Five Weeks of Sprint Interval Training Increase Absolute Power Output Within Severe‐Intensity Domain Without Altering Muscle Activation or Fatigability

ABSTRACT

The study aimed to investigate whether a 5‐week sprint interval training (SIT) could improve neuromuscular function (NMF), alter electromyographic activity (EMG), and attenuate fatigability, ultimately resulting in improved cycling performance during constant work rate (CWR) exercise within the severe‐intensity domain. Twenty‐one recreational active males were randomly assigned into a SIT group (SIT_G; n = 11) that underwent 5 weeks of SIT (4–6, 30‐s Wingate bouts) and a control group (CON; n = 10). The study was conducted in a parallel independent group study. At the pre‐ and post‐training periods, participants performed (i) a ramp incremental test; (ii) CWR exercise at 150% of the pre‐ and post‐training power output at the respiratory compensation point (RCP‐PO) until task failure, with the NMF assessed at baseline (NMF_BL) and immediately after exercise [isometric maximal voluntary contraction + peripheral nerve stimulations with measurement of the force and EMG]. Pre‐ and post‐training NMF_BL was used to assess isometric performance, whereas the difference between NMF_BL and NMF after exercise was used to assess performance fatigability. Furthermore, EMG was also measured during exercise. SIT_G showed a significant increase in absolute RCP‐PO compared to CON (+11.0% vs. −0.7%, respectively; p < 0.01), performing the post‐training CWR at a greater power output than in pre‐training. Despite this, no significative differences were observed for the NMF_BL, fatigability, and EMG during CWR exercise. These findings suggest that SIT may improve EMG activity efficiency (i.e., requiring the same level of activation to sustain a greater absolute power output) during CWR exercise within the severe‐intensity domain. However, NMF_BL, fatigability, and time‐to‐task failure performance remained unchanged.

Highlight Summary

• Five weeks of sprint interval training increased the absolute power output representing 150% of RCP‐PO, whereas muscle activation (EMG) was maintained, indicating an improved neuromuscular efficiency during severe‐intensity exercise.

• Despite greater total work, time‐to‐task failure and performance fatigability remained unchanged after training.

• SIT also increased PPO, $\dot{\mathrm{V}}$O_2max, and $\dot{\mathrm{V}}$O₂ at RCP and RCP‐PO parameters, consistent with previous findings.

1. Introduction

During cycling exercises, the neuromuscular system plays a pivotal role in providing the stimuli for muscular contractions and sustaining repetitive movements of the lower limbs (Faria et al. 2005; Takaishi et al. 1996). Given the stimulus‐dependent nature of neuromuscular adaptations (Rich and Cafarelli 2000), and with the goal of enhancing responses within the neuromuscular system, high‐intensity exercises are frequently utilized to maximize enhancements in neuromuscular function (NMF) (Cormie et al. 2011). In connection with this, all‐out sprint interval training (SIT) represents a common strategy to induce positive adaptations within the cardiovascular system (Burgomaster et al. 2008; Gibala et al. 2006; MacInnis and Gibala 2017) and emerges as a strong candidate to optimize neuromuscular gains as it elicits high peripheral demands (Buchheit and Laursen 2013; Creer et al. 2004; Hureau et al. 2016; Lloria‐Varella et al. 2023, 2024).

The SIT efforts (often comprising four to six trials of 30‐s Wingate all‐out bouts) exhibit an initial surge in power output within the first 5–10 s of exercise, followed by a significant performance decline (i.e., performance fatigability) until the end of the bout (Malta et al. 2020; Takei et al. 2021), and are characterized by very high metabolic and neuromuscular demands (Mendez‐Villanueva et al. 2008). During the execution of all‐out efforts, such as those performed during SIT, a notable accumulation of metabolites, such as hydrogen ions and inorganic phosphate (Bishop and Edge 2006; Gaitanos et al. 1993), leads to a significant disturbance in neuromuscular function, resulting in a decrease in motor unit recruitment and/or firing rate (Hunter et al. 2003), muscle fiber conduction (Stewart et al. 2011; Takei et al. 2021) and excitation‐contraction coupling during the effort (Mendez‐Villanueva et al. 2008; Stewart et al. 2011; Takei et al. 2021), ultimately inducing performance fatigability, mainly at the peripheral level (Fernandez‐del‐Olmo et al. 2013; Fiorenza et al. 2019; Hureau et al. 2016; Lloria‐Varella et al. 2023). Therefore, chronic exposure to SIT would be expected to result in metabolic adaptations (e.g., enhanced glycolytic enzyme activity, increased muscle glycogen stores, and improved buffering capacity) (de Poli et al. 2024) that should also lead to positive changes in neuromuscular function (i.e., lower motor unit recruitment and/or firing rate to produce the same work) and improved exercise performance during constant work rate (CWR) exercise to task failure.

However, possibly due to the technical challenges associated with evaluating neuromuscular function, few studies investigated neuromuscular adaptations elicited by SIT during CWR exercise (Bertschinger et al. 2020). Bertschinger et al. (2020) reported no improvement in any parameters of the neuromuscular function or performance fatigability (i.e., development of fatigability during exercise) as well as time‐to‐task failure (TTF) or work performed during moderate intensity (i.e., below gas exchange threshold) exercise. This lack of difference may be attributed to the selection of a CWR exercise within the moderate intensity domain that resulted in minimal metabolic disturbance for the investigation of neuromuscular function (Brownstein et al. 2022).

Then, exploring the role of SIT training on neuromuscular adaptations during CWR exercise performed at higher intensities that induce greater peripheral disturbances is needed. For example, exercise intensities within the severe‐intensity domain, witch correspond to workloads above the maximal metabolic steady state (MMSS) that, although allowing for achievement of $\dot{\mathrm{V}}$O_2max, they result in non‐oxidative reserves being gradually consumed, accumulation of metabolites, and disruption of cellular homeostasis, ultimately lead to task failure at maximal oxygen uptake (Burnley and Jones 2007), might present a different scenario (Black et al. 2017; Iannetta et al. 2022). Although the etiology of task failure during exercises within severe‐intensity domain remains a subject of debate (Amann and Secher 2010; Fitts 2016; Marcora 2010; Westerblad 2016; Zagatto et al. 2022, 2025), it appears that exceeding some “critical” concentrations of metabolites (e.g., inorganic phosphate, hydrogen, and potassium ions) as a result of the high skeletal muscle contractile actions and use of nonoxidative sources for resynthesis of adenosine triphosphate can reduce muscle contractile capacity at the peripheral level (i.e., peripheral fatigue) (Black et al. 2017; Carroll et al. 2017; Hureau et al. 2022; Thomas et al. 2016). Additionally, despite these peripheral disturbances, it seems that these metabolic perturbations, particularly the accumulation of hydrogen ions, can potentially interfere with the response of group III and IV neural afferents, leading to a compromised central drive for muscle activation (i.e., central fatigue) (Hureau et al. 2022), although central fatigue does not appear to be severely affected during short‐term exercise performed within the severe‐intensity domain (Thomas et al. 2016).

Therefore, considering the importance of neuromuscular function, including both peripheral and neural components, such as motor unit recruitment and voluntary activation, for SIT (Fernandez‐del‐Olmo et al. 2013; Hureau et al. 2016) and during CWR performed within the severe intensity domain, it is reasonable to hypothesize that SIT would lead to positive adaptations at the muscle level. These adaptations may include increased force production, changes in muscle activation consistent with improved neuromuscular activation economy (i.e., similar or lower EMG amplitude despite a higher absolute workload), and reduced performance fatigability, which together could contribute to improving exercise performance within the severe‐intensity domain. Thus, this study aimed to investigate whether a 5‐week SIT program would improve neuromuscular function by increasing force production and muscle electrical activation while decreasing performance fatigability during CWR exercise within the severe‐intensity domain.

