Quadriceps Muscle Length Differentially Affects Exercise Capacity, the Torque‐Duration Relationship, and Neuromuscular Fatigue

ABSTRACT

To determine the effects of muscle length on exercise capacity, critical torque/W′ and neuromuscular fatigue. Thirteen subjects performed 60 isometric maximal voluntary contractions (MVC) over a period of 5 min, at five knee angles to manipulate maximal torque (i.e., optimal, −15°, −30°, +15° and +30°). Exercise capacity was quantified by total impulse. Critical torque was calculated as the mean torque of the last 30s of exercise, and W′ as the impulse done above critical torque. Cardiometabolic and ventilatory responses were measured. MVC, potentiated and interpolated twitches were performed to quantify neuromuscular fatigue, muscle contractility, and voluntary activation failure. Compared to optimal muscle length (i.e., the length producing the greatest torque‐output), total impulse was reduced in all conditions (p < 0.001) but −15°. W′ was reduced in every condition whereas critical torque was reduced only in +15° and +30° (p < 0.001). Neuromuscular fatigue, contractile impairments, and oxygen cost of contraction were reduced in −15° and −30°. In contrast, no difference in the degree of neuromuscular and contractile fatigue was found at +15° and +30° exercise end. However, the rate at which neuromuscular and contractile fatigue accumulated, as well as the oxygen cost of contraction were increased. Compared to optimal length, muscle lengthening and shortening produced symmetrical effects on W′, suggesting that maximal torque plays an important role in determining W′. In contrast, the asymmetrical effects of changing muscle length on critical torque, neuromuscular fatigue, and its mechanisms suggest that the role of maximal torque in determining these variables is limited.

1. Introduction

Unraveling the physiological mechanisms that determine exercise performance is essential for counteracting the deleterious effects of sedentary behaviors and optimizing exercise prescription and performance. Compelling evidence has shown that exercise capacity follows a hyperbolic relationship when plotted against exercise intensity [1, 2, 3, 4]. This relationship is mathematically defined by two parameters: (1) critical torque, which corresponds to the asymptote of the hyperbolic relationship and (2) the curvature constant (W′), which represents a fixed amount of impulse that can be performed above critical torque [5].

Critical torque has been described as a phase transition between the heavy and severe intensity domains [6]. During exercise at or below critical torque, intramuscular metabolite concentrations [7], muscle activation [8, 9], oxygen uptake and blood lactate reach a steady state [1]. In contrast, they do not when exercise is conducted above critical torque. Under these conditions, the inability of the oxidative metabolism to meet the ATP demand results in progressive depletion of phosphocreatine and accumulation of inorganic phosphate and hydrogen ions [7]. These intramuscular disturbances impair participants' torque generating capacity, a phenomenon defined as neuromuscular fatigue [10] and recognized as a key determinant of exercise capacity [11, 12, 13, 14].

Neuromuscular fatigue has been proposed to depend, at least in part, on an individual's maximal torque generating capacity [15, 16, 17, 18, 19, 20]. For example, Massamba et al. [16] reported that knee extensors, which produce greater maximal torque than knee flexors, also exhibited both greater exercise capacity and greater neuromuscular fatigue. Likewise, Zarzissi et al. [17] found that young adults generated higher maximal torque and developed greater neuromuscular fatigue than older adults. However, whether maximal torque per se determines exercise capacity and fatigue remains unresolved. In these studies, greater maximal torque was associated with higher critical torque and W′, both of which independently influence exercise tolerance and neuromuscular fatigue and represent important confounding factors [20, 21, 22]. In addition, comparisons between different muscle groups or populations inherently involve differences in muscle phenotype [23], metabolic function [12], and other physiological characteristics that may also independently affect neuromuscular fatigue [16, 17, 19, 21, 24]. Isolating the role of maximal torque therefore requires an experimental approach that selectively manipulates torque generating capacity within the same participants and muscle group while minimizing these confounding factors.

In the present study, we manipulated muscle length to dose dependently change maximal torque generating capacity within the same participants and muscle group [25]. This procedure provides a more controlled methodology to determine the effects of maximal torque on neuromuscular fatigue, exercise capacity, and the parameters of the torque‐duration relationship. While exercising at short muscle length is typically associated with improved exercise capacity and a slower rate of neuromuscular fatigue, exercising at long muscle length is associated with reduced exercise capacity and a faster rate of neuromuscular fatigue [8, 26, 27]. Whether these effects are mediated by changes in critical torque, W′ or both remains unclear. To address this question, participants performed 5‐min all‐out exercises at five different muscle lengths to evaluate critical torque and W′ in a single visit [4]. We hypothesized that exercising at shorter than optimal muscle lengths would preserve exercise capacity by attenuating neuromuscular fatigue and increasing critical torque relative to maximal torque. Conversely, we hypothesized that exercising at longer than optimal muscle lengths would reduce exercise capacity by accelerating neuromuscular fatigue and decreasing critical torque relative to maximal torque.

2. Methods

2.1. Participants

Thirteen participants (four women; mean ± SD; age, 23 ± 3 yr.; height, 172 ± 9 cm; weight, 70 ± 9 kg; body fat, 17% ± 4%) took part in the present study. All participants gave their written informed consent and were engaged in regular physical activity. All participants were nonsmokers, nonmedicated, and arrived at the laboratory rested and hydrated. They were asked to avoid strenuous physical activity, caffeine and alcohol consumption 24 h prior to each visit. The study was approved by the University of Strasbourg ethics committee (2021‐A00412‐07) and conducted according to the latest Declaration of Helsinki for human experimentation except for registration in a database.

