A total of 16 (12 male and 4 female) participants who were regularly performing physical or recreational sporting activities were recruited for the study. Their average ± SD (range) age, height and body mass were 26.9 ± 4.0 (20–32) years, 1.73 ± 0.01 (1.52–1.84) m, and 80.4 ± 17.2 (59.2 – 111.7) kg, respectively. The sample size was based on the difference between MVC torque in eccentric (287.7 ± 47.0 Nm), isometric (246.9 ± 50.0 Nm) and concentric contractions (216.7 ± 45.2 Nm) of the knee extensors in young adults from a previous study (Ruas et al. 2018), which provided the Cohen’s effect size of 0.8. Using G*Power 3.1 (Institute for Experimental Psychology, Dusseldorf, Germany), with a power of 0.8 and a significance level of 0.05, it was estimated that at least 12 participants were necessary to identify potential differences in MVC torque among the three contraction types. Accounting for a potential estimation error in the sample size calculation, and possible withdrawal of some participants from the study, 16 participants were recruited.

Participants provided written informed consent and completed a pre-exercise medical questionnaire. To identify leg dominance, the Waterloo Footedness Questionnaire Revised (van Melick et al. 2017) was completed by the participants, and only the dominant leg was assessed in all tests. All participants were identified as right leg dominant. Ethical approval was obtained from the Edith Cowan University Human Research Ethics Committee (Project no. 00944).

All participants visited the laboratory on one occasion and completed the following assessments in order: (1) MVC torque of the knee extensors of the leg for isometric and isokinetic concentric and eccentric contractions; (2) isometric, concentric and eccentric contractions at 30% of MVC (i.e., submaximal contractions) with superimposed single- and paired-pulse TMS; (3) isometric, concentric and eccentric MVCs with superimposed peripheral electrical nerve stimulation; (4) additional isometric, concentric and eccentric MVCs with superimposed single-pulse TMS. The order of isometric, concentric and eccentric contractions was randomized for each block of the measures shown above and among participants.

Neurophysiological responses were recorded via surface EMG from the vastus lateralis muscle. Outcome measures included MVC torque of eccentric, isometric and concentric contractions, and other neuromuscular indices [i.e., MEP, SICI, CSP duration, resting twitch torque, superimposed twitch torque, VA, EMG activity, Maximal M wave (MMAX)] during maximal and/or submaximal contractions which were compared across contraction modes.

All participants sat upright on an isokinetic dynamometer (Biodex System 4 Pro, Shirley, NY) for all assessments with the hips positioned at 85° of hip flexion (0° = full hip extension), and had straps across their chest, hips and thighs to minimize movement of other parts of the body. The lateral epicondyle of the femur from the participant’s dominant leg was aligned to the dynamometer’s axis of rotation, with the leg attached to the lever arm 2 cm above the medial malleolus (Doguet et al. 2017).

During all testing, EMG activity from the vastus lateralis muscle was recorded by a PowerLab EMG system with a 16-bit analog-to-digital converter (PowerLab 16/35, ADInstruments, 457 Bella Vista, NSW, Australia) using surface electrodes (Ag–AgCl; Ambu Blue Sensor N-00-S/25, Ambu, Denmark). A LabChart software (ADInstruments, Bella Vista, NSW, Australia) was used to record EMG and torque signals at a sampling rate of 2-kHz (common mode rejection ratio > 85 dB, gain = 1000), which also automatically triggered the transcranial and peripheral nerve stimuli. For all participants, one electrode was placed at ~ 66% of the line between the inguinal crease and the patella and the second, 5 cm distal, in a pseudo-monopolar orientation (Ruas et al. 2022a, b). A ground electrode was placed over the tibial tuberosity of the tested leg, after the areas were shaved, abraded and cleaned to reduce impedance (Z < 5 kΩ).

Raw EMG signals were filtered (20–1000 Hz band pass filter) and amplified (1000x). Root mean square (RMS) EMG was calculated over a 500 ms period prior to stimulation during eccentric, isometric and concentric contractions. RMS was also calculated around the time of peak torque during MVCs without the stimulation to ensure the torque, EMG and knee angle recordings were matched. For the submaximal contractions, target torques of 30% MVC were set for each participant, because the 30% level allowed the participants to maintain the target torque for the entire range of motion (between 110° and 40° of knee flexion) tested. The EMG values prior to simulation were divided by the torque at 30% MVC for each contraction and averaged. For MVCs, the EMG values of the MVCs (without stimulation) that presented the highest peak torque of the three trials of each contraction type tested were selected and divided by the peak torque of that MVC (EMG/torque) (Duclay et al. 2011).