2. Materials and Methods

2.1. Participants

A minimum sample size of 8 participants per group was calculated to achieve a statistical power of 80% with an alpha level of 0.05, according to Cohen's guidelines. As no meta‐analysis with bias‐adjusted estimates is available on the neuromuscular effects of sprint interval training, the sample size estimation was based on a conservative smallest effect size of interest (SESOI) derived from roots mean square data (EMG‐RMS) adaptations after 4 weeks of training (Creer et al. 2004). From those data, a Cohen's f of 0.35 was estimated and used in G*Power 3.1 (Kiel, Germany) to compute the required sample size for a repeated‐measures ANOVA. A total of 21 healthy active males (i.e., exercising at least 3 times per week, who were participating in activities, such as football, futsal, jogging, and others, while not adhering to a regular training program), participated in the study (age 30 ± 4 years, stature 176.5 ± 4.7 cm, and body mass 76.4 ± 11.0 kg). No females took part in the present study as neuromuscular function and fatigability of the knee extensors can vary according to the menstrual cycle (Ansdell et al. 2019). Considering our experimental design (see Experimental design section for more details), it was possible that females might not have been assessed during the same phase of the menstrual cycle, thus potentially compromising the validity of the findings. Further, a subset of the volunteers in the current study also participated in a previous research investigation conducted by our research group as detailed elsewhere (de Poli et al. 2024).

Inclusion criteria required that participants were not involved in a periodized training regimen (i.e., performing physical exercise at least 3 times a week) and had no recent history of muscle or neuromuscular injury for at least 6 months prior to the study. Furthermore, only nonsmokers and individuals who did not undergo any medical treatment that could potentially affect their metabolic or neuromuscular responses to testing and exercise were enrolled. As required by the exclusion criteria, participants refrained from using performance ergogenic aids and any other training or exercise routine during the study period. Furthermore, participants had to complete at least 95% of the total scheduled training sessions to maintain eligibility for the study.

All participants were informed about the possible risks and benefits of the study and signed an informed consent document. All procedures were approved by the local Research Committee (Protocol #2.650.216/2018), and the study was conducted according to the Declaration of Helsinki.

2.2. Experimental Design

The study employed a parallel independent‐groups design with pre and post‐training measurements. During the pre‐training assessments, participants visited the laboratory three times in the same week: (i) the first visit involved familiarization with the CWR exercise and with NMF assessments to ensure maximum execution during subsequent evaluations; (ii) the second visit consisted of performing a ramp incremental test to task failure; and (iii) the third visit involved a CWR exercise at 150% of the power output associated with the respiratory compensation point (RCP‐PO). This test included surface electromyography (EMG) analysis during exercise and evaluation of neuromuscular function (including performance fatigability) before and after the trial. Following pre‐training measurements, participants were matched based on their performance in the ramp incremental test to ensure a balanced distribution of fitness levels and then randomly assigned to the SIT group (SIT_G; n = 11) or the control group (CON; n = 10), which did not engage in any regular training during the intervention period. Participants were instructed to refrain from performing any additional structured exercise outside the experimental sessions throughout the study and verbally confirmed adherence to these instructions. Although performance was considered prior to randomization, the allocation process was not stratified or restricted, which allowed for a slight imbalance in final group sizes. All analyses were conducted using an independent‐groups design.

After 5 weeks of intervention or control (post‐training), the same procedures applied in pre‐training measurements were repeated (i.e., the CWR exercise power output was performed at 150% of the posttraining RCP‐PO). A schematic of experimental design is shown in Figure 1.

FIGURE 1.

Study experimental design and training parameters. CWR, constant work rate cycling; RCP‐PO, respiratory compensation point power output; T1–T15, number of training sessions; and WT, Wingate.

All cycling exercise procedures were performed on an electromagnetic braked cycle ergometer (Excalibur Sport, Lode, Netherlands) at a self‐selected cadence between 80 and 100 rpm (±5 rpm of the chosen cadence during all exercises). All testing and training sessions were separated by a minimum interval of 48 h and a maximum of 72 h and were performed at a similar time of the day (±2 h) to avoid circadian influences. Participants were verbally encouraged to perform the maximum effort in all tests and training sessions. Task failure for all exercises was assumed as the inability to maintain the preselected cadence within 5 rpm for longer than 5‐s despite strong verbal encouragement or disengagement from the task.

2.3. Exercise Protocols

2.3.1. Ramp Incremental Test

The ramp incremental test was performed to determine $\dot{\mathrm{V}}$O₂ at the RCP and RCP‐PO as well as maximal oxygen consumption ($\dot{\mathrm{V}}$O_2max) and peak power output (PPO) (Iannetta, De Almeida Azevedo, et al. 2020). The test began with a 5‐min warm‐up at 75 W followed by a workload increase of 1 W every 2‐s period (i.e., 30 W per minute) until task failure (de Poli et al. 2024).

2.3.2. CWR Exercise Within the Severe‐Intensity Domain

The CWR was performed at 150% of RCP‐PO [i.e., corresponding to approximately 115% of PPO (Zagatto et al. 2017, 2023, 2025)] and this intensity was selected for two primary reasons: (i) the significant contribution of the glycolytic pathway (de Poli et al. 2024), enhancing the likelihood of inducing neuromuscular performance fatigability (de Poli et al. 2020), and (ii) ensuring a homogenous metabolic demand within and between individuals or testing performance (normalized by RCP‐PO). It should be noted that, although there may be some variation in relative intensity between participants, the severe‐intensity domain was previously confirmed in all participants by the achievement of $\dot{\mathrm{V}}$O_2max at the end of the trial performed at 150% of RCP‐PO (de Poli et al. 2024). In addition, CWR trials were performed at 150% of the RCP‐PO specific to the pre or post‐training evaluation to avoid neuromuscular influence from different relative intensities as varied relative intensities above RCP can affect exercise performance and neuromuscular response differently (Black et al. 2017).

The workload was applied after the participants achieved their self‐selected cadence, and then the TTF was recorded. The workload was acquired at 5 Hz by Lode test software (Excalibur Sport, Lode BV, Netherlands), and the total work performed was determined as the product of the constant load PO and exercise duration.

2.3.3. Sprint Interval Training

The SIT comprised four to six repeated bouts of 30‐s all‐out efforts (workload: 0.70 Nm·kg⁻¹ and rest: 4 min between each bout) (de Poli et al. 2024; Gibala et al. 2006). All sprints were performed in a seated position, and workload and performance variables (e.g., peak and mean power output, fatigue index) were automatically recorded using the Wingate Test Software (Lode BV). The entire training program included weekly adjustments in volume (i.e., increasing the number of bouts: week 1: 4 bouts; week 2: 5 bouts; and week 3: 6 bouts) and the application of a tapering period (i.e., reducing the training volume by approximately 16%–50%). During the tapering period (weeks 4 and 5), the participants underwent 6 bouts until the last four training sessions, during which they completed five, four, three, and three bouts, as previously reported by de Poli et al. (2024).