2.2. Experimental Design

All participants visited the laboratory 6 times to complete a familiarization visit and five experimental visits (detailed in “Experimental protocol” section). Testing took place over a 2‐ to 3‐week period, with at least 48 h between visits to allow complete recovery of the neuromuscular function [28]. To control for potential diurnal fluctuations in maximal torque‐generating capacity and muscle activation, the testing sessions were scheduled at the same time of day for each participant.

2.3. Experimental Protocol

2.3.1. Familiarization (Visit 1)

Anthropometric measurements were collected, participants were familiarized with the evaluation of neuromuscular function and evaluation of the optimal angle that produces maximal torque was determined. At the end of the visit, participants were familiarized with the 5‐min all‐out exercise performed during each experimental visit. This exercise consisted of 60 maximal voluntary contractions (MVCs) over a 5 min period (3 s on—2 s off) at the previously determined optimal angle [4].

2.3.2. Experimental Visits (Visit 2–6)

During the five experimental sessions, participants performed the 5‐min all‐out exercise at five different knee angles to modulate the quadriceps muscle length. The different knee angles, performed in randomized order, were (1) optimal angle (i.e., the angle that resulted in producing the maximal voluntary torque), optimal angle +15° (+15°), optimal angle +30° (+30°), optimal angle −15° (−15°) and optimal angle −30° (−30°). These different knee angles were selected to modulate maximal torque dose‐dependently around the optimal muscle length. Compared to the optimum, knee angles set at −15° and −30° placed the quadriceps muscles in a shortened position, whereas knee angles set at +15° and +30° placed the quadriceps muscles in a lengthened position. During the task, participants were securely fastened by seat belts strapped across the chest and hips to avoid lateral, vertical or frontal movements. Arms were folded across the chest and strong vocal encouragements were given throughout the exercise to maximize performance. Participants had no information on the time elapsed or number of contractions performed or remaining. A soundtrack indicated each contraction and relaxation phase, and a visual feedback of torque output was provided on the computer screen. To ensure that participants produced their best performance during the first contraction of 60 MVCs all participants had to reach at least 95% of their MVC peak torque performed during baseline measurement. If participants failed to reach this threshold, they were informed that the exercise would start over after a period of 10 min of recovery. Neuromuscular function was assessed before, during (every 10 contractions) and at 0 s and 1, 2, 4, 6, 10, and 15 min after the fatiguing exercise protocol.

2.4. Data Collection and Analysis

2.4.1. Determination of the Optimal Knee Angle for Torque Production

The optimal knee angle was determined using an isokinetic dynamometer (Biodex System 4, Biodex Medical Systems Inc., Shirley, NY, USA). The hip angle was set at 85° and the settings of the isokinetic dynamometer were adjusted and recorded to correspond to the subject's morphology. Participants performed in randomized order two MVCs at different knee angles of 65°, 70°, 75°, 80°, 85° and 90° (with 0° representing full knee extension). The optimal knee angle (76.2° ± 4.6°) was determined as the angle that elicited maximal torque production.

2.4.2. Neuromuscular Function Assessment

For the assessment of neuromuscular function, participants were seated on the isokinetic dynamometer, the rotation axis of the lever arm was aligned with the lateral epicondyle of the right tibia, and the knee angle was set at the specific angle of the visit. Importantly, the degree of neuromuscular fatigue has previously been shown to be unaffected by the knee angle at which it is assessed [26]. Accordingly, fatigue measurements acquired at the knee angle of the corresponding exercise can be compared (i.e., from optimal angle −30° to +30°).

The lever arm of the dynamometer was fixed to the subject's right ankle, just superior to the malleoli. The cathode (i.e., negative pole), a self‐adhesive electrode, was placed on the femoral triangle, at the site of stimulation that allowed both the production of the maximal torque output of the quadriceps twitch and the maximal amplitude of the compound muscle action potential (M _MAX) for the vastus lateralis (VL), vastus medialis (VM), and rectus femoris (RF). The anode (i.e., positive pole), a rectangular self‐adhesive electrode, was placed halfway between the greater trochanter and the superior iliac crest [29]. The position of these electrodes was marked with an indelible marker to ensure a reproducible stimulation site from one visit to the next. A constant‐current stimulator (DS7AH, Digitimer, Hertfordshire, UK) delivered a square‐wave stimulus (200 μs) at a maximum voltage of 400 V. To ensure maximum spatial recruitment of motor units during neuromuscular assessments and to avoid the effects of membrane excitability variation during exercise, stimulation intensity (158 ± 61 mA) was set at 120% of the stimulation intensity required to obtain maximal quadriceps twitch and M _MAX values [30].

Neuromuscular assessment of the quadriceps muscle was conducted before, during and after each exercise. Each measurement block consisted of a 3‐s MVC during which a single superimposed stimulation twitch (QT_single, _superimposed) was delivered at the peak torque of the MVC to determine voluntary activation of the quadriceps [31, 32]. Immediately after the MVC, a stimulation train consisting of single (QT_single) and double stimulations [10 Hz (QT₁₀) and 100 Hz (QT₁₀₀)] of the femoral nerve was delivered at 1, 3 and 5 s post‐MVC, respectively. These measurements were repeated six times before (the first two without electrical stimulations) to maximize twitch potentiation [33] and at 0 s and 1, 2, 4, 6, 10 and 15 min postexercise. During exercise, only single stimulations (QT_single) were delivered post MVC to maintain constant recovery time (2 s) between contractions. Peak torque was determined for MVC, QT₁₀ and QT₁₀₀. For QT_single, we determined peak torque, contraction time to peak torque, maximal rate of torque development (MRTD), half relaxation time, and M _MAX . The QT₁₀/QT₁₀₀ ratio (QT_10:100) was calculated, as a decrease in this ratio is interpreted as an index of prolonged low‐frequency fatigue [34, 35]. VA of knee extensors was calculated according to the following formula:

$$\mathrm{VA}\left(\%\right)=\left(1-{\mathrm{QT}}_{\text{single},\text{superimposed}}/{\mathrm{QT}}_{\text{single}}\right)\times 100.$$

Baseline values were calculated by averaging the preexercise values of the standardized sets of contractions. For each index, the percent change from baseline was calculated to quantify and characterize the origin of exercise‐induced neuromuscular fatigue.

