-up on the dynamometer, an initial MVIC measurement, three EQI muscle contraction and a final MVIC measurement, was repeated at a different intensity on the other leg.

The data was exported from the dynamometer software (version 9.8.0; CSMi–Computer Sports Medicine Inc., Chicago, United States) to a Microsoft Excel 2013 (Microsoft Corporation, Redmond, Washington, United States) file and then analyzed using MATLAB software (version R2020a; The MathWorks Inc., Natick, Massachusetts, United States). Position (angle plantarflexion/dorsiflexion angle), angular velocity, and torque signals were captured at 100 Hz and processed without prior filtering (Oranchuk et al., 2021). For the MVIC, we considered the peak torque of the best repetition as the only dependent variable.

The start of the EQI muscle contraction was identified at the first local maximum for angle data (i.e., maximal plantar flexion). The end of the EQI contraction was determined when either 1) the end of the RoM was reached or 2) torque values dropped below 50% of the prescribed threshold. All signals were visually inspected for artefacts that could affect the determination of EQI muscle contraction onset and offset, with manual corrections made where needed (n = 3 contractions). To mitigate the risk of type 1 error rates and associated false discovery rates, our analysis primarily concentrated on four outcome variables: 1) the total duration of the EQI muscle contraction (s), 2) the total torque impulse (area under torque-time curve determined with the trapezoid integration) (Nm⋅s); 3) the mean angular velocity (°/s) and 4) total RoM (°). Although the final angle (end-RoM) was expected to be fairly consistent among repetitions, the fatigue of prior repetitions could cause the participant to reach a smaller initial plantar flexion with the concentric contraction. Consequently, including RoM as a dependent variable in the study was considered crucial to obtain a holistic picture of the differences among intensities and subsequent repetitions.

The data are presented as means ± standard deviations. The normality of the data distributions for all variables was verified with the Shapiro-Wilk test and visual inspection of histograms and Q-Q plots. Preliminary analysis indicated no differences between left and right legs (F = 0.262–0.455; p = 0.411–0.615), nor any leg × sex interactions (F = 0.001–0.321; p = 0.771–0.985). Therefore, the side factor (left, right) was not further considered as a factor in the main analyses. MVIC torque was analyzed with a general linear model which included time (before and after EQI muscle action) and load (75% and 90% MVIC) as within-subject factors and sex (men, women) as a between-subject factor. For EQI muscle action, a general linear model with load (75% and 90% MVIC) and repetition (1st, 2nd, 3rd) as within-subjects factors and sex as between-subject factors was applied. Mauchly’s test was used to assess sphericity and the Greenhouse-Geisser correction was applied to adjust for any violations of sphericity. Bonferroni-corrected post hoc testing was used to assess pairwise differences among repetitions. Effect sizes were calculated as partial eta-squared (η_p ²) and were considered to indicate no effect (<0.01), a small effect (0.01–0.06), a medium effect (0.06–0.14), and a large effect (>0.14). In case of significant 2-way or 3-way interactions, estimated marginal means with 95% confidence intervals (CI) and post hoc 1-way analyses of variance were conducted. The threshold for statistical significance was set at α < 0.05 and all analyses were carried out in SPSS statistical software (version 25.0, IBM, United States).

The analysis indicated no statistically significant effect of EQI protocol on absolute MVIC torque (main effect of time: p = 0.636) (Table 2). No statistically significant main effect of time was observed between 75% and 90% MVIC loads (p = 0.589), and there was also no load × time interaction (p = 0.314). As expected from the baseline sex comparisons (Table 1), there was a large sex effect (F = 10.3; p = 0.005; η_p ² = 0.37). The remaining interaction effects analyzed in the general linear model (not shown in Table 2) were also not statistically significant (load × sex: p = 0.897; time × sex: p = 0.168; time × sex × load: p = 0.850).

For EQI muscle contraction time, there was no main effect of sex (p = 0.433), load × repetition interaction (p = 0.577), nor load × sex (p = 0.526), repetition × sex (p = 0.588), and repetition × sex × load (p = 0.342) interactions. There was a large and statistically significant effect of load (F = 29.2; p < 0.001; η_p ² = 0.60), as well as repetition (F = 8.5; p = 0.005; η_p ² = 0.31). Specifically, the total time was shorter with every subsequent repetition and during 90% compared to 75% MVIC load (Figure 2A). Post-hoc tests demonstrated that Repetition 3 was shorter compared to Repetition 1 (p = 0.013) and Repetition 2 (p = 0.003), while there was no statistically significant difference between Repetition 1 and 2 (p = 0.208).