Participants performed eccentric, isometric and concentric MVCs unilaterally with the dominant leg (the preferred leg for kicking a ball) in a randomized order on the isokinetic dynamometer. Isometric MVC consisted of three repetitions of 3 s with a 2-min rest between repetitions with the leg positioned at 75° of knee flexion (0° = full knee extension) (Doguet et al. 2017; Brown and Weir 2001). The angle (75°) has been reported to represent an intermediate muscle length of vastus lateralis and is within the optimal range for isometric knee extension torque production (Doguet et al. 2017; Becker and Awiszus 2001; Pietta-Dias et al. 2020).

Eccentric and concentric MVCs were performed at 30°/s through 70° of range of motion (between 110° and 40° of knee flexion) on the isokinetic dynamometer (Doguet et al. 2017). Both eccentric and concentric MVC measures consisted of three trials each with a 3-min rest between trials and conditions. Two practice contractions with a 3-min rest between each were given before the start of each set of muscle contractions to ensure that the maximal effort contractions were performed by each participant. In the eccentric trials, participants were asked to perform an isometric MVC at 40° of knee flexion first (Jensen et al. 1991) for ~ 1 s, and then resist the movement of the machine as hard as possible (Brown and Weir 2001). For the concentric trials, participants were asked to push the lever arm as hard and as fast as possible starting at 110° of knee flexion. The MVC that presented the highest peak torque of the three trials for each contraction mode was used for further analysis. To assess the magnitude of neuromuscular fatigue, the isometric MVC torque was also measured at the end of the exercise session following the same protocol described above. Based upon the highest peak torque, target torques of 30% of MVC were calculated for later assessments, and the ratio of eccentric MVC torque relative to isometric MVC torque (ECC/ISOM) was calculated for later analysis.

Peripheral nerve stimuli were delivered over the femoral nerve to evoke M-waves, resting and superimposed twitches. Electrical stimuli were delivered by a constant-current stimulator (DS7AH, Digitimer, Welwyn 369 Garden City, UK) using a cathode and anode (White Sensor 4560 M, 79 mm, Ambu, Ballerup, Denmark) placed over the femoral triangle and greater trochanter, respectively (Ruas et al. 2020). Single, electrical stimuli with a duration of 200 μs and increasing intensity were delivered until a maximal amplitude of the M-wave (MMAX) was reached for the vastus lateralis muscle with the knee at rest and passively supported at 75° of flexion (Doguet et al. 2017; Duclay et al. 2011). The superimposed twitch and resting twitch torques were assessed using a supramaximal stimulus intensity (equivalent to 150% of MMAX intensity). This stimulus was automatically delivered at 75° of knee flexion during isometric, and eccentric and concentric MVCs at 30°/s to elicit superimposed twitch, and also at ~ 2 s after each MVC with the muscles relaxed to elicit a potentiated resting twitch. After isometric MVCs, the leg remained at 75° of flexion, whereas after eccentric and concentric MVCs, the dynamometer returned the leg to its starting position and then moved it passively through the 70° range (lengthening or shortening the knee extensors also at 30°/s with stimuli automatically delivered at 75°). Three trials consisting of superimposed twitch and resting twitch stimuli were given for each contraction type. Resting twitch and superimposed twitch torque amplitudes were measured as the difference between the torque just prior to the onset of the twitch (i.e., approximately 12 ms after the femoral nerve stimulation) and the peak torque of the twitch. VA was further determined by calculating (1 – superimposed twitch torque/resting twitch torque) × 100. For each contraction type, three values of each parameter [i.e., resting twitch torque (Nm), superimposed twitch torque (Nm), peak-to-peak amplitude (mV) of MMAX, and VA (%)] were averaged and used for further comparisons.

