Neuromuscular electrical stimulation affects the change in stretch tolerance by static stretching Authors: Akira Saito, Takamasa Mizuno Published: 2026 ABSTRACT AbstractNeuromuscular electrical stimulation (NMES) generates contractions by repetitive depolarization of axons under stimulating electrodes, and neural pathways activating motor units are affected by the stimulation parameters and delivery location. The present study examined the differences in acute effects of NMES applied over the nerve trunk and over the muscle belly on changes in range of motion and passive torque by static stretching. Effects of NMES or prolonged resting and 5-min stretching on the maximal dorsiflexion angle and passive torque were investigated. NMES consisted of 60 stimulation trains (10 s ON, 10 s OFF) at 10 Hz to the tibial nerve or triceps surae. Passive ankle dorsiflexion was performed before and after NMES and stretching interventions. Stretch tolerance and stiffness of the muscle-tendon unit were evaluated by the peak and slope of passive torques, respectively. Increases in the maximal dorsiflexion angle by stretching did not differ significantly among NMES over the nerve trunk (3 ± 2°), the over the muscle belly (2 ± 3°), and prolonged resting (2 ± 2°). Decreases in stiffness of the muscle-tendon unit by stretching were not significantly different among NMES over the nerve trunk (− 13.2 ± 7.1%), over the muscle belly (− 8.8 ± 11.3%), and prolonged resting (− 11.5 ± 11.5%). Significant increases in stretch tolerance by stretching were observed with NMES over the nerve trunk (9.0 ± 10.2%) and over the muscle belly (6.9 ± 14.6%), but not with prolonged resting (1.5 ± 8.7%). NMES tended to facilitate greater stretch tolerance by static stretching regardless of electrode location.Keywords: Hoffmann-reflex, Sensory nerve, Ankle joint, Muscle-tendon unit stiffness, Flexibility Introduction Neuromuscular electrical stimulation (NMES) generates muscle contractions by repetitive depolarization of axons under the stimulating electrodes over nerves or muscles (Bergquist et al. 2011a, b, 2012). It has been demonstrated that the contribution made by central and peripheral pathways to motor unit recruitment during NMES is affected by the NMES parameters (i.e., stimulation intensity, frequency, pulse duration) and location of NMES delivery (i.e., over the nerve trunk or the muscle belly) (Collins 2007; Bergquist et al. 2011a, b; Bickel et al. 2011). In general, regardless of NMES location, generating contractions with a preferential contribution of the central pathway (i.e., sensory volley recruits motor units through reflex pathways) requires NMES delivered at low stimulation intensity, wide pulse duration, and high stimulation frequency (Bergquist et al. 2011b). Contractions generated by NMES over the nerve trunk, but not over the muscle belly, also enhance the central contribution through spinal reflex circuits even when the stimulation is delivered at low frequencies (Bergquist et al. 2011b). Furthermore, the electrically evoked sensory inputs during NMES over the nerve trunk increase cortical activation (Arienzo et al. 2006; Sutherland and Tang 2006), and the sensory inputs evoked by NMES lead to changes in cortical excitability (Ridding et al. 2000; Mang et al. 2010). It has been demonstrated that synchronization of movement-related cortical activity and NMES-evoked sensory inputs induce plastic changes in the motor cortex (Stefan et al. 2000; Wolters et al. 2003; Kaneko et al. 2022). Therefore, it is possible that NMES over the nerve trunk is more effective for inducing neural changes in spinal and cortical circuits than NMES over the muscle belly. Static stretching is commonly used by athletes and in clinical settings to improve the maximal range of motion (ROM) of human joints. Maximal ROM affects the ability of individuals to perform athletic tasks and activities of daily living (Oosterwijk et al. 2018), and it may impact soft tissue injury risk (Witvrouw et al. 2003; Gabbe et al. 2005). It also changes considerably with aging (Bassey et al. 1989) and disease (Schenkman et al. 2001). As one of the static stretching techniques, a brief voluntary contraction during stretching is suggested to induce greater changes in ROM after the stretch. For example, proprioceptive neuromuscular facilitation (PNF) stretching has been demonstrated to be more effective for increasing ROM than conventional static stretching (Funk et al. 2003; Hindle et al. 2012). In the contract-relax method of PNF, stretching is performed following contraction of the target muscles