The training duration (i.e., 4 weeks) was chosen based on previous findings by (Creer et al. 2004), who reported adaptations in EMG‐RMS during high‐intensity efforts after 4 weeks of SIT. Additionally, given the cumulative physiological demand of repeated SIT sessions, a fifth week was included as a tapering period to maximize performance adaptations (Bosquet et al. 2007).

2.4. Data Collection

2.4.1. Ramp Incremental Respiratory Data

During the ramp test, gas exchange, and ventilatory responses were measured breath‐by‐breath using a metabolic cart (Quark CPET; COSMED, Italy). Before each ramp incremental test, the flowmeter and the gas analyzer were calibrated with a 3‐L syringe, ambient air, and a gas mixture of known concentration (16% O₂; 4.993% CO₂; and balance N₂), respectively.

2.4.2. Surface Electromyography Signal (EMG) During CWR Exercise

EMG signals from the vastus lateralis, rectus femoris, biceps femoris, and gluteus maximus muscles of the self‐reported propulsive leg were recorded during CWR exercise using a wireless device (Mini Wave, Cometa System, Italy). Before the electrodes were placed, the skin surface was gently shaved and abraded to remove dead skin cells. Subsequently, the area was cleaned and degreased with isopropyl alcohol to reduce the impedance (< 10 kΩ) (Hermens et al. 2000). Then, silver/silver chloride (Ag/AgCl) electrodes (recording area, 78.5 mm²; Medi‐trace, Dublin, Ireland) were placed lengthwise in the belly of each muscle to noninvasively measure surface EMG following established recommendations (Hermens et al. 2000). EMG signal data were acquired at 2000 Hz with PowerLab signal acquisition hardware (ADInstruments, New Zealand).

2.4.3. Neuromuscular Function

To assess NMF, a 5‐s maximal isometric voluntary contraction (iMVC) of dominant knee extension, combined with peripheral nerve stimulation (PNS) and recording of the vastus lateralis EMG signal, was performed both before (baseline) and immediately after CWR task failure (delay of 25 ± 7 s; post‐exercise) (Millet et al. 2011). This procedure was used to identify potential changes in both peripheral and central components of neuromuscular fatigability (Gandevia 2001; Millet et al. 2011).

2.4.3.1. Maximal Isometric Voluntary Contraction

Before beginning each trial, participants underwent a standardized ‘warm‐up' comprising 5‐min cycling at 75 W, followed by two iMVC intersected by 1‐min and concluding with 2‐min passive rest (Dutra et al. 2023). For iMVC, participants were positioned on a custom‐designed chair that maintained the knee and hip flexed at 90° and with the trunk and thighs fixed with a belt (de Poli et al. 2020; Dutra et al. 2023). An inextensible bar was attached above the region of the medial malleolus (e.g., ∼1–2 cm above; ankle region) using heavy‐duty Velcro straps, connected to a load cell (strain gauge sensibility: 2.00 ± 0.02 mV·V; maximum capacity: 100 kgf; 0.03% rated output accuracy; 0.98 N resolution; and MKControle, Brazil) (Dutra et al. 2023). The load cell signal was acquired by a PowerLab 16/35 digital‐to‐analog converter (ADInstruments, New Zealand) with a sampling rate of 2000 Hz. If the baseline iMVC performance was lower than that achieved during the ‘warm‐up,' the iMVC trial was repeated.

2.4.3.2. Peripheral Nerve Stimulation

Stimulations of the femoral nerve of the dominant knee extensors (pulse duration: 1‐ms; square wave) were delivered by a high‐voltage electrical stimulator (Biostimulator V3, 400 V peak to peak; Insight, Brazil) through Ag/AgCl electrodes (78.5 mm²; Medi‐trace, Dublin, Ireland) on the femoral nerve (cathode) and through carbon rubber electrode (32.5 mm of diameter, CARCI, Brazil) at the femoral greater trochanter level (anode). Prior to each NMF assessment, a consecutive incremental single stimulus on relaxed muscle was applied until the force plateau verification. Then, all electrical stimuli were delivered at 120% of the stimulus at which the force plateau was reached (Dutra et al. 2023), which corresponded to 302 ± 52 mA. Doublet high‐frequency stimulus (i.e., square wave, 100 Hz, and pulse duration of 1‐ms) was delivered during the iMVC (∼ during the 3^th second of iMVC; superimposed), followed by another doublet high‐frequency stimulus, doublet low‐frequency stimulus (10 Hz) and single (Q _tw), each applied at 5‐s interval after the completion of the iMVC in relaxed muscle.

2.4.3.3. Vastus Lateralis Surface Electromyography During the iMVCs

The Ag/AgCl electrodes were positioned 20 mm apart (center to center) in the direction of the muscle fibers and immediately above the region of the myotendinous junction detected visually after an isometric voluntary contraction (Dutra et al. 2023). Although the reference electrode was placed on the patella. All preparations (i.e., to reduce impedance) and signal acquisition was performed as previously described in surface electromyography during CWR exercise section.

2.5. Data Analysis

2.5.1. Ramp Incremental Respiratory Data

After data collection, raw gas exchange data were processed to eliminate outliers. A local mean of 5‐s bins were calculated, and data that exceeded more than 3 standard deviations from the local mean were removed (Lamarra et al. 1987). Then, the data were interpolated to 1‐s intervals using OriginPro 8.0 software (OriginLab, USA) (Iannetta et al. 2019).

The end‐tidal partial pressure of CO₂ (PCO₂), ventilation ($\dot{\mathrm{V}}$ _E), and ventilation/carbon dioxide production curve ($\dot{\mathrm{V}}$ _E/$\dot{\mathrm{V}}$CO₂) values were plotted against oxygen consumption ($\dot{\mathrm{V}}$O₂) data for RCP determination (Iannetta, Inglis, et al. 2020). The RCP was estimated by consensus of two experienced researchers and defined as the point at which the end‐tidal PCO₂ started to decrease, simultaneously with the second and first break points of $\dot{\mathrm{V}}$ _E and $\dot{\mathrm{V}}$ _E/$\dot{\mathrm{V}}$CO₂, respectively (Keir et al. 2022). After the $\dot{\mathrm{V}}$O₂ at RCP was identified, the RCP‐PO was determined based on the ramp incremental test $\dot{\mathrm{V}}$O₂ (y‐axis) versus PO (x‐axis) relationship. It should be noted that RCP‐PO was based on the linear relationship between $\dot{\mathrm{V}}$O₂ and PO observed during ramp incremental test, without correcting for the mean response time and slow component of $\dot{\mathrm{V}}$O₂, as recommended elsewhere (Iannetta et al. 2023; Iannetta, Inglis, et al. 2020). Despite that, the severe‐intensity domain was confirmed by the $\dot{\mathrm{V}}$O_2max values attained during CWR exercise in both the pre and posttraining phases, as demonstrated by research conducted by our group, detailed elsewhere (de Poli et al. 2024).

Finally, the $\dot{\mathrm{V}}$O_2max was calculated as the highest $\dot{\mathrm{V}}$O₂ value from a 20‐s rolling average, whereas the PPO was accepted as the highest intensity reached during the ramp incremental test (Iannetta, De Almeida Azevedo, et al. 2020).

2.5.2. Surface Electromyography (EMG) Signal During CWR Exercise

The EMG signal during CWR exercise was filtered with a band pass filter of 20–500 Hz and a notch (60 Hz) (Dutra et al. 2023). Evaluation of muscular activity was estimated by EMG‐RMS, which was calculated for four‐quarter time window according to TTF (i.e., 0%–25%, 25%–50%, 50%–75%, and 75%–100%).

