The effect of transcutaneous electrical nerve stimulation on fatigue recovery in recreational rowers

Participants

Twenty-eight participants were recruited and randomly divided into a TENS group and a sham-TENS group using a computer-generated randomization program implemented in SPSS (Version 20; IBM Corp., Armonk, NY, USA). The inclusion criteria for the participants were as follows: (1) college recreational rowers (training 2–3 times a week for 2 h each time); (2) aged 19–24 years. The exclusion criteria were as follows: (1) participants with lower limb muscle or bone injuries in the past 3 months; (2) participants who had ingested caffeine, alcohol, or any sports drinks within 24 h before the testing. The general characteristics of the participants are included in Table 1. Written informed consent was obtained from all participants. The study was approved by the Ethical Committee of the Shanghai University of Sport before testing commenced (Reference No. 102772021RT081).

Experimental design

The participants arrived at the laboratory on the experimental day 2 h after eating a meal. The familiarization was performed prior to the assessment but with a 20-min rest before the actual testing was executed, and this was mandatory for all participants. A single-blind experimental design was employed, with each trial involving only one participant. The participants were not aware of the specific sensations of the TENS intervention in this experiment prior to the trial. In the sham-TENS group, the placebo was administered with the device powered on but not activated. Therefore, the procedure was consistent for both groups. First, a standardized warm-up (consisting of running at 9 km/h for 5-min, followed by dynamic lower-limb stretching exercises (Silva et al., 2017)). The following data were collected before and immediately after a 2000-m rowing ergometer exercise and after a 20-min recovery intervention: (1) ratings of perceived exertion (RPE), (2) blood lactate (BLa), (3) peak torque (PT) and surface electromyography (sEMG) of the knee extensors for both limbs, and (4) heart rate (HR).

Fatiguing exercise was performed on a Concept II rowing ergometer (Model D; Concept II Inc., Rochester, NY, USA). Research has identified significant fatigue in the voluntary activation of knee extensors in athletes following a 2000-m rowing session (Husmann et al., 2017). Participants were asked to perform a 2000-m rowing ergometer exercise, which is the standard distance at the Olympic Games and the World Rowing Championships and successfully mimics competition-induced fatigue (Husmann et al., 2017). The Concept II rowing ergometer provides a highly realistic simulation of water-based rowing and is a staple in the training regimen of rowing athletes. The exercise began with the initiation of a paddle stroke with a resistance coefficient of four for females and five for males, according to their respective training habits. Verbal encouragement was consistently provided throughout the entire process.

Measurements

Surface electromyography

sEMG was recorded from the vastus lateralis (VL) and the long head of the rectus femoris (RF) muscles during each PT test (Bagnoli™ Desktop System; Delsys, Natick, MA, USA). Prior to application of the electrode, the skin was shaved, scrubbed, wiped with alcohol pads and marked with a mark pen. Electrodes were placed on the muscle belly of the VL (2/3 of the line from the anterior superior iliac spina to the lateral side of the patella) and the RF (50% of the line from the anterior superior iliac spina to the superior part of the patella) according to the SENIAM guidelines (Hermens et al., 1999).

Surface EMG signals were sampled at 1,111 Hz and processed using custom-written programs (MATLAB R2023a; MathWorks, Natick, MA, USA). Surface EMG signals were first off-line band-pass filtered between 20 and 500 Hz using a fourth-order zero-lag Butterworth filter. To quantify the surface EMG time and frequency components at PT, a moving root mean square (RMS) was performed to detect the plateau of torque. Then, the RMS and median frequency (MDF) were calculated over a period of 1,000 ms corresponding to the previously identified plateau of torque. The EMG RMS and MDF of the VL and RF muscles recorded from PT after fatiguing exercise and recovery intervention, were normalized relative to their RMS and MDF values obtained from pre-fatigue PT.

TENS intervention

TENS interventions can be achieved by utilizing various stimulation parameters, namely, frequency and intensity (DeSantana et al., 2008). These can be categorized into three types: conventional TENS (high-frequency, low-intensity), acupuncture-like TENS (low-frequency, high-intensity), and intense TENS (high-frequency, high-intensity) (Mokhtari et al., 2020). Conventional TENS has been shown to block the transmission of nociceptive afferent fibers in the spinal cord by stimulating large-diameter group II myelinated afferent fibers (Mokhtari et al., 2020). Acupuncture-like TENS stimulates both small diameter myelinated (Aδ) and unmyelinated (C) afferent nerves. This subsequently activates extra segmental descending pain inhibitory pathways, thereby producing a spatially diffuse analgesic effect (Johnson, 2021). Intense TENS aims to activate Aδ fibers and block harmful messages through extra-nodal analgesic mechanisms, similar to acupuncture-like TENS (Johnson, 2021). In addition, it has been demonstrated to produce rapid and significant analgesia both during and after stimulation (Johnson, 2021). Accordingly, the stimulus variables selected for this study were high frequency and high intensity—specifically, 120 mA, 200 μs, and 150 Hz (Mokhtari et al., 2020).