isometrically and immediately after the muscles are more stretched (Magnusson et al. 1996a; Kay et al. 2015, 2016). It has been suggested that some neuromuscular mechanisms are associated with changes in ROM with PNF stretching (e.g., autogenic inhibition and gate control theory) (Funk et al. 2003; Guissard and Duchateau 2006; Sharman et al. 2006; Hindle et al. 2012). Autogenic inhibition may occur during muscle contraction in PNF stretching (Sharman et al. 2006; Hindle et al. 2012). This neural pathway involves increased activity of Ib afferents from Golgi tendon organs inhibiting the spinal excitability of the stretched muscles (Sharman et al. 2006). In the gate control theory, it has been suggested that pressure receptors activated during contractions could inhibit pain perception because pressure receptors are associated with larger myelinated neurons that connect to the same spinal interneurons as unmyelinated nociceptive fibers within the spinal horn (Moayedi and Davis 2013; Capobianco et al. 2018). Thus, the increased activity of pressure receptors could dampen nociceptive signals and thereby decrease pain perception, which allows individuals to stretch further (Kay et al. 2016). However, PNF stretching techniques require a brief and strong voluntary contraction and an assistant. The electrically evoked contraction by NMES over the nerve trunk may substitute for a brief contraction of PNF stretching and improve changes in ROM after static stretching. Although previous studies used NMES over the muscle belly to examine the acute effect of NMES on changes in ROM of static stretching, their results were inconsistent (Piqueras-Rodríguez et al. 2016; Capobianco et al. 2018; Mizuno 2021, 2023). Using patterned NMES over the common peroneal nerve, we previously examined whether the modulation of reciprocal Ia inhibition affects changes in ankle ROM and passive torque after static stretching (Saito and Mizuno 2023). The modulation of reciprocal Ia inhibition did not affect the change in ROM by stretching, and greater changes in stretch tolerance than static stretching without NMES were observed in both patterned NMES and uniform NMES (ineffective for alternating reciprocal Ia inhibition) (Saito and Mizuno 2023). These results suggest that afferent inputs evoked by NMES to the sensory nerve enhance the changes in stretch tolerance after static stretching. However, whether changes in ROM and stretch tolerance improve when static stretching is performed following NMES-induced muscle contraction of target muscles remains unclear. Furthermore, the effects of stimulus location of NMES on changes in ROM by static stretching also remain unclear. Therefore, the present study compared the additional effects of NMES applied over the nerve trunk with those of NMES applied over the muscle belly on changes in ROM by static stretching, to test whether electrically induced sensory volleys are a specific neural factor improving ROM and stretch tolerance acutely. The purpose of the present study was to compare the acute effects of NMES applied over the nerve trunk with those of NMES applied over the muscle belly on changes in ROM and passive torque by static stretching. We hypothesized that combining NMES over the nerve trunk and static stretching induces greater changes in ROM and stretch tolerance than NMES over the muscle belly and static stretching. To test this hypothesis, changes in the maximal dorsiflexion angle and passive torque by static stretching were compared among NMES over the nerve trunk, NMES over the muscle belly, and prolonged resting without NMES. Moreover, H-reflex responses from triceps surae during NMES were recorded to quantify the activity of spinal reflex circuits. Materials and methods Subjects Seventeen men (values are given as mean ± standard deviation [SD]; age, 20.1 ± 4.1 years; height, 172.1 ± 6.5 cm; body mass, 64.2 ± 8.1 kg) participated in the present study. The sample size was estimated for two-way analysis of variance (ANOVA) (α error = 0.05, statistical power = 0.80, effect size = 0.4) using G*Power statistical power analysis software. The calculated sample size was 15. No subjects had any recent history of lower limb musculoskeletal injuries or neuromuscular disorders. The procedure, purpose, risks, and benefits associated with the present study were explained to each subject, and written, informed consent was obtained from each prior to enrolment. The ethics review committee for experimental research involving human subjects at Kyushu Sangyo University approved the experimental protocols (Approval No. 2022-0010), which