To quantify the rate of recovery over time, the time‐course of recovery was divided into the following three periods [20]: (R1) from 0 s to 2 min, (R2) from 2 to 4 min, and (R3) from 4 to 15 min. The recovery rate of each period was calculated as follows:

$$\text{Recovery}\left(\%.{\min }^{-1}\right)=100\times \left[\left({\mathit{Xt}}_1-{\mathit{Xt}}_2\right)\right]/\text{Duration recovery}$$

with Xt ₁ and Xt ₂ corresponding to the exercise‐induced reduction in the index (as a percentage) measured at the beginning (t ₁) and end (t ₂) of the recovery period and duration recovery being 2, 2 and 11 min for the first, second, and third recovery periods, respectively.

2.4.3. Surface Electromyography

Electrical activity of quadriceps was assessed using surface EMG (Octal Bio‐Amp‐ ML138 and PowerLab 16/35, AdInstrument, Bella‐Vista, Australia). Bipolar, self‐adhesive Ag/AgCl electrodes (Diameter, 10 mm; interelectrode distance, 20 mm, Contrôle Graphique Medical, Brie‐Comte‐Robert, France) were placed over the right vastus lateralis (VL), vastus medialis (VM), rectus femoris (RF) and biceps femoris (BF). A reference electrode was placed on the lateral condyle of the right tibia. For each visit, the skin was shaved, abraded with emery paper, and cleaned with alcohol to reduce skin impedance below 3 kΩ. Electrodes were positioned according to SENIAM recommendations [36]. Final positions were adjusted to optimize M _MAX shape and amplitude and were marked with an indelible marker to ensure identical placement at subsequent visits. EMG signals were amplified (gain, 20), filtered (first‐order filter, 10 Hz; fourth‐order Bessel filter, 500 Hz) and recorded at a sampling frequency of 4 kHz using commercially available software (Labchart 8, ADInstruments, Bella‐Vista, Australia; RRID:SCR_023643).

Maximal muscle activation was calculated during each contraction as the highest 500‐ms of EMG activity. Muscle activation was estimated by calculating the root mean square (RMS) of the EMG signals which was normalized by the amplitude of the muscle M wave. For biceps femoris activation, muscle activation was normalized by the RMS recorded during pre‐exercise flexion MVC.

To evaluate knee flexors coactivation during knee extension MVCs, we calculated the coactivation index using the following formula [37]:

$$\mathrm{BF}\text{coactivation index}\left(\%\right)=\frac{\mathrm{BF}\mathrm{RMS}\text{knee extension}}{\mathrm{BF}\mathrm{RMS}\text{knee flexion}}$$

2.4.4. Determination of Critical Torque and Work Done Above Critical Torque

Critical torque was estimated by averaging the last six MVCs of the all‐out exercise [4]. W′ was calculated according to the following formula:

$$\begin{array}{c}{\mathrm{W}}^{\prime }=\text{total torque impulse}-\left[\text{critical torque}\times \text{contraction time}\right].\hfill \end{array}$$

Total torque impulse was calculated by integrating the torque signal over the 60 contractions [4].

2.4.5. Systemic Response to Exercise

Pulmonary ventilation and gas exchange indices were measured breath‐by‐breath at rest and throughout exercises using a stationary automatic ergospirometer connected to a facemask (Vyntus MS‐CPX, Viasys, San Diego, CA). Before each test, gas analyzers were calibrated using a certified gas preparation (O₂, 15.9%; CO₂, 5%) and an accurate volume of ambient air (2 L) was used to adjust the pneumotachograph. Heart rate (HR) was calculated from a heart rate monitor (V800; Polar Electro, Kempele, Finland). Oxygen uptake (V̇O₂), carbon dioxide output (V̇CO₂), respiratory exchange ratio (V̇CO₂.V̇O₂ ⁻¹), minute ventilation (V̇_E), breathing frequency (fB), tidal volume (V_T), and HR measured during exercise were averaged every 30 s of the exercise duration. Blood lactate concentration ([La]_b) was quantified using an electrochemical method (LactateScout; SensLab GmbH, Germany) from capillary blood samples (5 μL) collected from a fingertip at rest and 3 min postexercise. To evaluate the oxygen cost of contraction between conditions, V̇O₂ gain was calculated by normalizing oxygen consumption to the mean torque recorded during its corresponding time period.

2.4.6. Near‐Infrared Spectroscopy

A near‐infrared spectroscopy device (NIRS; Portamon, Artinis Medical Systems, Netherlands) was placed on participants' right vastus lateralis and rectus femoris to assess muscle tissue oxy‐, deoxy‐, and total hemoglobin/myoglobin relative concentrations [heme] and tissue absolute oxygenation (StO₂). The NIRS sensors use three pairs of light emitting diodes that emit continuous light at 760 and 850 nm. Light emitting diodes are placed in a spatially resolved spectroscopy configuration with a source–detector spacing of 30, 35, and 40 mm. The sensors were placed on the vastus lateralis and rectus femoris as close as possible to the EMG electrode locations and proximal to the greater trochanter. Sensor location was marked with indelible ink for reproducible placement between experimental visits and covered with black cloths to avoid signal contamination with ambient light. As recommended by the manufacturer, a differential path length factor of 4 was used, and data were sampled at 10 Hz. All signals were collected in OxySoft (Artinis Medical Systems, Netherlands; RRID:SCR_024666).