Torque impulse also showed statistically significant and large effects of load (F = 17.2; p = 0.001; η_p ² = 0.48) and repetition (F = 7.0; p = 0.011; η_p ² = 0.27), without statistically significant differences between sexes (p = 0.331) and any statistically significant interactions (p = 0.251–0.779). Torque impulse was larger at 75% MVIC compared to 90% MVIC and was decreasing with subsequent repetitions (Figure 2B); similar to total contraction time, post hoc tests showed that Repetition 3 had lower torque impulse compared to Repetition 1 (p = 0.024) and Repetition 2 (p = 0.001), while there was no statistically significant difference between Repetition 1 and 2 (p = 0.335).

For mean angular velocity (Figure 2C), there was no effect of load (p = 0.244), repetition (p = 0.743), and sex (p = 0.942). In addition, there were no statistically significant 2-way or 3-way interactions (p = 0.154–0.893).

For total RoM, there was a large and statistically significant effect of load (F = 25.8; p < 0.001; η_p ² = 0.58) and repetition (F = 9.72; p = 0.001; η_p ² = 0.34). Total RoM was larger in 75% MVIC compared to 90% MVIC and decreased with subsequent repetitions (Figure 2D). Post-hoc tests showed that Repetition 3 was characterized by a smaller RoM compared to Repetition 1 (p = 0.005) and Repetition 2 (p = 0.015), while there was no statistically significant difference between Repetition 1 and 2 (p = 0.174). There was no load × repetition interaction (p = 0.307). There was a statistically significant effect of sex (F = 5.35; p = 0.032; η_p ² = 0.22) and a statistically significant sex × load interaction (F = 4.65; p = 0.044; η_p ² = 0.20). There was no repetition × sex (p = 0.794) and no repetition × sex × load (p = 0.888) interactions. Additional analysis of estimated marginal means (pooled across repetitions) showed that women had larger RoM compared to men at 75% MVIC (23.4° (95% CI = 20.5–26.4°) compared to 15.6° (95% CI = 13.1–18.1°); p = 0.009), but not at 90% MVIC (15.6° (95% CI = 13.1–18.1°) compared to 14.1° (95% CI = 11.2–16.9); p = 0.413).

Further analysis was done to discern whether the starting angle (i.e., the plantar flexion angle to which the participants were able to push concentrically before starting the EQI muscle action) or the final angle was driving the differences in RoM. Starting angle displayed only statistically significant effects of load (F = 13.1; p = 0.002; η_p ² = 0.40) and repetition (F = 14.5; p < 0.001; η_p ² = 0.40), with no sex effects (p = 0.784) nor any interaction (p = 0.101–0.530). Additional analysis of values between loads showed that the starting angle was statistically significantly larger (i.e., more plantaflexed) for the 75% MVIC load compared to 90% MVIC (mean difference: repetition 1: +6.1 ± 7.4°, p = 0.001; repetition 2: 3.7 ± 4.7°, p = 0.002; repetition 3: 3.6 ± 6.2°, p = 0.017). Conversely, the final angle did not statistically significantly differ among intensities (p = 0.137) nor across repetitions (p = 0.441). This was expected, as the participants were required to perform the EQI muscle contraction to the maximal available RoM. However, the final angle showed a statistically significant effect of sex (F = 9.66; p = 0.006; η_p ² = 34), as well as a statistically significant sex × load interaction (F = 5.71; p = 0.027; η_p ² = 0.23). There were no other statistically significant interactions (p = 0.301–0.631). Additional analysis of estimated marginal means (pooled across repetitions) showed that women had larger final angle than men at 75% MVIC (15.8° (95% CI = 13.7–17.9°) compared to 9.8° (95% CI = 7.4–12.2°); p = 0.001), but not at 90% MVIC (12.9° (95% CI = 10.8–15.1°) compared to 10.4° (95% CI = 7.9–12.8°); p = 0.128). Overall, this suggests that different starting angles caused the differences in RoM among loads and repetitions, while a different final angle was the primary cause of sex differences and sex × load interactions.