were conducted in accordance with the guidelines of the Declaration of Helsinki. Study design Subjects visited the laboratory on four occasions, with an interval between each visit of at least 48 h. The first visit involved a familiarization trial, and the subsequent three visits assessed the acute effect of NMES on changes in ankle ROM and passive torque after static stretching. One experiment was performed to test the effect of NMES over the nerve trunk on changes in ROM and passive torque by stretching, and another experiment was performed to test the effect of NMES over the muscle belly on changes in ROM and passive torque by stretching. The third experiment was performed to test the effect of prolonged sitting on changes in ROM and passive torque by stretching. The order of experiments was randomized for each subject. The subjects were asked to maintain their usual level of physical activity during the periods between experiments. Subjects were secured in an isokinetic dynamometer (Humac Norm CN77, CSMI, Stoughton, MA, USA) with their right knees fully extended (Fig. 1A). The center of the lateral malleolus of the right foot was visually aligned to the rotational axis of the dynamometer. The right foot was firmly secured to the footplate of the dynamometer with a non-elastic strap. The angle of the back of the seat was 70° to the floor. In the present study, the ankle angle was defined as 0° when the footplate was perpendicular to the floor. Experimental setup and procedure. A Experimental setup. Subjects lay in the supine position with the right knee fully extended. Electrode locations for neuromuscular electrical stimulation (NMES) and electromyographic recordings for the triceps surae are shown. Black squares and circles indicate stimulating electrodes, and the white circles indicate recording electrodes. B Experimental procedure. The combined interventions of NMES and static stretching are performed. The passive ankle dorsiflexion test is measured before and after the interventions. Peripheral nerve stimulation (PNS) and maximal voluntary contractions (MVC) are performed before the interventions. NMES over the nerve trunk or muscle belly, or prolonged resting are performed on different days Surface electromyographic recordings Surface electromyographic (EMG) signals were recorded from soleus (SOL), medial gastrocnemius (MG), lateral gastrocnemius (LG), and tibialis anterior (TA) of the right lower leg (Fig. 1A). Ag-AgCl electrodes (Vitrode F-150 S, Nihon Kohden, Tokyo, Japan) with an inter-electrode distance of 20 mm were used for EMG acquisition from each muscle. The amplifier was set to a gain of 1000-fold with a bandpass filter between 5 Hz and 1 kHz (AB-611 J, Nihon Kohden). The EMG signals and torque signals were simultaneously sampled at 4 kHz using an analog-to-digital converter (PowerLab, ADInstruments, Melbourne, Australia) and stored on a personal computer using software (LabChart 7, ADInstruments). Root-mean-square (RMS) values of EMG signals from SOL, MG, LG, and TA during the passive ankle dorsiflexion test were determined for the initial 5° and final 5° of dorsiflexion, respectively (Mizuno 2023; Saito and Mizuno 2025). Surface EMG recordings were performed to ensure that the subjects relaxed their lower legs during the passive ankle dorsiflexion test. Consequently, no significant changes in RMS values of any muscles were identified between the initial and final 5° of dorsiflexion during the passive ankle dorsiflexion test in each condition (p = 0.084–0.981, r = 0.005–0.419). M-wave and H-reflex responses in EMG signals were processed using LabChart 7 (ADInstruments) and MATLAB (The Mathworks, Natick, MA) software. Peak-to-peak amplitudes of M-wave and H-reflex responses were calculated for each muscle. Neuromuscular electrical stimulation NMES was delivered either over the posterior tibial nerve or over the triceps surae using a constant current electrical stimulator (DS7R, Digitimer, Hertfordshire, UK) (Fig. 1A). The anode (5 × 5 cm2) was placed on the patella, and the cathode (diameter, 1 cm) was placed at the popliteal fossa. NMES over the triceps surae muscle belly was delivered through two electrodes. The anode (5 × 9 cm2) was placed over the proximal part of both MG and LG (5 cm below the popliteal fossa). The cathode (5 × 5 cm2) was placed over the superficial aspect of SOL and about 5 cm distal from where gastrocnemius joins the Achilles tendon. Experimental procedures Peripheral nerve stimulation was delivered to the posterior tibial nerve using the same electrodes for NMES over the nerve trunk (Fig. 