To account for change in blood volume, oxy[heme] and deoxy[heme] signals were corrected using the following formula [38]:

$$\beta \left(t\right)=\frac{\mathrm{oxy}\left[\text{heme}\right]}{\mathrm{oxy}\left[\text{heme}\right]+\text{deoxy}\left[\text{heme}\right]}$$

$$\mathrm{oxy}{\left[\text{heme}\right]}_{\mathrm{c}}\left(t\right)=\mathrm{oxy}\left[\text{heme}\right]\left(t\right)-\text{total}\left[\text{heme}\left(t\right)\right]\times \left(1-\beta \left(t\right)\right)$$

$$\text{deoxy}{\left[\text{heme}\right]}_{\mathrm{c}}\left(t\right)=\text{deoxy}\left[\text{heme}\right]\left(t\right)-\text{total}\left[\text{heme}\left(t\right)\right]\times \beta \left(t\right)$$

where oxy[heme]_c and deoxy[heme]_c are the corrected oxygenated and deoxygenated signals, respectively, total[heme] is the blood volume signal from the NIRS device, β is the blood volume correction factor, and t is time. The raw oxy[heme] signal was corrected by subtracting the proportion of the blood volume change attributed to oxy[heme], while the raw deoxy[heme] signal was corrected by subtracting the proportion of blood volume change attributed to deoxy[heme].

At the end of each experimental visit, a blood pressure cuff was placed proximally around the right thigh. The cuff was inflated at suprasystolic pressure (> 300 mmHg) to progressively deoxygenate the tissue under the optodes until a plateau in oxy[heme] and deoxy[heme] was observed (~5 min). The minimal and maximal oxy[heme] or deoxy[heme] values obtained during this occlusion period and the subsequent hyperemic response were considered as 0% and 100% oxy[heme] or deoxy[heme] [39]. Although this procedure neglects the influence of adipose tissue on NIRS data [39], subcutaneous adipose tissue thickness was measured using a skinfold caliper (Harpenden, London, United Kingdom) which observed values was divided by 2. All signals were averaged every 30 s of exercise duration.

2.5. Statistical Analysis

Data presented in the results section are expressed as mean ± SD. Normality and homogeneity of the variance of every dependent variable were verified using the Shapiro–Wilk test and the Bartlett's test, respectively. To determine the effects of muscle length on parameters measured during exercise (e.g., torque production, EMG, and cardiometabolic data) or to assess differences in neuromuscular fatigue indices between PRE to POST‐exercise (e.g., ∆MVC, ∆QT_single, ∆VA), we used a two‐way ANOVA with repeated measures (muscle length × time effects). To determine the effects of muscle length on maximal torque, total impulse, critical torque, W′, and blood lactate, we used a one‐way ANOVA with repeated measures (muscle length effect). A Greenhouse–Geisser correction was used when sphericity was violated. When a significant difference was found, multiple‐comparison analysis between conditions was performed using Tukey's honestly significant difference post hoc test. Effect size was assessed using partial ƞ ² (ƞ _p ²). Effect size was considered as small when ƞ _p ² was lower than 0.07, medium when ƞ _p ² comprised between 0.07 and 0.20, and large when ƞ _p ² was greater than 0.20 [40]. Associations between maximal torque, V̇O₂ gain, muscle activation indices, fatigue indices, W′ and critical torque were tested using linear mixed models with a random intercept. The regression coefficients (β) were computed with 95% confidence intervals (95% CI), and goodness of the fit from the predictor effect was computed using the r ² coefficient. Statistical analyses were conducted using Prism (10.3.0; GraphPad, La Jolla, CA; RRID:SCR_002798). Correlation analyses were conducted using R studio (2025.05.1 + 513; Posit Software, PBC; Boston, USA; RRID:SCR_000432) with lme4, lmerTest, and performance packages. Statistical significance was set at p < 0.05.

3. Results

3.1. Baseline Neuromuscular Function

Optimal muscle length produced the greatest maximal voluntary torque, MRTD, ∆QT10 and ∆QT100 compared to all other conditions which decreased length‐dependently when muscle length increased or decreased (p < 0.0008, η _p ²  > 0.457; Figure 1; Table 1). Compared to optimal muscle length, QTsingle decreased when muscle length decreased (i.e., −15° and −30°, p < 0.0001, η _p ²  = 0.668; post hoc: p < 0.0018) but remained unchanged when muscle length increased (Figure 1; post hoc: p > 0.553). In contrast, voluntary activation as well as VL, VM, RF muscle activations, and BF coactivation were the lowest at −30° (i.e., shortest muscle length) and increased together with muscle length to reach the highest value at +30° (p < 0.0031, η _p ²  > 0.369; Figure 1). Knee flexors torque was not different between optimal (93 ± 19 Nm) and reduced quadriceps muscle length (i.e., lengthened hamstring muscles; −30°: 91 ± 20 Nm; −15°: 95 ± 22 Nm; post hoc: p > 0.986) but decreased when quadriceps muscle length increased (+15°: 83 ± 19 Nm; +30°: 54 ± 20 Nm; post hoc: p < 0.0168; muscle length effect: p < 0.0001 η _p ²  > 0.708).