The TENS pain management program of the Compex SP 2.0 (DJO Global Products Inc., Carlsbad, CA, USA) electrical stimulation trainer was utilized for recovery intervention in the TENS group for 20 min. Self-adhesive electrode pairs (2 pairs of 5 cm × 5 cm, 1 pair of 10 cm × 5 cm) were positioned on the VL (2/3 of the line from the anterior superior iliac spina to the lateral side of the patella), the vastus medialis (80% of the line between the anterior superior iliac spina and the joint space in front of the anterior border of the medial ligament), and the RF (50% of the line from the anterior superior iliac spina to the superior part of the patella) on both legs (Arumugam et al., 2020). Current intensity was determined according to the participant’s maximum tolerance to the electrical stimulation (Cuesta-Gómez et al., 2019). Every 5 min, the researchers asked the participants whether the sensation was comfortable or decreasing and adjusted the intensity of TENS accordingly based on the feedback. The sham-TENS group underwent a 20-min placebo TENS intervention with electrodes placed on the skin but no electric current, and was told “This type of stimulation is supposed to help recovery by using a subthreshold stimulus that you will not able to perceive”.

Statistical analysis

Statistical analyses were conducted using IBM SPSS Statistics (Version 20.0; IBM Corp., Armonk, NY, USA). Normal distribution of the data was tested using the Kolmogorov–Smirnov test, showing that all variables were normally distributed. An analysis of variance (ANOVA) test with repeated measures was used to determine significant interactions between time [prefatigue (T1), postfatigue (T2), and recovery (T3)] and group (TENS and sham-TENS) on RPE, BLa, PT of knee extensors, and sEMG activities. A two-way ANOVA with repeated measures was performed with two between-subject factors (muscle and group) and one within-subject factor (time: T1, T2 and T3) on EMG MDF and RMS. When a significant interaction effect was detected, post hoc analysis was performed for further analysis. When there was no interaction effect, a simple effects analysis was performed. Effect size was calculated as partial eta squared. Statistical significance was set at p < 0.05. Data are reported as means ± SD within the text and displayed as means ± SEM in the figures.

RPE

There was no significant interaction between time and group in RPE (F_{(2, 52)} = 1.069, p = 0.350, η²_p = 0.039). However, a main effect of time was observed for RPE scores, showing a similar increase for both groups at T2 (TENS:17.8 ± 1.1 vs. sham-TENS: 18.5 ± 1.5, p < 0.001) and a decrease (p < 0.001) at T3 (TENS: 9.7 ± 2.2 vs. sham-TENS: 9.4 ± 1.8).

HR

Only a main effect of time was observed for HR (F_{(2, 52) =} 794.355, p < 0.001, η²_p = 0.968), showing a similar HR increase immediately at T2 (p < 0.001) (TENS: 184.1 ± 9.2 bpm vs. sham-TENS: 180.4 ± 5.5 bpm) and a decrease at T3 (p < 0.001) for both groups (TENS: 102.7 ± 13.1 bpm vs. sham-TENS: 102.9 ± 10.1 bpm).

BLa

A significant interaction was observed in BLa between the two groups from T2 to T3 (F_{(1, 26)} = 6.768, p = 0.015, η²_p = 0.207), showing that the TENS group had a greater decrease (17.76 to 9.83 mmol/L) compared with the sham-TENS (17.70 to 11.36 mmol/L) (Fig. 1).

PT of knee extensors

A significant group-by-time interaction was observed in terms of PT (F_{(2, 106)} = 11.549, p < 0.001, η²_p = 0.179). Post hoc testing revealed a significant difference in PT between the two groups from T2 to T3 (F_(1,53) = 17.378, p < 0.001, η²_p = 0.247) but not between T1 and T2 (p = 0.131). PT showed a similar reduction from T1 to T2 in both groups (TENS: 87.6 ± 9.4 Nm vs. sham-TENS: 82.8 ± 13.6 Nm; p < 0.001) but a different increase from T2 to T3; a greater increase/recovery was observed in TENS (108.4 ± 13.9 Nm, p < 0.001) compared to sham-TENS (91.7 ± 16.0 Nm, p < 0.01) group (Fig. 2).

sEMG

RMS

No time by group interaction was observed in EMG RMS (F_{(2, 214)} = 1.268, p = 0.268, η²_p = 0.012). Only a time main effect was observed in RMS (F_{(2, 206)} = 5.797, p = 0.004, η²_p = 0.053), showing that RMS at T3 was greater than at either T2 (p < 0.001) or T1 (p < 0.001) in both groups (Fig. 3).