1A). Subjects were instructed to relax completely during peripheral nerve stimulation. The recruitment curves of M-wave and H-reflex responses from SOL, MG, and LG were collected. Stimulation intensity increased gradually until no further increase was observed in the M-wave amplitude of SOL, MG, and LG using 1-mA increments. Maximal M-wave (Mmax) and twitch torque of the plantarflexors were evoked by setting the supramaximal stimulus intensity to 20% above the maximum experimental measurements. Mean supramaximal intensities to evoke Mmax for NMES over the nerve trunk, NMES over the muscle belly, and prolonged sitting conditions were 33.4 ± 10.3 mA, 32.2 ± 9.1 mA, and 32.8 ± 9.2 mA, respectively. Three Mmax values were collected at plantarflexion of 20° as baseline recordings. The passive ankle dorsiflexion test was performed before NMES and after static stretching. Passive dorsiflexion at 1°/s from plantarflexion of 20° to the maximal dorsiflexion angle determined based on the onset of discomfort for each subject was performed three times to define ankle ROM. The ankle joint angle was returned immediately to the plantarflexion position to avoid a stretching effect on stiffness of the muscle-tendon unit and stretch tolerance. Simultaneously, the passive torque on the footplate was recorded throughout the passive ankle dorsiflexion test. The subjects were asked to completely relax and not to exert any voluntary resistance. The maximal value of ankle ROM during the three passive ankle dorsiflexion tests was used for further analysis. The torque data were filtered with a zero-lag, 6 Hz Butterworth low-pass filter before passive torque was determined. Peak passive torque (i.e., stretch tolerance) was measured within a 250 ms epoch during the passive ankle dorsiflexion test (Kay et al. 2015). The slope of the passive torque-dorsiflexion angle curve (i.e., stiffness of the muscle-tendon unit) was measured at every fourth degree during the final 13° (i.e., at 1°, 5°, 9°, and 13°), which are common joint angles, before and after the intervention (Ryan et al. 2008, 2009). The mean value of the slope of the passive torque-dorsiflexion angle curve during the final 13° was used for further analysis. Following several warm-up trials, the subjects performed an isometric maximal voluntary contraction (MVC) of the plantarflexors for approximately 3 s. During MVCs, the right ankle joint was plantarflexed to 20°. Three MVC trials were performed to obtain the maximal torque. The interval between trials was set to 1 min. The averaged value of the highest two plantarflexion torques during MVCs was used for the representative value of each subject. NMES was applied to the nerve trunk or muscle belly for 20 min. NMES consisted of 60 stimulation trains (10 s ON, 10 s OFF). Pulse duration was 1 ms, because a wide pulse duration of 1 ms could activate relatively more sensory axons than narrow durations (0.05 and 0.25 ms) during NMES (Collins et al. 2002). Stimulation frequency was 10 Hz. The 10 Hz frequency was chosen because the SOL H-reflex remained markedly depressed throughout 10 s NMES at 20 Hz by post-activation depression, but the H-reflex amplitudes were initially depressed, and then the amplitudes recovered completely by the end of the train during 10 Hz NMES (Clair et al. 2011). Stimulation intensity was set before each NMES trial and adjusted to generate plantarflexion torque of approximately 5% of the MVC during NMES. If the NMES was uncomfortable, NMES was conducted with a stimulation intensity set lower than the target stimulation current. The stimulation intensity of NMES over the nerve trunk at the initial and last stimulation trains was 12.1 ± 3.7 mA and 13.2 ± 3.9 mA, respectively. The stimulation intensity of NMES over the muscle belly at the initial and last stimulation trains was 13.7 ± 5.9 and 14.3 ± 6.0 mA, respectively. NMES-induced plantarflexion torque generated during each 10 s stimulation train was measured over the centered 8 s period during the torque generation phase and normalized to each subject’s torque value during MVC. Peak-to-peak amplitudes of M-wave and H-reflex responses evoked by NMES were calculated using custom-written MATLAB software (The Mathworks) and then normalized to each Mmax. Static stretching was performed using the dynamometer after NMES or prolonged resting. The right ankle was passively dorsiflexed from plantarflexion of 20° to maximal dorsiflexion, fixed there for 60 s, and then moved to plantarflexion of 20°. The stretch protocol was then repeated four times with no rest interval between repetitions, giving