FIGURE 1.

Effects of muscle length on baseline neuromuscular function. Data are presented as group mean (bars) and individuals (white circles and dotted lines) for neuromuscular indices and were analyzed using one‐way ANOVA (n = 13). OPT, optimal muscle length; −30°, optimal muscle length minus 30° of knee extension; −15°, optimal muscle length minus 15° of knee extension; +15°, optimal muscle length plus 15° of knee extension; +30°, optimal muscle length plus 30° of knee extension. QT_single, QT₁₀ and QT₁₀₀, potentiated twitch peak torque evoked by single, 10 Hz paired and 100 Hz paired electrical stimulation of the femoral nerve, respectively; VL, vastus lateralis; VM, vastus medialis; RF, rectus femoris. *, p < 0.05; **, p < 0.01; ***, p < 0.001; For clarity, only the most relevant significant differences are presented.

TABLE 1.

Effects of different muscle lengths during exercise on neuromuscular function.

Index	Units	−30°			−15°			OPT			+15°			+30°
		Pre	Post 0 s	15 min	Pre	Post 0 s	15 min	Pre	Post 0 s	15 min	Pre	Post 0 s	15 min	Pre	Post 0 s	15 min
MVC	Nm	182 ± 65*	115 ± 32	190 ± 69*	237 ± 75*	116 ± 24	228 ± 70	275 ± 77	113 ± 34	257 ± 91	231 ± 46*	93 ± 29	205 ± 52*	183 ± 41*	77 ± 20	156 ± 46*
VA	%	90 ± 8	79 ± 9	87 ± 10	91 ± 5	73 ± 11	87 ± 6	93 ± 5	75 ± 11	91 ± 6	96 ± 3	82 ± 13*	92 ± 5	98 ± 2	82 ± 11*	95 ± 3
QTsingle	Nm	39 ± 13*	24 ± 9	28 ± 8	48 ± 15*	22 ± 8	32 ± 7	54 ± 15	19 ± 7	33 ± 7	51 ± 12	16 ± 8	28 ± 5	52 ± 13	17 ± 9	27 ± 7
QT10	Nm	72 ± 20*	45 ± 17*	51 ± 14	89 ± 23*	38 ± 14*	55 ± 12	97 ± 22	31 ± 13	52 ± 12	91 ± 16	26 ± 15	44 ± 12*	82 ± 19*	27 ± 15	42 ± 15*
QT100	Nm	78 ± 26*	62 ± 18*	62 ± 15*	92 ± 27	58 ± 14*	71 ± 13	96 ± 22	49 ± 13	70 ± 12	88 ± 18*	41 ± 15*	62 ± 11*	81 ± 19*	41 ± 14*	57 ± 16*
QT10:100	Ø	0.93 ± 0.08	0.72 ± 0.12*	0.83 ± 0.13	0.97 ± 0.07	0.66 ± 0.16	0.77 ± 0.13	1.01 ± 0.06	0.61 ± 0.15	0.74 ± 0.14	1.04 ± 0.07	0.63 ± 0.24	0.71 ± 0.15	1.02 ± 0.06	0.62 ± 0.14	0.73 ± 0.12
CT	ms	97 ± 19	81 ± 13	79 ± 20	101 ± 12	76 ± 11	77 ± 8	94 ± 8	80 ± 14	74 ± 11	103 ± 14	93 ± 19	76 ± 11	104 ± 10	100 ± 14*	79 ± 16
HRT	ms	67 ± 20	82 ± 17	74 ± 15	66 ± 12	89 ± 25	73 ± 24	70 ± 14	93 ± 22	69 ± 13	74 ± 14	91 ± 23	73 ± 15	70 ± 9	82 ± 18	71 ± 13
MRTD	Nm.s−1	967 ± 391*	619 ± 214*	693 ± 188*	1200 ± 434	559 ± 228	853 ± 208	1295 ± 333	469 ± 201	893 ± 147	1140 ± 198	363 ± 172	701 ± 128*	979 ± 201*	363 ± 190*	605 ± 149*
VL Mmax	mV	14.1 ± 3.4	12.5 ± 4.7	12.6 ± 4.5	14.3 ± 3.7	11.9 ± 4	12.1 ± 3.6	14.4 ± 3.9	13.5 ± 4.1	12.9 ± 3.6	14.1 ± 3.9	13 ± 3.6	12.6 ± 3.8	13.5 ± 3.7	12.7 ± 3.8	11.9 ± 3.4
VM Mmax	mV	15.8 ± 4.7	14.3 ± 3.9	14 ± 6.2	16 ± 4.9	14.2 ± 7.2	13.2 ± 6.5	16.0 ± 4.6	16.0 ± 4.2	14.5 ± 4.6	16.0 ± 4.7	16 ± 5.0	14.4 ± 5.1	15.9 ± 4.6	16.0 ± 4.3	14.4 ± 4.9
RF Mmax	mV	5.6 ± 1.7	4.6 ± 2.0	4.5 ± 2.1	5.7 ± 1.9	4.2 ± 1.4	4.4 ± 1.4	5.2 ± 1.8	4.7 ± 1.6	4.7 ± 2.1	5.7 ± 1.9	4.5 ± 1.4	4.5 ± 1.7	5.2 ± 2.0	4.3 ± 1.6	4.4 ± 1.8

Note: Results are presented as the mean ± SD and were analyzed using two‐way ANOVA (n = 13). *p < 0.05 vs. OPT.