Then, the raw EMG‐RMS data for each time window were normalized to the last 5‐s of the warm‐up phase at 75W (EMG‐RMS%), using a normalization protocol adapted from MacDougall et al. (2025). Normalization using the first 5‐s of warm‐up cycling was chosen because it may provide a more accurate representation of the neural drive during cycling compared to RMS from an iMVC (ecological response) (Rouffet and Hautier 2008), and fatigability within this time window is expected to be minimal (Dutra et al. 2022). Considering that the RMS amplitude increases with power output, normalizing to a fixed workload provides a fairer basis for detecting training‐induced changes (Bijker et al. 2002).

Additionally, an average over time windows was also performed (Ducrocq and Blain 2022).

2.5.3. Neuromuscular Function

2.5.3.1. Force at iMVC Plus PNS

Force data were filtered by a fourth‐order Butterworth filter in LabChart 11 (ADInstruments, New Zealand). The highest voluntary force recorded during iMVC force plateau (iMVC_Force) was assumed as mean of 100‐ms around peak value (± 50‐ms mean around peak value) (Behm et al. 1996). The force amplitude for each electrical stimuli (onset torque) was measured for determination of the force level at doublets high‐frequency stimulus (Db₁₀₀), low‐frequency stimulus (Db₁₀), and Q _tw. Then, low‐frequency muscle function (Db_10:100) was calculated as ratio between low and high frequency stimuli. In addition, the voluntary activation level (VA; central parameter) was assessed using Equation (1) (Strojnik and Komi 1998).

$$\text{VA}=1-\left({\text{Db}}_{100‐\text{superimposed}}\times \left(\frac{\text{Force}\text{level}\text{at}\text{stimulus}\text{moment}}{\text{iMVCForce}}\right)/{\text{Db}}_{100}\right)\times 100$$ (1)

The baseline (pre‐exercise) data at pre‐ and post‐training (i.e., 5‐week) periods were used to compare possible effects of intervention on baseline neuromuscular function (NMF_BL) (Dutra et al. 2023; Giuriato et al. 2024). To assess performance fatigability, both the absolute difference and the percentage change between pre‐ and post‐exercise values were used (Azevedo et al. 2022; Iannetta et al. 2022).

2.5.3.2. Vastus Lateralis Surface Electromyography During the iMVCs

The EMG data were analyzed in LabChart 11 software (ADinstruments, New Zealand) offline, filtered with a band pass filter of 20–500 Hz and notch (60 Hz). Subsequently, the RMS from 500‐ms of iMVC_Force (RMS_iMVC) was measured. In addition, without the band pass filter, the maximal amplitude and duration of M‐Wave evoked by the single stimuli was measured. The maximal amplitude of M‐Wave was also used to normalize RMS values (RMS_iMVC/M‐Wave) (Millet et al. 2011).

All parameters measured during the iMVC with peripheral neuromuscular stimulation (including force and EMG responses) showed good reliability and repeatability as recently demonstrated by our study group (Dutra et al. 2023).

2.6. Statistical Analysis

Data are presented as mean ± standard deviation (SD) and 95% confidence intervals (CIs 95%). Shapiro–Wilk and Levene tests were used to assess the normality distribution and homoscedasticity of the data. A two‐way mixed‐design general linear model ANOVA with repeated measures was conducted to examine potential differences in NMF_BL, performance fatigability (delta data), TTF, total work performed [training time (pre‐training × post‐training) × group (SIT_G × CON)], on EMG‐RMS signal during exercise [exercise time window (0–25 × 25%–50% × 50%–75% × 75%–100%) × training time (pre‐training × post‐training) × group (SIT_G × CON)], and on raw fatigability data [training time (pre‐training × post‐training) × exercise time (baseline × post‐exercise) × group (SIT_G × CON)]. Mauchly's sphericity test was applied, and sphericity was considered violated if the F‐test was significant. In cases where sphericity was violated, the Greenhouse–Geisser epsilon correction was applied. Post hoc tests were performed using the “Bonferroni” method. To improve clarity in the text, when multiple independent variables were analyzed and yielded similar statistical outcomes (e.g., all nonsignificant), the most representative values (i.e., lowest p‐value and corresponding F and η _p ²) were reported in the main text.

Additionally, when necessary, an independent samples t‐test was used to compare deltas (Δ%; percentage of change between pre and posttraining; or percentage of change between baseline and post‐exercise). The significance level for all analysis was set at 5% (p < 0.05). All statistical analysis were conducted using the SPSS software package, version 23 (IBM Corp., USA).

3. Results

3.1. Ramp Incremental Test

The cardiorespiratory values and performance parameters measured during the ramp incremental test are shown in Table 1. Only SIT_G showed a significant training time × group interaction (p ≤ 0.02; F_(1,19) ≥ 6.10; and η _p2 ≥ 0.24), indicating an increase in PPO (post hoc p < 0.01), absolute $\dot{\mathrm{V}}$O_2max (post hoc p < 0.01), RCP‐PO (post hoc p < 0.01), and absolute $\dot{\mathrm{V}}$O₂ at RCP (post hoc p < 0.01). In addition, both the Δ% between pre and posttraining showed a significative increase for PPO, absolute $\dot{\mathrm{V}}$O_2max, RCP‐PO, and absolute $\dot{\mathrm{V}}$O₂ at RCP parameters for SIT_G when compared with CON (t‐test p < 0.01).

TABLE 1.

Variables of the ramp incremental test measured at pre‐ and post‐training in SIT group and control group.

	SIT group (SITG)			Control group (CON)			ANOVA
	Pre‐training	Post‐training	Δ%	Pre‐training	Post‐training	Δ%	Interaction p (F) η p 2	Time p (F) η p 2	Group p (F) η p 2
PPO (watts)	267 ± 37 (242–292)	290 ± 34 (267–314) a , b	+9.4 c	267 ± 43 (237–298)	269 ± 46 (237–302)	+0.6	< 0.01 (15.30) η 2 = 0.45	< 0.01 (21.43) η 2 = 0.53	0.56 (0.36) η 2 = 0.02
$\dot{\mathrm{V}}$O2max (L·min−1)	2.8 ± 0.4 (2.6–3.1)	3.0 ± 0.4 (2.8–3.3) a , b	+7.1 c	2.8 ± 0.4 (2.4–3.1)	2. 8 ± 0.4 (2.5–3.1)	+1.0	0.01 (7.25) η 2 = 0.28	< 0.01 (11.46) η 2 = 0.38	0.38 (0.82) η 2 = 0.04
$\dot{\mathrm{V}}$O2max (mL·kg−1 min−1)	36.1 ± 4.3 (33.3–39.0)	37.5 ± 3.5 (35.2–39.9)	+4.3	37.8 ± 4.8 (34.3–41.2)	37.8 ± 5.5 (33.9–41.8)	+0.05	0.10 (3.05) η 2 = 0.14	0.06 (3.84) η 2 = 0.14	0.63 (0.24) η 2 = 0.01
RCP‐PO (watts)	210 ± 29 (190–229)	232 ± 26 (214–250) a , b	+11.0 c	210 ± 27 (190–230)	209 ± 32 (186–231)	−0.7	< 0.01 (62.22) η 2 = 0.77	< 0.01 (51.94) η 2 = 0.73	0.37 (0.85) η 2 = 0.04
$\dot{\mathrm{V}}$O2 at RCP (L·min−1)	2.35 ± 0.35 (2.12–2.59)	2.55 ± 0.35 (2.32–2.78) a	+8.8 c	2.33 ± 0.35 (2.08–2.58)	2.31 ± 0.39 (2.02–2.59)	−1.0	0.02 (6.10) η 2 = 0.24	0.07 (3.60) η 2 = 0.16	0.39 (0.79) η 2 = 0.04
$\dot{\mathrm{V}}$O2 at RCP (mL·kg−1 min−1)	30.0 ± 3.8 (27.4–32.6)	31.6 ± 3.8 (29.1–34.2)	+5.9 c	31.9 ± 4.0 (29.0–34.8)	31.4 ± 5.7 (27.4–35.4)	−2.0	0.06 (4.04) η 2 = 0.18	0.30 (1.15) η 2 = 0.06	0.66 (0.21) η 2 = 0.01

Note: Values are expressed as the mean ± SD (CI95%).