a total stretch duration of 5 min. The maximal dorsiflexion angle was determined by passive ankle dorsiflexion testing before each stretch trial. Each subject was asked to relax the lower limb and not exert any voluntary resistance. Statistical analysis Data are provided as mean ± SD values. The normality of data distributions was investigated using the Shapiro-Wilk test for all data. When normality of data was not observed, nonparametric statistical tests were used. Differences in H-reflex amplitudes between the 1st and 60th 10 s stimulation trains of each muscle during NMES over the nerve trunk were compared using the Wilcoxon signed-rank test. The mean response amplitudes of the M-wave and H-reflex between muscles during NMES were analyzed using the Friedman test. When a significant effect was found, the Wilcoxon signed-rank test was performed as a post hoc test. Differences in NMES-induced plantarflexion torque between over the nerve trunk and over the muscle belly were compared using the Wilcoxon signed-rank test. The plantarflexion torques during MVC among NMES over the nerve trunk, NMES over the muscle belly, and with prolonged resting were analyzed using the Friedman test. The dorsiflexion angles during stretching were analyzed using two-way ANOVA with repeated measures, with stretch duration and NMES type used as factors. Differences in the maximal dorsiflexion angle, peak passive torque, and the slope of passive torque in NMES over the nerve trunk and over the muscle belly conditions were compared before and after interventions using a paired t-test. In the prolonged resting condition, differences in the maximal dorsiflexion angle were compared before and after interventions using a paired t-test, and the peak passive torque and the slope of passive torque were compared using the Wilcoxon signed-rank test. Changes in the maximal dorsiflexion angle and peak passive torque were compared among the three interventions by one-way ANOVA with repeated measures. Changes in the slope of passive torque were compared among the three interventions by the Friedman test. Differences in RMS values between the initial and final 5° of dorsiflexion during the passive ankle dorsiflexion test were analyzed for each muscle using the Wilcoxon signed-rank test. Significance was accepted for a value of p < 0.05, and p values for multiple comparisons were adjusted by the Bonferroni correction. For the paired t-test, the Wilcoxon signed-rank test, and ANOVA, Cohen’s d, r, and partial η2 were respectively calculated as the effect sizes indices. Results NMES The H-reflex amplitudes of SOL, MG, and LG evoked at the 1st 10 s stimulation train of NMES were 17.1 ± 13.3%, 4.3 ± 3.1%, and 9.0 ± 9.3% of Mmax, respectively, and the H-reflex amplitudes of SOL, MG, and LG evoked at the 60th 10 s stimulation train of NMES were 14.2 ± 14.7%, 3.6 ± 2.5%, and 8.4 ± 10.0% of Mmax, respectively. No significant differences in H-reflex amplitudes between the 1st and 60th 10 s stimulation trains were observed in SOL (p = 0.687, r = 0.097), MG (p = 0.795, r = 0.063), and LG (p = 0.723, r = 0.086). Regarding all H-reflex responses throughout NMES over the nerve trunk, the mean response amplitudes of the H-reflex were higher in SOL (p = 0.003, r = 0.843) and LG (p = 0.003, r = 0.820) than in MG. The mean response M-wave amplitudes of LG were higher than those of SOL (p = 0.003, r = 0.820). Unfortunately, M-wave and H-reflex responses of each muscle were not collected during NMES over the muscle belly because the EMG signals were saturated by stimulus artifact. During NMES over the nerve trunk, plantarflexion torque oscillated, and H-reflex amplitudes alternated between large and small within a 10 s stimulation train, but H-reflex responses were observed in each muscle from the initial to the last 10 s stimulation trains (Fig. 2A and B). During NMES over the muscle belly, plantarflexion torque was relatively stable compared with NMES over the nerve trunk throughout NMES (Fig. 2G and H). For example, one subject, whose data are shown in Fig. 2, had plantarflexion torque at each 10 s stimulation train during NMES over the muscle belly that ranged from 3.2% to 5.2%MVC, and the torque during NMES over the nerve trunk ranged from 1.3% to 6.1%MVC. Mean plantarflexion torques during NMES over the nerve trunk and muscle belly were 3.1 ± 0.8 Nm and 3.3 ± 1.2 Nm, respectively. Relative values of mean plantarflexion torque during NMES over the nerve trunk and muscle belly were 4.0 ± 0.9% and 4.2 ± 1.4% of MVC, respectively. No significant differences between NMES over the nerve