Abbreviations: CT, contraction time; MRTD, maximal rate of torque development; MVC, maximal voluntary contraction; QT10:100, low frequency fatigue ratio (QT10/QT100); QTsingle, QT10 and QT100 potentiated twitch peak torque evoked by single, double at low frequency (10 Hz) and double at high frequency (100 Hz) electrical stimulation of the femoral nerve, respectively; RF, rectus femoris; VA, voluntary activation; VL, vastus lateralis; VM, vastus medialis.

3.2. Exercise Performance, Critical Torque and W′

Total impulse performed during the 5‐min all‐out exercises was the highest at optimal knee angle and −15° compared to other knee angles (p < 0.0001, η _p ²  = 0.754; Figure 2). Total impulse then decreased length‐dependently when muscle length decreased or increased. The slopes of torque of the last 30s of each exercise were not different than 0 Nm/s indicating that a plateau was reached at exercise end (−30°: −0.06 ± 0.08 Nm/s; −15°: −0.05 ± 0.15 Nm/s; OPT: −0.05 ± 0.14 Nm/s; +15°: 0.05 ± 0.29 Nm/s; +30°: −0.04 ± 0.13 Nm/s; p = 0.347, η _p ²  = 0.085). When expressed in absolute units, critical torque was not different during −15° and −30°, but decreased during +15° and +30° compared to optimal muscle length (p < 0.0001, η _p ²  = 0.660; Figure 2). When critical torque was expressed relative to maximal torque, no difference was found between optimal muscle length, +15° and +30°, but relative critical torque increased when muscle length was shortened (p < 0.0001, η _p ²  = 0.717; Figure 2). W′ was the greatest at optimal muscle length compared to all other conditions and decreased length‐dependently when muscle length increased or decreased (p < 0.0001, η _p ²  = 0.581; Figure 2). Relative to optimal muscle length, critical torque and W′ were associated with maximal torque.

FIGURE 2.

Effects of muscle length on exercise capacity, critical torque and impulse done above critical torque. Time courses of maximal torque are presented as mean group values on a contraction‐by‐contraction basis and were analyzed using two‐way ANOVA (n = 13). For clarity, error bars were omitted. On the figures presenting total impulse, critical torque and W′, data are presented as group mean (bars) and individuals (white circles and dotted lines) and were analyzed using one‐way ANOVA (n = 13). On the figures presenting the association between maximal torque and critical torque or W′, data are presented as individual changes relative to the value measured during OPT. data were analyzed using mixed‐linear models with random intercept (n = 52). OPT, optimal muscle length; −30°, optimal muscle length minus 30° of knee extension; −15°, optimal muscle length minus 15° of knee extension; +15°, optimal muscle length plus 15° of knee extension; +30°, optimal muscle length plus 30° of knee extension; W′, impulse done above critical torque; *, p < 0.05; **, p < 0.01; ***, p < 0.001; †, p < 0.05 between OPT and −15°; ‡, p < 0.05 between OPT and −30°; +, p < 0.05 between OPT and +15°; #, p < 0.05 between OPT and +30°. For clarity, only the most relevant significant differences are presented.

3.3. Muscle Activation During the 5‐Min All‐Out Exercise

No pre‐ to postexercise difference was found in VL, VM, and RF M_MAX, suggesting that membrane excitability was not impaired (muscle length × time effect, p > 0.375, η _p ²  = 0.083; Table 1). Except during exercise at −30°, VL, VM, and RF RMS_Mmax decreased with time. Mean VL and VM muscle activation as well as BF coactivation increased from −30° to +30° knee angles (Figure 3). Besides a lower RF muscle activation found during the first 20 MVCs at −30° compared to optimal muscle length, no difference in RF muscle activation was found between conditions.

FIGURE 3.

Effects of muscle length on muscle activation during exercise. Data are presented as group mean ± SD for VL, VM, RF and BF muscle activation measured during the 5‐min all‐out exercises. For clarity, only the upper and lower error bars are presented. Data were analyzed using two‐way ANOVA (n = 13). On the figures presenting the association between muscle activation and muscle fatigue (∆QTsingle), data are presented as individual changes relative to the value measured during OPT and were analyzed using mixed‐linear models with random intercept (n = 52). QT_single, potentiated twitch peak torque evoked by single electrical stimulation of the femoral nerve; OPT, optimal muscle length; −30°, optimal muscle length minus 30° of knee extension; −15°, optimal muscle length minus 15° of knee extension; +15°, optimal muscle length plus 15° of knee extension; +30°, optimal muscle length plus 30° of knee extension; VL, vastus lateralis; VM, vastus medialis; RF, rectus femoris; BF, Biceps femoris; ‡, p < 0.05 between OPT and −30°; #, p < 0.05 between OPT and +30°; ∑, p < 0.05 between −30° and +30°; Z, p < 0.05 between −15° and +30°; $, p < 0.05 between −15° and +15°; §, p < 0.05 between −30° and +15°; &, p < 0.05 between −30° and −15°.

3.4. Neuromuscular Fatigue and Recovery Induced by Exercise

Excluding half‐relaxation time which increased, all mechanical indices of neuromuscular fatigue decreased at the end of 5‐min all‐out exercises (time effect, p < 0.0078, η _p ²  = 0.189; Figure 4; Table 1). The exercise‐induced decrease in ∆MVC and ∆QTsingle was nonlinear in all conditions, reaching a plateau on average at the 50th and 40th contractions, respectively (Figure 4). ∆QTsingle increased after the first 10 contractions of the −30° exercise but decreased afterwards. Conversely, the exercise‐induced decrease in voluntary activation was linear throughout exercises (Figure 4).

FIGURE 4.