Abbreviations: Δ%, percentage of change between pre‐ and post‐training for the same group; PPO, peak power output; RCP, respiratory compensation point and $\dot{\mathrm{V}}$O_2max, maximum oxygen consumption.

^a Significant difference for procedure time × group interaction (pre‐training < post‐training).

^b Main time effect (p < 0.05; pre‐training < post‐training).

^c Significant independent t‐test difference (SIT_G > CON). The degrees of freedom for all F were (1,19).

3.2. Time to Task Failure and Total Work Performed at CWR

No significant training time × group interaction (p = 0.52; F_(1,19) = 1.51; and η _p2 = 0.02), training time (p = 0.84; F_(1,19) = 0.04; and η _p2 < 0.01), or group main effects (p = 0.65; F_(1,19) = 0.21; and η _p2 = 0.01) were observed for TTF during the CWR exercises (Figure 2). Similarly, no differences were observed in the Δ% between pre‐training and post‐training for the CWR TTF (t test p = 0.56).

FIGURE 2.

Pre‐ and post‐training comparison of time to task failure and total work performed during constant work rate cycling at 150% of the respiratory compensation point. *Significant pre‐to‐post‐training difference within group (post‐training > pre‐training; p < 0.05). **Significantly different between groups (SIT_G > CON and p < 0.05).

However, the total work performed during the CWR presented a significant training time × group interaction (p = 0.049; F_(1,19) = 4.37; and η _p2 = 0.19), with greater post than pre‐training values observed only in the SIT_G group (post hoc p < 0.01). In addition, a significant main effect of training time was found for total work, indicating that, overall, participants increased total work following training (p = 0.01; F_(1,19) = 8.48; η _p2 = 0.31; and Figure 2). No group main effect was verified for total work performed (p = 0.53; F_(1,19) = 0.41; and η _p2 = 0.02) In line with this result, the comparison of Δ% between pre and post‐training total work showed a significant difference between groups, with SIT_G presenting greater values than CON (t‐test p = 0.04 and Figure 2).

3.3. EMG‐RMS Signal Responses During the CWR

For all muscles, warm‐up EMG‐RMS showed no differences between pre and posttraining and moderate to high intraclass correlation coefficients (ICCs) were observed for these variables (Supporting Information S1). The EMG‐%RMS values during CWR exercise are presented in Figure 3. None of the analyzed muscles exhibited a significant exercise window × training time × group interaction (p ≥ 0.78; F_(4,1,19) ≤ 0.35; and η _p2 ≤ 0.01), training time effect (p ≥ 0.21; F_(4,1,19) ≤ 1.61; and η _p2 ≤ 0.12), or a group effect (p ≥ 0.14; F_(4,1,19) ≤ 5.45; and η _p2 ≤ 0.01). However, a main effect for exercise window was observed for %RMS in all analyzed muscles (p < 0.04; F_(4,1,19) ≥ 0.44; and η _p2 ≥ 0.04), indicating a significant continuous increase in EMG‐%RMS from the beginning to the end of the exercise (i.e., 0%–25% different from the others time windows) for rectus femoris, vastus lateralis, and biceps femoris (post hoc p < 0.01). For gluteus maximus, RMS increased shortly after the onset of exercise and then remained elevated throughout (i.e., 50%–75% is higher than 0%–25%; post hoc p < 0.01; and 75%–100% higher from the others time windows; p < 0.01).

FIGURE 3.

Comparison between EMG‐RMS relativized by warm‐up RMS (EMG‐%RMS) measured during exercise (i.e., fragmented every 25% of the time to exhaustion) at pre‐ and post‐intervention (training and control) for rectus femoris (Panel A), vastus lateralis (Panel B), biceps femoris (Panel C), and gluteus maximum (Panel D). ^†Main time window effect (≠ of all percentage moments; p < 0.05). ^$Main time window effect (≠ of moment 0; p < 0.05).

No significant training time × group interaction (p ≥ 0.24; F_(1,19) ≤ 2.96; and η _p2 ≤ 0.18), training time effect (p ≥ 0.20; F_(1,19) ≤ 1.79; and η _p2 ≤ 0.09), or group effect (p ≥ 0.48; F_(1,19) ≤ 0.57; and η _p2 ≤ 0.03) were observed for the exercise mean values of %RMS.

3.4. Neuromuscular Function—Maximal Voluntary Contractions and Peripheral Neuromuscular Stimulus at Baseline (NMF_BL)

Parameters of NMF_BL in pre and post‐intervention are reported in Table 2. There was no significant difference evidenced by training time effect × group interaction for neuromuscular function variables (p ≥ 0.17; F_(1,19)≤ 2.00; and η _p2 ≤ 0.09). Only Q _tw showed a significant training time effect (p = 0.01; F_(1,19) = 10.9; and η _p2 = 0.36), where the post‐training values were greater compared to pre‐training (post hoc p < 0.01). In addition, no significant group effects were observed for any of measured variables (p ≥ 0.13; F_(1,19)≤ 2.54; and η _p2 ≤ 0.12). Additionally, no significant differences were observed in the Δ% or the logarithmic ratio between pre and posttraining for any NMF_BL data (t‐test p ≥ 0.19).

TABLE 2.

Baseline neuromuscular function (NMF_BL) at pre‐ and post‐training in the SIT group and control group.

	SIT group (SITG)			Control group (CON)			ANOVA
	Pre‐training	Post‐training	Δ%	Pre‐training	Post‐training	Δ%	Interaction p (F) η p2	Time p (F) η p2	Group p (F) η p2
iMVCForce (N)	611 ± 120 (561–692)	614 ± 132 (526–702)	+1.6	600 ± 86 (538–662)	568 ± 87 (506–630)	−5.0	0.34 (0.95) η 2 = 0.05	0.42 (0.68) η 2 = 0.03	0.52 (0.42) η 2 = 0.02
Q tw (N)	169 ± 63 (127–212)	156 ± 64 (113–198)	−8.3	170 ± 40 (141–198)	150 ± 33 (126–173)	−10.6	0.54 (0.40) η 2 = 0.02	0.01 (10.9) η 2 = 0.36	0.90 (0.02) η 2 = 0.01
Db10:100 (a.u.)	0.86 ± 0.11 (0.79–0.94)	0.84 ± 0.16 (0.73–0.95)	−2.7	0.84 ± 0.17 (0.71–0.96)	0.82 ± 0.13 (0.72–0.92)	+1.2	0.91 (0.01) η 2 = 0.01	0.50 (0.48) η 2 = 0.03	0.71 (0.15) η 2 = 0.01
VA (%)	91 ± 5 (87–94)	92 ± 4 (89–95)	+1.4	92 ± 6 (87–96)	92 ± 3 (89–94)	+0.1	0.58 (0.31) η 2 = 0.02	0.72 (0.13) η 2 = 0.01	0.80 (0.07) η 2 = 0.01
RMSMVC (μV)	272 ± 117 (193–350)	332 ± 163 (223–442)	+20.0	207 ± 106 (131–283)	219 ± 141 (118–319)	+7.7	0.17 (2.00) η 2 = 0.09	0.06 (4.24) η 2 = 0.18	0.13 (2.54) η 2 = 0.12
M‐wave max. (mV)	14.1 ± 3.0 (12.0–16.1)	15.7 ± 4.0 (12.3–19.0)	+12.9	14.5 ± 3.2 (12.2–16.7)	14.1 ± 5.3 (10.3–17.9)	−0.8	0.38 (0.82) η 2 = 0.04	0.57 (0.33) η 2 = 0.02	0.70 (0.15) η 2 = 0.01
RMSMVC/M‐wave (a.u.)	0.020 ± 0.012 (0.012–0.028)	0.022 ± 0.007 (0.017–0.026)	+14.8	0.015 ± 0.011 (0.009–0.021)	0.016 ± 0.010 (0.009–0.022)	+18.8	0.80 (0.06) η 2 = 0.01	0.40 (0.75) η 2 = 0.04	0.18 (1.99) η 2 = 0.09