Based on the previous studies and our findings, we believe that EQI exercise could have potential for enhancing exercise-based programs. While the EQI protocols investigated in this study (75% and 90% of MVIC torque) can be classified as moderate- and high-load conditions, our findings suggest that both protocols may be appropriate for regularly active individuals. However, the 75% MVIC protocol appears more favorable, as it resulted in a greater total torque impulse and a larger range of motion, indicating a potentially more effective mechanical stimulus despite the lower load. The findings of our study could be applied to a practical situation in individuals who do not have access to an isokinetic dynamometer to perform finely controlled EQI muscle actions For instance, an alternative approach using a Smith machine for unilateral EQI muscle actions of the ankle plantar flexors could be used. After determining the unilateral one repetition maximum (1 RM), participants would calculate 75% of the determined 1 RM load and perform a bilateral concentric contraction of the ankle plantar flexors, immediately followed by an EQI contraction through the full RoM. The Smith machine would facilitate a safer and as optimized as possible execution of this protocol outside of a laboratory setting. However, future studies are warranted to investigate the biomechanical characteristics of EQI muscle actions performed on a Smith machine.

This detailed examination of EQI muscle contraction torque, time under tension, velocity, and RoM can serve as a potential starting point to explore the efficacy of EQI exercsie for tendon rehabilitation. In case of the Achilles tendon, injuries are most frequently referred to Achilles tendinopathy (Sharma and Maffulli, 2006). The latter is characterized by swelling, pain, and impaired performance of the Achilles tendon (Alfredson, 2003; Maffulli et al., 1998; van Dijk et al., 2011). The field of conservative management of Achilles tendinopathy is relatively well-researched. While some protocols now incorporate well-defined, objective loading criteria (e.g., Radovanović et al., 2022), the main criticism of most exercise-therapeutic protocols (e.g., Alfredson protocol, HSR protocol, Silbernagel combined protocol, etc.) is the imprecise determination of the load that needs to be added at later stages, where increasing the load is mostly based on the subjective sensation of pain level, which should be around 5 on a 10-point scale, and subjective feeling that the patient will be able to perform a certain number of quality repetitions (Luan et al., 2019). Since EQI exercise is based on the well-defined load (even in case of Smith machine variant) and because 70%–75% MVIC is considered sufficient to induce an Achilles tendon strain of approximately 4.0%–4.5% (McCrum et al., 2018; Wiesinger et al., 2015), both intensities can potentially be considered adequate to promote tendon adaptations such as tendon cross-sectional area, stiffness and Young’s modulus increase (Arampatzis et al., 2007; Bohm et al., 2015; Wiesinger et al., 2015). However, it is important to note that not only the magnitude of strain, but also its duration, and the strain-time integral is potentially critical in triggering tendon adaptation. Before providing a firmer conclusion on the potential use of the EQI exercise for rehabilitation, more detailed studies investigating the morphological, biomechanical and fatigue effects after the implementation of at least 12-week isometric, eccentric or EQI protocol should be executed. In addition, studies to determine the most appropriate intensity and the most appropriate rules for increasing loads within EQI protocol should be carried out prior to a study examining the effects on specific pathologies.

Some limitations of our study must be acknowledged. First, because the study participants were young and regularly physically active, their results are not directly transferable to the general, older, and less physically active population. Furthermore, we found out that neither the load at 75% nor at 90% of MVIC caused statistically significant effects on post-protocol MVIC torque. We believe that these results are since, as mentioned above, the participants were regularly physically active and only one series of three EQI repetitions for each intensity was performed. Therefore, the fatiguing effects of a more extensive EQI protocol remain to be explored. An additional limitation of the study is the non-standardization of knee position between subjects. The knee was slightly flexed due to discomfort when the protocol was performed and this level was not precisely controlled, which may have influenced the results. A further limitation is that EQI contractions began at the end of active concentric plantarflexion rather than full anatomical plantarflexion, to reduce the risk of cramping. While this approach improved safety, it may affect comparability with previous protocols starting from a more shortened muscle position.