Effects of muscle length on neuromuscular fatigue. Data are presented as group mean ± SD for neuromuscular fatigue indices measured during the 5‐min all‐out exercises. For clarity, only the upper and lower error bars are presented. Data were analyzed using two‐way ANOVA (n = 13). MVC, maximal voluntary contraction peak torque; VA, voluntary activation; QT_single, QT₁₀ and QT₁₀₀, potentiated twitch peak torque evoked by single, 10 Hz paired and 100 Hz paired electrical stimulation of the femoral nerve, respectively; QT_10:100, Low frequency fatigue ratio. (QT10/QT100); OPT, optimal muscle length; −30°, optimal muscle length minus 30° of knee extension; −15°, optimal muscle length minus 15° of knee extension; +15°, optimal muscle length plus 15° of knee extension; +30°, optimal muscle length plus 30° of knee extension; †, p < 0.05 between OPT and −15°; ‡, p < 0.05 between OPT and −30°; +, p < 0.05 between OPT and +15°; #, p < 0.05 between OPT and +30°.

A muscle length × time effect was found for ∆MVC, ∆QTsingle, ∆QT10, ∆QT100, ∆QT10:100 and ∆MRTD (p < 0.001; η _p ²  > 0.335). No difference was found in the reduction of these indices at the end of +15° and +30° exercises compared to optimal muscle length. In contrast, a lower reduction was found following exercises at −15° and −30° compared to the other conditions (Figure 4; Table 1). No interaction effect was found in voluntary activation suggesting that there was no difference between conditions across the different time‐points (p = 0.218, η _p ²  = 0.103). The reduction in ∆QTsingle was associated with VL, VM, and RF muscle activation (Figure 3). No association was found between ∆QTsingle and W′ (β = 0.19, 95% CI [−0.217; 0.601]; p = 0.361; r ² = 0.016).

A muscle length × recovery time effect was found for ∆QT100 (p = 0.0008, η _p ²  = 0.349; Table 2). No effect was found for the ∆MVC, voluntary activation, ∆QTsingle, ∆QT10 and ∆QT10:100 (p > 0.104, η _p ²  < 0.164). Regardless of muscle length, all the neuromuscular fatigue indices displayed a bi‐phasic recovery, with faster rates of recovery during the first recovery phase (R1; Time effect: p < 0.0019, η _p ²  > 0.568). Recovery then stopped or declined during the next recovery periods (R2, R3). The rates of recovery of ∆QT100 were slower at R1 and R2 following exercise at −30° compared to optimal muscle length, +15° and +30° (Table 2). In addition, the rates of recovery of ∆QT100 were slower at R1 following exercise at −15° compared to optimal muscle length. No difference was found between the rates of recovery that followed exercises at optimal muscle length, +15° and +30°.

TABLE 2.

Rates of recovery of neuromuscular indices following 5‐min all‐out exercises with different quadriceps muscle lengths.

Index (%.min−1)	−30°			−15°			OPT			+15°			+30°
	R1	R2	R3	R1	R2	R3	R1	R2	R3	R1	R2	R3	R1	R2	R3
MVC	13.9 ± 5.6	3.6 ± 3.0	0.3 ± 0.5	16.7 ± 5.9	4.2 ± 2.8	0.3 ± 0.6	15.8 ± 6.3	4.5 ± 4.2	0.9 ± 0.8	14.3 ± 5.3	3.8 ± 4.0	1.1 ± 0.7	12.7 ± 6.9	4.2 ± 3.2	0.7 ± 0.5
VA	3.8 ± 3.7	1.4 ± 1.9	−0.3 ± 0.6	6.6 ± 4.9	1.4 ± 2.3	−0.2 ± 0.6	7.1 ± 5.1	0.9 ± 1.9	0.0 ± 0.5	4.1 ± 7.5	0.8 ± 1.8	0.0 ± 0.4	4.7 ± 6.2	1.3 ± 2.1	0.1 ± 0.2
QTSingle	17.0 ± 6.6	−3.1 ± 4.7	−2.0 ± 1.0	19.0 ± 7.4	0.9 ± 5.0	−1.8 ± 1.0	19.5 ± 7.2	1.1 ± 4.8	−1.4 ± 1.1	19 ± 7.3	1.1 ± 4.3	−1.5 ± 0.9	16.1 ± 5.8	0.5 ± 3.8	−1.2 ± 1.0
QT10	14.4 ± 5.0	−0.8 ± 4.6	−1.8 ± 0.8	17.8 ± 6.0	0.7 ± 5.5	−1.8 ± 0.8	17.7 ± 6.1	1.3 ± 4.7	−1.5 ± 0.9	15.9 ± 5.8	1.2 ± 3.4	−1.3 ± 0.7	13.7 ± 5.0	0.7 ± 2.8	−1.0 ± 0.7
QT100	4.5 ± 4.0*	−1.0 ± 2.7*	−0.8 ± 0.4	8.9 ± 3.4*	0.9 ± 3.6	−0.5 ± 0.5	10.7 ± 3.0	1.8 ± 2.6	−0.4 ± 0.6	11.7 ± 3.4‡	1.8 ± 2.3‡	−0.3 ± 0.5	9.7 ± 3.1‡	1.4 ± 2.2‡	−0.3 ± 0.5
QT10:100	12.8 ± 4.9	0.5 ± 3.2	−1.3 ± 1.1	14.7 ± 5.0	0.4 ± 4.6	−1.7 ± 1.0	15.0 ± 4.5	0.3 ± 4.7	−1.6 ± 1.0	11.8 ± 9.8	0.0 ± 3.0	−1.5 ± 0.7	10.9 ± 4.4	−0.2 ± 2.8	−1.0 ± 0.6

Note: Results are presented as the mean ± SD and were analyzed using two‐way ANOVA (n = 13). QT_single, QT₁₀ and QT₁₀₀, potentiated twitch peak torque evoked by single, 10 Hz paired and 100 Hz paired electrical stimulation of the femoral nerve, respectively; QT_10:100, Low frequency fatigue ratio. The time course of recovery was divided into the following three periods: (R1) from 0 s to 2 min; (R2) from 2 to 4 min; and (R3) from 4 to 15 min. *p < 0.05 vs. OPT; ‡p < 0.05 vs. −30.