Note: Values are expressed as the mean ± SD. The degrees of freedom for all F were (1, 19).

Abbreviations: Δ%, percentage of change between pre‐ and post‐training for the same group; Db_10:100, ratio between force level at 10 and 100 Hz doublets stimulus; iMVC_Force, peak force peak measured during isometric maximal voluntary contraction; M‐Wave max, maximal M‐Wave amplitude; Q _tw, electrical stimuli; RMS_MVC, root mean square from 500‐ms iMVC peak force; RMS_MVC/M‐Wave, RMS_MVC normalized by maximal M‐Wave amplitude; and VA, voluntary activation.

3.5. Performance Fatigability

Raw data related to performance fatigability are presented in Table 3. No significant differences were observed for training time × exercise time × group interaction for any of analyzed variables (p ≥ 0.16; F_(1,1,19) ≤ 2.09; and η _p2 ≤ 0.10) as well for r group effect (p ≥ 0.14; F_(1,1,19) ≤ 2.31; and η _p2 ≤ 0.11). A significant exercise time effect was observed for iMVC_Force, Q _tw, Db_10:100, VA, RMS_iMVC, and RMS_iMVC/M‐Wave (p ≤ 0.02; F_(1,1,19) ≥ 6.33; and η _p2 ≥ 0.39), where postexercise values were lower than baseline values (post hoc p < 0.02). In addition a training time effect was observed only for Q _tw (p < 0.01; F_(1,1,19) = 11.78; and η _p2 = 0.38), where pretraining showed greater values than post‐training (post hoc p = 0.03).

TABLE 3.

Raw data of neuromuscular system function indicators pre and post constant workload severe domain cycling.

		Pre‐training		Post‐training		ANOVA
		Baseline	Post‐exercise	Baseline	Post‐exercise	Interaction p (F) η p2	Exercise time p (F) η p2	Training time p (F) η p2	Group p (F) η p2
iMVCForce (N)	SITG	611 ± 120 (561–692)	436 ± 153 (334–539) a	614 ± 132 (526–702)	461 ± 143 (345–557) a	0.16 (2.09)	< 0.01 (149.43)	0.60 (0.28)	0.98 (0.01)
	CON	600 ± 86 (538–662)	491 ± 74 (438–544) a	568 ± 87 (506–630)	459 ± 79 (402–516) a	η 2 = 0.10	η 2 = 0.89	η 2 = 0.01	η 2 = 0.00
Q tw (N)	SITG	169 ± 63 (127–212)	89 ± 44 (59–118) a	156 ± 64 (113–198)	84 ± 45 (54–114) a	0.18 (1.95)	< 0.01 (161.73)	< 0.01 (11.78)	0.80 (0.07)
	CON	170 ± 40 (141–198)	109 ± 32 (86–132) a	150 ± 33 (126–173)	89 ± 29 (68–110) a	η 2 = 0.09	η 2 = 0.89	η 2 = 0.38	η 2 = 0.01
Db10:100 (a.u.)	SITG	0.86 ± 0.11 (0.79–0.94)	0.56 ± 0.18 (0.44–0.68) a	0.84 ± 0.16 (0.73–0.95)	0.57 ± 0.17 (0.45–0.68) a	0.43 (0.65)	< 0.01 (116.77)	0.65 (0.22)	0.83 (0.05)
	CON	0.84 ± 0.17 (0.71–0.96)	0.62 ± 0.20 (0.48–0.76) a	0.82 ± 0.13 (0.72–0.92)	0.60 ± 0.16 (0.49–0.71) a	η 2 = 0.03	η 2 = 0.86	η 2 = 0.01	η 2 = 0.00
VA (%)	SITG	91 ± 5 (87–94)	84 ± 11 (77–91) a	92 ± 4 (89–95)	85 ± 10 (79–92) a	0.94 (0.01)	0.02 (6.33)	0.65 (0.21)	0.45 (0.61)
	CON	92 ± 6 (87–96)	88 ± 12 (80–96) a	92 ± 3 (89–94)	88 ± 10 (81–95) a	η 2 = 0.00	η 2 = 0.25	η 2 = 0.01	η 2 = 0.03
RMSiMVC (μV)	SITG	272 ± 117 (193–350)	213 ± 114 (137–290) a	332 ± 163 (223–442)	263 ± 205 (126–401) a	0.99 (0.00)	< 0.01 (21.07)	0.10 (3.04)	0.14 (2.31)
	CON	207 ± 106 (131–283)	160 ± 113 (79–241) a	219 ± 141 (118–319)	161 ± 97 (91–230) a	η 2 = 0.00	η 2 = 0.53	η 2 = 0.14	η 2 = 0.11
M‐wave max. (mV)	SITG	14.1 ± 3.0 (12.0–16.1)	14.0 ± 2.8 (12.1–15.9)	15.7 ± 4.0 (12.3–19.0)	14.3 ± 3.6 (11.8–16.7)	0.19 (1.81)	0.56 (0.34)	0.75 (0.10)	0.98 (0.00)
	CON	14.5 ± 3.2 (12.2–16.7)	14.9 ± 4.0 (12.0–17.7)	14.1 ± 5.3 (10.3–17.9)	14.5 ± 5.3 (10.7–18.3)	η 2 = 0.09	η 2 = 0.02	η 2 = 0.01	η 2 = 0.00
RMSiMVC/M‐wave (a.u.)	SITG	0.020 ± 0.012 (0.012–0.028)	0.016 ± 0.009 (0.009–0.022) a	0.022 ± 0.007 (0.017–0.026)	0.018 ± 0.013 (0.009–0.027) a	0.95 (0.01)	< 0.01 (12.33)	0.25 (1.44)	0.15 (2.21)
	CON	0.015 ± 0.011 (0.009–0.021)	0.011 ± 0.006 (0.006–0.016) a	0.016 ± 0.010 (0.009–0.022)	0.012 ± 0.007 (0.007–0.017) a	η 2 = 0.00	η 2 = 0.39	η 2 = 0.07	η 2 = 0.10

Note: Values are expressed as the mean ± SD. The degrees of freedom for all F were (1, 1, 19).