MDF

There was no significant interaction between time and group (F_(2,214) = 0.997, p = 0.371, η²_p = 0.009) in MDF. However, an interaction effect between time and muscle (F_(2,214) = 9.560, p < 0.001, η²_p = 0.082) was observed, showing that the VL did not change throughout the testing, but the RF decreased at T2 (p < 0.001) then increased from T2 to T3 (p < 0.001) (Fig. 4). Moreover, at T1, the MDF of the RF was greater than that of the VL (84.62 ± 20.58 vs. 71.80 ± 16.68 Hz; F_(1,107) = 13.685, p < 0.001, η²_p = 0.113) (Fig. 4).

RPE and HR

The 2000-m rowing ergometer exercise resulted in a similar increase in RPE and HR between the TENS and sham-TENS groups. After the 2000-m rowing ergometer exercise on the Concept II rowing ergometer, RPE surged to 17 or higher, while the immediate HR of both groups was over 180 bpm. These results demonstrated that comparable fatigue was successfully induced in both groups. RPE is a commonly employed subjective measure of exertion during physical activity, demonstrating correlations with a range of physiological metrics, including muscular force, heart rate, ventilation, respiratory rate, oxygen uptake, and BLa concentrations (Tucker, 2009). Borg (1982) affirmed that RPE serves as the foremost indicator of physical strain, integrating signals from peripheral working muscles and joints as well as from the central cardiovascular system, respiratory functions, and nervous system. Moreover, RPE is comprised of not only physiological factors but also psychological and affective components, alongside other mediators (Coquart et al., 2012). Therefore, the overall sensation of exertion experienced during exercise can be considered the conscious verbal expression of the integration of physiological and psychological factors (Tucker, 2009). The competence for rating perceived effort is typically evaluated by correlating RPE with objective measures, such as HR (Reinke & Schmitz, 2023). A high correlation between HR and RPE has been reported (Reinke & Schmitz, 2023). Therefore, RPE and HR changes proved that successful fatigue was induced by the 2000-m rowing ergometer exercise in both the TENS and sham-TENS groups.

The TENS intervention did not result in any significantly different alterations in RPE and HR between the TENS and sham-TENS groups following the 20-min intervention. However, previous findings on the effect of TENS on RPE and HR changes after fatigue were not consistent with the results of the present study. Cheung & Jones (2007) found that TENS significantly reduced perceived exertion during physical activity. The participants reported lower RPE scores when TENS was applied, indicating that TENS can help alleviate the perception of effort during exercise. Since the Borg scale index is used to evaluate subjective fatigue, it may be difficult to compare only these ratings between subjects, as perceptions of fatigue vary greatly for each individual (Al-Mulla, Sepulveda & Colley, 2011). Haddad et al. (2013) suggested that there may be additional psychobiological factors beyond fatigue, stress, delayed onset of muscle soreness and sleep that influence RPE during exercise. Indeed, TENS has been shown to significantly accelerated the recovery of HR after treadmill running, with participants who received TENS showing a faster reduction in HR compared to those who did not receive the treatment (Cheung & Jones, 2007). Psychological factors can also influence HR. For example, emotional states, such as anger, fear, and dread, have been demonstrated to significantly affect HR (Al-Mulla, Sepulveda & Colley, 2011). In this case, the sham-TENS group received a placebo intervention and performed the same operation as the other group, with members of both groups having electrodes placed on their legs. This produced a psychological placebo effect, which may have contributed to the lack of a significant difference in RPE and HR between the two groups after the intervention.

BLa

Following the 2000-m rowing ergometer exercise, both the TENS and sham-TENS groups exhibited a comparable rapid increase in BLa levels (exceeding 18 mmol/L) (84.32% vs. 84.34%). This phenomenon can be attributed to the fact that lactate is produced at a faster rate than the body’s capacity to remove it, resulting in its accumulation during high-intensity exercise (Johnson, 2021). Furthermore, the accumulation of lactate lowers the pH of the blood, thereby exacerbating muscle fatigue (Flotyńska et al., 2021). Therefore, it is important to ensure the timely reduce of lactate to maintain a high level of competitive performance (Hinojosa, Hearon & Kowalsky, 2021).