trunk and over the muscle belly were observed in the mean values (p = 0.496, r = 0.176) and relative values of plantarflexion torque (p = 0.427, r = 0.205). Plantarflexion torque during MVC in NMES over the nerve trunk was 80.4 ± 20.4 Nm, in NMES over the muscle belly was 80.9 ± 19.7 Nm, and with prolonged resting was 79.7 ± 22.9 Nm. No significant difference in MVC torque was observed among these conditions (Friedman test p = 0.253). Typical examples of the exerted torque and H-reflex responses during NMES from a single subject. The exerted torque traces (A and B) and H-reflex amplitudes of triceps surae (C and D) during the 1st 10 s train and the 60th 10 s train of NMES over the nerve trunk are shown. H-reflex responses of triceps surae are expressed as a percentage of each maximal M-wave (Mmax). The waveforms indicate the evoked responses of triceps surae during the middle 10 s train of NMES over the nerve trunk (E and F). The exerted torque traces during the 1st 10 s train and the 60th 10 s train of NMES over the muscle belly are shown (G and H). SOL: soleus; MG: medial gastrocnemius; LG: lateral gastrocnemius Static stretching A significant main effect of stretch duration was observed in dorsiflexion angles during static stretching on two-way ANOVA (F(2.792,44.668) = 8.206, p < 0.001, partial η2 = 0.339), and no significant main effect of NMES type (F(2,32) = 2.180, p = 0.130, partial η2 = 0.120) and no interaction (F(3.957,63.307) = 0.828, p = 0.511, partial η2 = 0.049) were found. In NMES over the muscle belly, the dorsiflexion angle during stretching was higher with the 3rd (p = 0.026, d = 0.342) and 5th stretches (p = 0.034, d = 0.336) than with the 1st stretch, and the dorsiflexion angle was higher with the 3rd stretch than with the 2nd stretch (p = 0.029, d = 0.204). Passive ankle dorsiflexion test A significant change in the maximal dorsiflexion angle was obtained before and after NMES over the nerve trunk (p < 0.001, d = 0.458), NMES over the muscle belly (p = 0.013, d = 0.281), and with prolonged resting (p = 0.001, d = 0.285). No significant difference in the change in the maximal dorsiflexion angle was found among NMES over the nerve trunk (3 ± 2°), NMES over the muscle belly (2 ± 3°), and with prolonged resting (2 ± 2°) on ANOVA (F = 1.249, p = 0.300, partial η2 = 0.072). A significant increase in peak passive torque by stretching was observed in NMES over the nerve trunk (p = 0.011, d = 0.225) and NMES over the muscle belly (p = 0.045, d = 0.172), but not with prolonged resting (p = 0.586, r = 0.131). A significant difference in the change in peak passive torque was not seen among NMES over the nerve trunk (9.0 ± 10.2%), NMES over the muscle belly (6.9 ± 14.6%), and with prolonged resting (1.5 ± 8.7%) on ANOVA (F = 1.944, p = 0.160, partial η2 = 0.108). The slope of passive torque was significantly decreased by NMES over the nerve trunk (p < 0.001, d = 0.289), NMES over the muscle belly (p = 0.002, d = 0.152), and with prolonged resting (p = 0.002, r = 0.752). A significant difference in the change in the slope of passive torque was not observed among NMES over the nerve trunk (− 13.2 ± 7.1%), NMES over the muscle belly (− 8.8 ± 11.3%), and with prolonged resting (− 11.5 ± 11.5%) (Friedman test p = 0.465). Discussion The purpose of the present study was to compare the acute effects of NMES applied over the nerve trunk with those of NMES applied over the muscle belly on changes in ROM and passive torque by static stretching. The main findings were that (1) stretch tolerance increased after stretching with NMES over the nerve trunk and NMES over the muscle belly, and (2) increases in ankle ROM and decreases in stiffness of the muscle-tendon unit after stretching were observed with each intervention. These results did not support the initial hypothesis. When delivering wide pulses of NMES over the nerve trunk at 10 Hz for 20 min, substantial H-reflex responses of three synergists of the triceps surae were observed (Fig. 2C and D). The H-reflex amplitudes of SOL during NMES over the nerve trunk were different between the present study (13.6 ± 12.6% of Mmax) and previous NMES studies (Bergquist et al. 2011a; Clair et al. 2011). The previous study reported that NMES over the tibial nerve for 8 s at 20 Hz produced contractions at 10% of MVC torque, and the H-reflex amplitudes of SOL were approximately 5% of Mmax (Bergquist et al. 2011a). The size of H-reflex responses during NMES would be affected by NMES parameters. It has been demonstrated that spinal excitability of SOL was differentially modulated by the stimulation