Abbreviations: Db_10:100, ratio between force level at 10 and 100 Hz doublets stimulus; iMVC_Force, peak force peak measured during isometric maximal voluntary contraction; M‐Wave max, maximal M‐Wave amplitude; Q _tw, electrical stimuli; RMS_MVC, root mean square from 500‐ms iMVC peak force; RMS_iMVC/M‐Wave, RMS_MVC normalized by maximal M‐Wave amplitude; and VA, voluntary activation.

^a Significative exercise time effect (pre exercise > postexercise).

The Δ% data (Figure 4) showed no significant training time effect × group interaction for any variable (p ≥ 0.13; F_(1,19) ≤ 2.48; and η _p2 ≤ 0.12). Neither training time effect (p ≥ 0.47; F_(1,19) ≤ 0.55; and η _p2 ≤ 0.03) nor group effect (p ≥ 0.06; F_(1,19) ≤ 4.30; and η _p2 ≤ 0.18) was observed for any of the variables.

FIGURE 4.

Pre and posttraining percentage comparison of performance fatigability force parameters and electromyograph parameters (Panel A). Pre‐ and post‐training comparison of performance fatigability force parameters and electromyograph parameters normalized by total work (Panel B). Significantly pre‐ and post‐training difference within group (post‐training < pre‐training; p < 0.05).

4. Discussion

This study aimed to assess the impact of a 5‐week SIT intervention on neuromuscular function, muscular activation, and performance fatigability during a bout of CWR exercise performed within severe‐intensity domain. Novel findings from this study indicate that the SIT intervention: (i) did not alter the muscle activation patterns or average EMG‐RMS values during exercise, despite the increase in the PO associated with 150% of RCP performed post‐training, suggesting improved neuromuscular activation economy; (ii) did not change performance fatigability; (iii) did not result in significant alterations in NMF_BL; and (iv) increased the capacity to generate external work for a similar tolerance time during CWR exercise at a physiologically matched intensity (i.e., 150% of RCP‐PO).

In addition to these main outcomes, the SIT intervention also improved physiological parameters (e.g., PPO, $\dot{\mathrm{V}}$O_2max, and RCP), in line with previous reports on SIT induced cardiorespiratory and metabolic adaptations (Burgomaster et al. 2008; Gibala et al. 2006).

4.1. EMG‐RMS During CWR Exercise at 150% of RCP‐PO

Although SIT has been proposed to induce improvements in neuromuscular parameters due to its high demand for skeletal muscle strain (Fernandez‐del‐Olmo et al. 2013; Fiorenza et al. 2019; Riesco‐Villar et al. 2025), only one study has investigated the impact of SIT on EMG‐RMS parameters during cycling exercises, showing a significant increase in the activation of the quadriceps femoris muscles during exercise performed within the extreme‐intensity domain (i.e., all‐out exercise) (Creer et al. 2004). However, the present study is the first to investigate the impact of a SIT program (5 weeks) on neuromuscular adaptations during CWR exercise within severe‐intensity domain, providing novel information that SIT can enhance neuromuscular activation economy.

Whereas the analyzed muscles play different roles during high‐intensity continuous cycling exercises (i.e., vastus lateralis and the gluteus maximus act as prime movers during the downstroke phase, contributing predominantly to knee and hip extension, respectively; the rectus femoris acts as a biarticular motor, assisting both knee extension and hip flexion across the pedal cycle; and the biceps femoris acts as a biarticular antagonist to knee extension while contributing to hip extension) (Hautier et al. 2000), previous studies have shown that, despite these distinct roles, the involvement of these muscle groups is essential for sustaining neuromuscular performance during CWR exercise within the severe‐intensity domain until task failure (e.g., 100%–110% of PPO) (Camata et al. 2011; Diefenthaeler et al. 2012; Dingwell et al. 2008; Dutra et al. 2022). Thus, the maintenance of similar muscle activation levels despite a higher absolute workload (+11% for SIT_G after 5 weeks), without a significant change in TTF, may indicate an improved neuromuscular efficiency, possibly reflecting a reduced neural drive requirement to sustain a given power output.

This interpretation can be supported based on previous evidence that showed that muscle activity during CWR cycling exercise is influenced by the workload, where greater power outputs may reflect higher muscle activity to sustain the task (Petrofsky 1979). Similarly, Bijker et al. (2002) reported significant correlations between muscle activity EMG and absolute power output in cycling, with strong associations for both the vastus lateralis (r = 0.92) and the biceps femoris (r = 0.64), strengthening the notion that greater power outputs are generally accompanied by greater EMG‐RMS amplitudes. Nevertheless, despite this evidences, Azevedo et al. (2022), reported that small variations in power output, achieved by modifying both absolute workload and exercise intensity domains (i.e., from −15% to +15% of maximal lactate steady state; which is comparable to the ∼11% increase on the absolute workload showed on the present study), did not result in significant differences in vastus lateralis activation. It is worth noting that EMG was not assessed in other muscle groups in that study. Thus, although the present results support the interpretation of improved neuromuscular efficiency following SIT, this conclusion should be interpreted with caution. To confirm this hypothesis, future studies should examine whether SIT reduces neuromuscular activation during exercise performed at the same absolute power output as before training, thereby providing direct evidence of improved neuromuscular economy.

4.2. Neuromuscular Performance Fatigability During CWR Exercise at 150% of RCP‐PO

Our findings revealed no differences in performance fatigability presented as raw data or as Δ% (Table 3 and Figure 4A) between groups after training. Both groups exhibited significant reductions in Q_tw, indicating a high level of peripheral fatigue at task failure, both before and after the 5‐week training or control period. This response is consistent with the typical characteristics of the severe‐intensity domain, where neuromuscular fatigue develops rapidly and is primarily of peripheral origin (Azevedo et al. 2022; de Poli et al. 2020; Iannetta et al. 2022), generally attributed for a higher accumulation of inorganic phosphate and hydrogen ions (Hureau et al. 2022; Laginestra et al. 2025).

The absence of post‐training differences in fatigability, despite the increased total work performed by the SIT group, may be explained by the nature of the exercise protocol. The CWR bouts were prescribed at 150% of each participant's current RCP‐PO, meaning that improvements in RCP‐PO after training resulted in proportionally greater absolute workloads (Table 1). This design aimed to preserve the relative physiological strain during posttraining assessments (de Poli et al. 2024), thereby eliciting comparable internal loads despite the greater power output.

No previous studies have directly assessed performance fatigability specifically during a CWR exercise within the severe‐intensity domain following SIT. Although primarily focused on metabolic responses, we recently demonstrated that SIT did not alter muscle glycogen utilization or TTF failure during severe‐intensity CWR cycling at 150% RCP‐PO, despite significant metabolic adaptations (e.g., increased muscle glycogen content and enhanced glycolytic enzyme content and activity) (de Poli et al. 2024). Similar findings have been reported following high‐intensity interval training, where neuromuscular fatigue responses remained unchanged despite increases in speed and performance markers during repeated sprint protocols (Milioni et al. 2019, 2024). These findings suggest that, when exercise is performed at a physiologically matched intensity, the internal demand remains similar even when absolute power output and total work are increased.