Compared to the sham-TENS group, the TENS group exhibited a much higher lactate removal rate (47.03% vs. 36.25%) after the recovery phase. After maximal intensity exercise, active recovery removes accumulated lactic acid from the blood at a faster rate than passive recovery, thereby facilitating the release of lactic acid produced by the muscles for uptake by the exercising and nonexercising muscles (Bangsbo et al., 1993). TENS has been reported to stimulate damaged frontal nerve cell tissue and enhance local blood circulation (Preetam et al., 2024). Additionally, TENS-induced vasodilation improves oxygen delivery and facilitates more effective blood lactate reduce (Naik et al., 1999). The lactate is oxidized and used as fuel by the muscles, heart, and brain (Proia et al., 2016). It has been demonstrated that electrical stimulation is effective in increasing the return of venous blood to the heart, thereby increasing cardiac output (Grunovas et al., 2007). Therefore, compared to natural recovery from fatigue, TENS intervention offers significant advantages.

PT of knee extensors

After completing the 2000-m rowing ergometer exercise, the TENS and sham-TENS groups experienced a similar decline (12.41% vs. 17.22%) in PT of both limbs (Fig. 2). Muscle fatigue primarily stems from diminished neural control over muscle activities, thereby reducing force-generating and joint motion capability (Liu et al., 2021). Furthermore, the 2000-m ergometer exercise on the Concept II rowing ergometer likely prompted a substantial glycolytic response among athletes, possibly leading to excessive H⁺ accumulation and subsequent cellular pH reduction (Lancha Junior et al., 2015). This acidic environment may disrupt myosin and actin interactions, consequently impairing athletes’ strength and performance quality.

Compared to the sham-TENS group, the TENS group exhibited a superior recovery (23.7% vs. 10.49%) in PT of the knee extensors of both limbs. This finding confirms our hypothesis and may be attributed to several factors. First, the direct activation of muscle fibers by TENS increases muscle activity levels, thus increasing muscle strength after the intervention (Sartori et al., 2024). Second, TENS enhances neural adaptation by increasing motor unit recruitment and firing rates to provide additional neural stimulation, potentially increasing muscle strength (Bax, Staes & Verhagen, 2005). Third, TENS modulates cardiovascular function by acting on the autonomic nervous system, indirectly improving circulation and increasing blood flow to the skin, which helps accelerate blood flow and recovery from postexercise fatigue (Tomasi et al., 2015). Fourth, TENS reduces pain by activating the body’s inhibitory mechanisms to reduce the excitability of the central nervous system (Dailey et al., 2013); this effect is achieved by stimulating large afferent fibers that, in turn activating downstream inhibitory systems, thereby reducing pain sensations (Dailey et al., 2013). Meanwhile, TENS stimulates the release of endogenous opioid peptides (e.g., endorphins), which reduce pain perception (Astokorki & Mauger, 2017). Another potential mechanism is that TENS may play a role in reducing peripheral sensitization of muscle afferent nerves, a process that may occur after intense exercise and contribute to delayed onset muscle soreness. In addition, TENS may influence both peripheral and central fatigue, as well as blood flow to the muscle (Zambolin et al., 2024). Therefore, TENS may facilitate fatigue recovery and enhance muscle strength by increasing muscle activation and neural adaptation, modulating cardiovascular function and reducing pain and peripheral sensitization of muscle afferent nerves (Johnson et al., 2022).

Surface EMG

The sEMG analysis revealed changes in the quadriceps muscles of both groups following fatigue without time or group interaction. RMS is employed in the analysis of muscle activity, with a particular focus on the evaluation of muscle fatigue (Rampichini et al., 2020). It represents a prevalent metric for assessing the characteristics of sEMG time-domain signals, quantifying signal power in a manner that reflects alterations in signal amplitude (Del Vecchio et al., 2017). This approach offers a means of evaluating muscle activation patterns (Del Vecchio et al., 2017). The MDF is an index used in studies of spectral shifts and can be defined as “the frequency which divides the power spectrum in two parts with equal areas” (Al-Mulla, Sepulveda & Colley, 2011). A reduction in this MDF is frequently associated with muscle fatigue, as fatigue results in a shift in the frequency component of the EMG signal toward lower frequencies (Husmann et al., 2017). As muscle strength declines and fatigue sets in, the time-domain features of sEMG (RMS) rise, while the frequency-domain features (MDF) decline (Luttmann, Jger & Laurig, 2000).