TMS (Magstim 2002, Magstim Co, Dyfed, UK) was delivered with a 110 mm double cone coil. First, the ‘hotspot’ of the vastus lateralis was found using a stimulation intensity that elicited a small response in the muscle with the knee relaxed and passively supported at 75° of flexion (Doguet et al. 2017). The hotspot was determined as the area that evoked the greatest MEP amplitude with a given stimulation intensity. To determine the hotspot, the coil was placed at the vertex and then moved in medio-lateral and/or anterior–posterior directions by 1-cm steps until finding the greatest stimuli response (Ruas et al. 2020). Then, the active motor threshold was determined as the intensity at which at least 5 of 10 stimuli evoked a MEP (peak-to-peak amplitude > 200 µV) (Rothwell et al. 1999) while individuals performed a knee extensor isometric contraction equal to 10% of isometric MVC at 75° of knee flexion. The TMS intensity was then increased to 140% of active motor threshold (considered as an intensity within the rising phase of stimulus response curve) and kept constant throughout the protocol regardless of contraction type (Doguet et al. 2017; Rossini et al. 2015).

During testing, TMS was delivered during submaximal contractions. Five single- and five paired-pulse TMS were delivered during 10 separate 30% MVCs for each type of muscle contraction (i.e., eccentric, isometric and concentric). Paired-pulse TMS used a subthreshold of 74% of the active motor threshold conditioning pulse and 140% of the active motor threshold test pulse (interstimulus interval of 2 ms) to assess SICI. The conditioning pulse intensity was based on our previous study (Ruas et al. 2020) that found an average conditioning intensity of 74% of the active motor threshold eliciting ~ 50% of maximal inhibition for individuals when measuring SICI related to the vastus lateralis. Contraction torque targets for each individual were set as 30% of the isometric, concentric and eccentric MVCs recorded at the start of the experimental session. For the eccentric and concentric isokinetic contractions, TMS and peripheral nerve stimuli were externally triggered so that they were delivered automatically for each repetition at 75° of knee flexion (Doguet et al. 2017). Single pulse MEP peak-to-peak amplitudes were averaged. Paired pulse MEP amplitudes were averaged and expressed as a percentage of the single pulse MEPs for each contraction mode.

Single pulse TMS was also delivered during isometric MVCs and eccentric and concentric MVCs at 30°/s (five repetitions for each MVC in a block randomized order). Based on the study by Hahn et al. (2012), at least 3 min of rest between MVCs and between contraction modes was provided in order to minimize fatigue. MEP peak-to-peak amplitude during MVCs was averaged for each contraction type.

CSP duration was also calculated from single-pulse TMS delivered during submaximal and maximal contractions as the time interval between the stimulus and the return of EMG activity (i.e., 50% of its background value over 100 ms period prior to stimulation) (Butler et al. 2012). The average CSP duration for each contraction mode tested was calculated.

Data were first screened using a Shapiro–Wilk test, which confirmed that all data were normally distributed. The absolute values of MVC torque, peak torque angle of MVC, MEP, SICI, CSP duration, EMG (RMS), MMAX, superimposed twitch torque, resting twitch torque and VA of participants were compared between eccentric, isometric and concentric contractions by one-way repeated measures ANOVA for each variable. Based on the eccentric MVC torque of the knee extensors in relation to the isometric MVC torque, the same dependent variables were compared between two groups of individuals according to their ECC/ISOM MVC torques.

In order to standardize the groups, the ECC/ISOM MVC of the participants was transformed and standardized to Z-scores, resulting in a common group standard mean value = 0, and a standard deviation value = 1. Individuals presenting a z-score smaller than the mean of 0 (i.e., – 0.411 to – 1.689) were considered as individuals with low ECC/ISOM MVC torque (Group A; n = 7; range 90.0 to 109.7%), and those that presented a z-score greater than the mean of 0 (i.e., 0.001 to 2.88) were considered as individuals with high ECC/ISOM MVC torque (Group B; n = 9; range 111.1 to 186.1%). The variables were compared between the two groups by two-way ANOVAs with contraction type as a repeated measures factor. If significant F values were found, results were followed up with least significant difference post-hoc analysis. A Greenhouse–Geisser correction was used if sphericity was violated. Isometric MVC torques at the beginning and end of the session were also compared by a paired t-test to determine if fatigue had occurred. Significance level was set at p < 0.05. All analyses were performed with SPSS 21.0 (Statistical Package for Social Sciences, Chicago, IL, USA). Percentage differences among the three contraction modes are also reported in the results.