Additional evidence from literature reinforces this interpretation. Fatigability appears to be primarily determined by the exercise intensity domain. For example, lower fatigability levels are consistently observed in the moderate domain, despite longer exercise durations and potentially greater total work, whereas severe‐intensity efforts lead to greater neuromuscular impairment in considerably shorter time frames (Iannetta et al. 2022). Although these studies did not explicitly assess the relationship between total work and fatigability, they indicate that lower‐intensity domains allow for more accumulated work while inducing less neuromuscular fatigue. Consistent with this, even small manipulations in power output around MMSS (−10% of MMSS vs. MMSS vs. +10% of MMSS) have elicited distinct fatigue responses only when changing the exercise intensity domain. For instance, exercising at +10% MMSS, which falls within the severe‐intensity domain, results in greater peripheral fatigue compared to MMSS or −10% MMSS, both of which are within the heavy‐intensity domain (upper boundary). This indicates that neuromuscular function is highly sensitive to relative intensity (Azevedo et al. 2021). This domain‐dependent response is further supported by findings showing that, whereas fatigability was not different when exercising at 70% and 90% of the gas exchange threshold (i.e., moderate intensity domain), it was significantly greater when exercising at 110% of this threshold under total work matched conditions (Brownstein et al. 2022). Additionally, fatigability has been shown to remain unchanged when comparing continuous and intermittent severe‐intensity exercise with equal total work, with no significant association with total work (MacDougall et al. 2024).

Despite these insights, our findings are limited to assessments performed at a physiologically matched intensity. Future studies should evaluate performance fatigability at the same absolute pretraining workload to determine whether short‐term SIT induces meaningful improvements in fatigue resistance. Such approaches are necessary to further elucidate the specific effects of training on neuromuscular function, independently of exercise intensity adjustments.

4.3. Baseline Neuromuscular Function Measured Pre‐ and Post‐Training

Our results also reported no significant differences between NMF_BL outcomes assessed pre and posttraining. Interestingly, other studies have similarly reported no changes in isometric knee‐extension following a 2‐week SIT regimen (four to six efforts per training session) in recreationally active participants (Astorino et al. 2012; Bertschinger et al. 2020). This could lead to an idea that SIT might not cause neuromuscular alterations if observed in isolation (without others neuromuscular parameters, e.g., EMG‐RMS during exercise and performance fatigability). However, considering that in present study the SIT group decreased the mean values of biceps femoris and gluteus maximus activation during CWR exercise performed at 150% of RCP‐PO when analyzed as a function of the pretraining matched work, the nonsignificant change on NMF_BL outcomes may be explained by the specificity of neural adaptation to exertion (Ansdell et al. 2020). For instance, whereas cycling dynamic exercise may have an influence of full use of synergist muscles to sustain exercise performance, iMVC “isolates” some muscular groups to exert force of a single joint (e.g., knee extension) (Gandevia 2001). Therefore, an improvement in neuromuscular function during a cycling dynamic exercise may not necessarily translate into improved performance in isometric efforts such as iMVC (i.e., this adaptations are “task specific”) (Gandevia 2001).

4.4. Improvement of Total Work Without Changes in Exercise Time Tolerance

The present data showed an increase in the total work performed during CWR exercise within severe‐intensity domain. As mentioned in previous sections, the heightened metabolic stress following SIT [e.g., acid‐base imbalance, probably due to a high accumulation of hydrogen ions (Fiorenza et al. 2019) and high accumulation of inorganic phosphate (Glaister 2005)] associated with high peripheral neuromuscular stress (Fernandez‐del‐Olmo et al. 2013; Fiorenza et al. 2019; Hureau et al. 2016), could promote neuro‐metabolic adaptations [e.g., improved buffering capacity and ion regulation (de Poli et al. 2024)] that may reduce peripheral fatigue for a given workload, thereby extending performance until the sensory tolerance limit” (Hureau et al. 2018). This would allow for a greater amount of total work to be performed during CWR exercise within the severe‐intensity domain.

However, it is important to consider that in the current study, despite the increased total work, no significant differences were observed in TTF. This finding is consistent with another study by our research group (de Poli et al. 2024), which demonstrated that despite the observed enhancements in biomolecular glycolytic metabolic markers after SIT (e.g., glycogen content, buffer capacity; phosphofructokinase content and activity), and the increase in total work during severe‐intensity domain CWR following 5 weeks of SIT, no differences were observed in TTF. Therefore, considering the neuromuscular adaptations for the same total work performed observed in the current study, along with the previously identified glycolytic pathway adaptations after 5‐week of SIT by our study group (de Poli et al. 2024), it can be suggested that TTF per se may not be a suitable parameter for assessing improvements in performance or tolerance during exercise performed at the same relative intensity within the severe‐intensity domain (e.g., 150% of RCP‐PO). Nevertheless, although TTF remained unchanged, it is important to highlight that the absolute intensity of the exercise was greater after SIT, while maintaining the same relative intensity (i.e., same percentage of RCP‐PO). When considered together with the increased total work, this indicates a meaningful improvement in performance capacity as participants were able to sustain a higher external load for the same duration within the severe‐intensity domain.

4.5. Limitations

Despite the relevance of the current findings, it is important to acknowledge that post‐training assessments were conducted at the same relative intensity (i.e., 150% of RCP‐PO), resulting in a higher absolute workload after SIT (e.g., higher RCP‐PO). As fatigability is influenced more by the exercise intensity domain than by absolute power output per se (Azevedo et al. 2021), the use of a physiologically matched intensity may have masked potential improvements in fatigability pattern. Performing the same exercise protocol at the pre‐training absolute workload for the same test duration (i.e., isotime trial) could provide further insights into neuromuscular adaptations to SIT, independently of the exercise intensity domain. Therefore, additional studies are warranted to fully elucidate the effects of SIT on performance fatigability and neuromuscular function across different intensity domains.

5. Conclusion

In addition to the expected benefits of SIT in improving PPO, $\dot{\mathrm{V}}$O_2max, and $\dot{\mathrm{V}}$O₂ at RCP and RCP‐PO, the present study provides novel evidence that SIT may promote more efficient muscle recruitment during severe‐intensity CWR exercise as suggested by maintained activation levels despite a higher absolute workload after training period. However, development of neuromuscular fatigue during 150% RCP‐PO CWR exercise within severe‐intensity domain did not change after SIT nor did the TTF at the same percentage intensity determined by RCP‐PO. Taken together, these findings indicate that SIT may enhance neuromuscular activation economy without altering the end‐exercise fatigue response within the severe‐intensity domain.

Author Contributions

R.A.B.d.P. and A.M.Z. designed the study. R.A.B.d.P. and Y.M.D. participated in data acquisition. R.A.B.d.P., Y.M.D., J.M.M., and A.M.Z. analyzed and interpreted the data. R.A.B.d.P., A.M.Z., J.M.M., and Y.M.D. draft the manuscript. All authors critically reviewed the manuscript and approved the final manuscript as submitted. All authors agree to be accountable for all aspects of the work to ensure that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. All persons designated as authors qualify for authorship, and all those who qualify for authorship are listed.

Funding

De Poli RAB, Dutra YM, and Zagatto AM received Sao Paulo Research Foundation (FAPESP) grants (Process numbers # 19/17445‐1; # 21/08479‐0; and #18/21126‐6, respectively). In addition, this work was supported by the National Council for Scientific and Technological Development (CNPq)—Brazil [Grant No. 422193/2016‐0].

Ethics Statement

The São Paulo State University Ethics Research Committee approved all procedures involving human participants (Protocol #2.650.216/2018).

Consent

All participants were carefully informed about the possible risk and benefits of the project and they provided written informed consent to participate in the study.

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Supporting Information S1

Data Availability Statement

The data that support the findings of this study are available from the corresponding author upon reasonable request.