No significant time or group interaction was observed in RMS following the recovery. A study by Behm et al. (2019) found that pre-exercise TENS significantly reduced RMS in the biceps femoris, which may be related to the delayed onset of fatigue. The main reason for the inconsistent discoveries may be explained by the different intervention timing of TENS. However, the reason why TENS applied before and after fatigue produced these controversial results remains unclear. It is speculated that variations in the intensity, duration and frequency of TENS might contribute to its effect on fatigue recovery. Further investigation is needed to valid this. However, RMS increased similarly following the 20-min recovery between both groups. The interference EMG signals can only provide a more ‘global’ view on the EMG responses in the time and frequency and therefore do not necessarily truly reflect motor unit (MU) recruitment strategies and individual MU properties (Keenan et al., 2006; Del Vecchio et al., 2017). Although the similar increase at T3 was observed in EMG RMS between two groups, the underlying neurophysiological mechanisms may differ. In the TENS group, this increase could be attributed to the enhanced muscle activation and pain-reducing effects of TENS (Tousignant-Laflamme et al., 2017). Indeed, Chan et al. (1999) have reported that TENS can increase the fatigue resistance of the larger and more fatigable MUs, potentially leading to an increase in overall EMG amplitude. On the other hand, in the Sham-TENS group, the increase in EMG amplitude may have resulted from natural recovery, in which the MUs with smaller amplitudes were recruited (e.g., slow-twitch MUs) (Jensen, Pilegaard & Sjøgaard, 2000). This suggests that, in sham-TENS group, more but small MUs may have been activated to achieve the same increase in EMG amplitude as observed in TENS group. In addition, an interaction effect between time and muscle observed in EMG MDF showing that VL did not change throughout the test, but the RF decreased after the fatigue exercise and then increased after the recovery phase. This is consistent with the findings of previous study (Mullany et al., 2002). First, it may be attributed to potential differences in muscle fiber type between the RF and the VL muscles (Wu et al., 2019). The proportion of type II muscle fibres in RF is greater than in VL muscles, which are less fatigue-resistant and more easily fatigued during longer and higher-loaded contractions (Mathur, Eng & MacIntyre, 2005). This could lead to a decrease in the post-fatigue MDF. Second, in rowing (i.e., multi-articular movements), the RF, as a bi-articular muscle, may recruit more motor units at an earlier stage during knee extension than a mono-articular biarticular muscle (i.e., VL), resulting in greater fatigue (Mathur, Eng & MacIntyre, 2005). This distinct recruitment strategy may result in a more pronounced decline in the MDF of the RF following fatigue. Furthermore, research has demonstrated that the MDF performance of the RF and VL may vary in response to different contraction intensities (Vøllestad, 1997). Specifically, the MDF decline in the RF may be more pronounced during high-load (80% MVC) contractions, while the difference may be less evident during low-load (20% MVC) contractions (Mathur, Eng & MacIntyre, 2005). Additionally, the observation that the MDF of the RF declines after fatigue and then increases after recovery may be linked to the physiological recovery process of the muscle following fatigue. This suggests that the RF may undergo a rebound in its EMG activity characteristics after a period of recovery, while the VL may not exhibit this apparent dynamic change (Vøllestad, 1997). In conclusion, these factors contribute to the differences in the sEMG amplitudes between the RF and VL under specific conditions.

Limitation

The present study comes with some limitations. Firstly, the study adopted a medium effect size (0.25) to determine the sample size, resulting in a total requirement of 28 participants. This approach may lack methodological rigor. Secondly, Green, Parro & Gabriel (2014) reported a minimum of 10, but no more than 15 contractions were necessary to constitute a familiarization session of isometric dorsiflexion contractions. In this study, only one familiarization session was implemented before testing. Future studies should incorporate adequate familiarization procedures. Thirdly, blood lactate was assessed from capillary blood taken from the fingertip. Feliu et al. (1999) reported that blood lactate concentrations collected from fingertip and earlobe could be different. Future research should consider collecting the blood sample from the earlobe for blood lactate analysis. The fourth limitation is that we had to temporarily remove the EMG electrodes during the TENS recovery intervention for the placement of the TENS electrodes. Removing and reattaching EMG electrodes likely adds unexplained variance, although the location of the EMG electrodes was marked on the skin to ensure the consistency. Future investigations should systematically address this concern to minimize potential data discrepancies induced by EMG electrode detachment. Lastly, this study focused exclusively on the immediate effects of TENS interventions on fatigue recovery. Future research could explore the long-term effects of TENS on fatigue recovery as well as its application in other muscle groups.