frequency of NMES (Vitry et al. 2019, 2021). Inhibition of the spinal excitability of SOL was observed after a 20 s train of NMES at 20 Hz that could be attributed to presynaptic inhibitory mechanisms involving increased homosynaptic post-activation depression (i.e., reduction of neurotransmitter release at Ia terminals), whereas facilitation of spinal excitability occurred after the train of NMES at 100 Hz that could be ascribed to enhanced motoneuron excitability as a result of persistent inward current activation (Vitry et al. 2021). In the case of repetitive deliveries of NMES applied over the nerve trunk for a long time, frequency-dependent NMES effects on corticospinal excitability were observed (Mang et al. 2010). A previous study examined the effect of NMES over the common peroneal nerve for 40 min (20 s ON, 20 s OFF cycle) on changes in corticospinal excitability of TA, and it showed that facilitation of corticospinal excitability occurred after 24 min during NMES at 100 Hz, but these modulations were not observed during NMES at 10 Hz (Mang et al. 2010). Although modulation of homosynaptic post-activation depression may be involved in modulation of spinal reflex excitability over time during low-frequency NMES, H-reflex amplitudes of each muscle remained between the initial and last trains of NMES in the present study. Ankle ROM increased with NMES over the nerve trunk (Fig. 3A), and the increase in ROM by stretching (3.4°) was not different with NMES over the muscle belly (2.2°) and with prolonged resting (2.2°). These results suggest that NMES-induced muscle contractions are not effective for improving ROM by static stretching. The evidence to support the effect of NMES over the muscle belly on changes in ROM of stretching has not shown consistency. The use of NMES applied over MG at 70 Hz (Mizuno 2023) and over the hamstrings at 50 Hz (Piqueras-Rodríguez et al. 2016) with stretching resulted in greater changes in ROM than stretching alone, but no added effect was observed with NMES applied over MG at 70 Hz (Mizuno 2021) and over triceps surae at 100 Hz (Capobianco et al. 2018) during stretching to increase ROM. It was demonstrated that stretching intensity determines changes in ROM of static stretching as the underlying mechanism that affects ROM (Freitas et al. 2015). Another study reported that static stretching performed to increase ROM had a greater effect on the maximal tolerable joint angle than on the submaximal angle (Kataura et al. 2017). In the present study, stretching intensity did not differ among the three interventions. Therefore, stretching intensity is a potential factor that may have caused no effect of NMES over the nerve trunk on ankle ROM prior to stretching. Effects of interventions on ankle joint flexibility The bar graphs show the maximal dorsiflexion angle (A), peak passive torque (B), and slope of passive torque (C) before and after the interventions. Each black circle represents a data point for each subject. NMES: neuromuscular electrical stimulation Stretch tolerance increased after static stretching with NMES over the nerve trunk and over the muscle belly, but not with prolonged resting (Fig. 3B). These results suggest that NMES increases stretch tolerance by static stretching regardless of electrode location. Our previous study examined the effects of patterned electrical stimulation (10 pulses at 100 Hz every 1.5 s) or uniform electrical stimulation (one pulse every 150 ms) to the common peroneal nerve for 20 min on changes in ROM of static stretching. Consequently, NMES improved stretch tolerance regardless of stimulation pattern. Thus, stimulation location and the pattern of NMES may not be related to the increase in stretch tolerance after stretching. Improvement of ROM by static stretching was caused by increasing stretch tolerance rather than a change in muscle-tendon unit stiffness (Magnusson et al. 1996b), and subjects with greater joint flexibility had greater stretch tolerance than those with less joint flexibility (Halbertsma and Göeken 1994; Blazevich et al. 2012). As the underlying mechanisms associated with increased ROM following PNF stretching, the gate control theory has been proposed (Kay et al. 2015, 2016). The gate control theory of pain asserts that the perceived level of pain is derived from a balance between nociceptive and non-nociceptive afferent inputs (Melzack and Wall 1965). It has been suggested that pressure receptors activated during contractions could inhibit pain perception because pressure receptors are associated with larger myelinated neurons that connect to the same spinal interneurons as unmyelinated nociceptive fibers within the spinal horn (Moayedi and Davis 2013; Capobianco et al. 2018). In the context of stretch tolerance, the increased activity of pressure receptors could dampen nociceptive signals and thereby decrease pain perception, which allows individuals to stretch further (Kay et al. 2016). Since the plantarflexion torque during NMES over the nerve trunk was unstable compared with that during NMES over the muscle belly, the changes in intramuscular tension, pressure, and blood flow during NMES over time may have altered the afferent inputs of pressure receptors. However, the relative value of mean plantarflexion torque during NMES was similar between these two NMES conditions. Therefore, the greater stretch tolerance by static stretching with NMES may be partly explained by the gate control theory. Stiffness of the muscle-tendon unit decreased in NMES over the nerve trunk (Fig. 3C), and the reduction in muscle-tendon unit stiffness by stretching (− 13.2%) was similar to that with NMES over the muscle belly (− 8.8%) and with prolonged resting (− 11.5%). The previous study showed that 1-min static stretching acutely reduces muscle stiffness, and PNF stretching reduces both muscle stiffness and tendon stiffness (Kay et al. 2015). Moreover, NMES over the muscle belly has an additional effect on changes in tendon displacement at the maximal dorsiflexion angle by 5 min static stretching (Mizuno 2023). Unfortunately, there was no greater NMES effect on the change in stiffness of the muscle-tendon unit after 5 min stretching (Fig. 3C). Our previous study also showed that reductions in stiffness of the muscle-tendon unit after NMES over the nerve trunk and 3 min stretching were similar to those with stretching without NMES (Saito and Mizuno 2023). It has been suggested that 1 min static stretching sufficiently decreased the stiffness of the muscle-tendon unit of the plantarflexors (Ryan et al. 2009). Although greater mechanical changes might be applied to the tendon after NMES and stretching interventions, the present study could not distinguish changes in the stiffness of the muscle and tendon. This study had several methodological limitations. The hips of the subjects were flexed 70° during the experiment to tightly fix the hips to the seats of the dynamometer during the passive ankle dorsiflexion test. Various tissues in the lower limb muscles and nerves are stretched in the hip flexed position compared with the neutral position. For example, changing the hip joint angle alters the degree of nerve slackness, which affects fibular nerve conduction velocity while keeping the knee joint extended (Broadhurst and Robinson 2017). Therefore, posture may have affected the changes in ROM and passive torque by static stretching in the present study. Furthermore, M-wave and H-reflex responses were not evoked during NMES over the muscle belly due to stimulus artifact. The present study could not collect clear reflex responses from triceps surae during NMES over the muscle belly in contrast to the previous study (Bergquist et al. 2011a). Finally, the differences in the neural pathways between NMES over the nerve trunk and NMES over the muscle belly were not clarified in the present study. Conclusion The present study examined the difference in the acute effects of NMES applied over the nerve trunk and of NMES applied over the muscle belly on changes in ROM and passive torque by static stretching. The findings indicate that NMES tended to facilitate greater stretch tolerance by static stretching regardless of the electrode location over the nerve trunk or muscle belly. Author contributions Akira Saito conceived and designed the research, performed the experiments, analyzed the data, interpreted the results of the experiments, prepared the figures, drafted the manuscript, edited and revised the manuscript, and approved the final version of the manuscript. Takamasa Mizuno conceived and designed the research, interpreted the results of the experiments, edited and revised the manuscript, and approved the final version of the manuscript. Funding The present study was supported by the CASIO Science Promotion Foundation (No.58) and JSPS KAKENHI (25K03000) to AS. Data availability No datasets were generated or analysed during the current study. Declarations Conflict of interest The authors declare no conflict of interests. Footnotes Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. References Associated Data Data Availability Statement No datasets were generated or analysed during the current study.