Abbreviations: MVC, maximal voluntary contraction; VA, voluntary activation.

3.5. Systemic Responses to Exercise

No difference was found between conditions for all cardiometabolic and ventilatory indices (p > 0.171, η _p ²  < 0.135). V̇O₂ gain was higher during the first 30s of +30° compared to the other conditions (Figure 5). V̇O₂ gain then increased in all conditions throughout exercise. Mean V̇O₂ gain measured during optimal muscle length was not different from the mean V̇O₂ gain measured during −15° and −30° but lower than the mean V̇O₂ gain measured during +15° and +30°. Relative to optimal muscle length, the change in V̇O₂ gain was significantly associated with the change in critical torque and W′ (Figure 5).

FIGURE 5.

Effects of muscle length on oxygen cost of contraction. Data are presented as group mean ± SD for the oxygen cost of contraction measured during the 5‐min all‐out exercises. For clarity, only the upper and lower error bars are presented. Data were analyzed using two‐way ANOVA (n = 13). On the figures presenting the association between $\dot{\mathrm{V}}$O₂ gain and critical torque or W′, data are presented as individual changes relative to the value measured during OPT and were analyzed using mixed‐linear models with random intercept (n = 52). OPT, optimal muscle length; −30°, optimal muscle length minus 30° of knee extension; −15°, optimal muscle length minus 15° of knee extension; +15°, optimal muscle length plus 15° of knee extension; +30°, optimal muscle length plus 30° of knee extension; †, p < 0.05 between OPT and −15°; ‡, p < 0.05 between OPT and −30°; +, p < 0.05 between OPT and +15°; #, p < 0.05 between OPT and +30°.

3.6. Muscle Oxygenation

Except for total[heme] in the vastus lateralis (muscle length effect: p = 0.0013, η _p ²  = 0.774; Figure 6), no muscle length effect was found in all NIRS indices measured in the vastus lateralis or rectus femoris muscles (p > 0.0927, η _p ²  < 0.530). In the vastus lateralis, total[heme] was lower during exercise at −30° compared with the total[heme] recorded during exercise at +30°.

FIGURE 6.

Effects of muscle length on NIRS variables recorded during exercises. Data are presented as group mean ± SD for near infrared spectroscopy indices measured during the 5‐min all‐out exercises. For clarity, only the upper and lower error bars are presented. Data were analyzed using two‐way ANOVA (n = 13). Oxy[heme], concentration of oxygenated heme (hemoglobin and myoglobin); deoxy[heme], concentration of deoxygenated heme; total[heme], total amount of heme (oxygenated and deoxygenated); StO₂, tissue oxygenation index; OPT, optimal muscle length; −30°, optimal muscle length minus 30° of knee extension; −15°, optimal muscle length minus 15° of knee extension; +15°, optimal muscle length plus 15° of knee extension; +30°, optimal muscle length plus 30° of knee extension; ‡, p < 0.05 between OPT and −30°.

4. Discussion

We modulated muscle length to dose‐dependently manipulate quadriceps torque generating capacity and investigate the role of maximal torque in determining exercise capacity (i.e., total impulse), neuromuscular fatigue and the parameters of the torque duration‐relationship. Consistent with our hypothesis, reducing maximal torque through muscle lengthening dose‐dependently decreased total impulse by decreasing critical torque, W′ and accelerating neuromuscular fatigue. In contrast, reducing maximal torque through muscle shortening only reduced total impulse at the shortest muscle length primarily as a consequence of reduced W′. Neuromuscular fatigue induced by exercising at shorter than optimal muscle lengths was preserved length‐dependently due to reduced impairment in muscle contractility, whereas no difference in either global or contractile fatigue was found at longer muscle lengths. Beyond improving our understanding of the effects of muscle length on exercise capacity and neuromuscular fatigue, these findings should provide valuable insights for prescribing exercise interventions at short or long muscle length and for predicting the physiological consequences to these exercises.

4.1. Muscle Length Differentially Affects Critical Torque and Exercise Capacity

One of the original findings of the present experiment is that reductions in maximal torque reduced exercise capacity through different mechanisms depending on whether they were induced by muscle shortening or lengthening. Specifically, reducing maximal torque by muscle shortening had no effect on total impulse done at −15° whereas it decreased total impulse by ~12% at −30°. Conversely, muscle lengthening decreased total impulse by ~21% and ~37% at +15° and +30° compared to optimal angle, respectively. The former results challenge the widely accepted view that exercise capacity at a given relative intensity is largely enhanced at short compared with longer muscle lengths [8, 26, 27]. A potential explanation for this apparent discrepancy is that participants in previous studies may have exercised in different exercise intensity domains despite performing contractions at similar relative intensities. Here, we show that critical torque, the lower boundary of the severe intensity domain [5], was improved by ~25% and ~55% with muscle shortening (Figure 2). Consequently, critical torque represented 45% and 57% of maximal torque at shortened muscle lengths which contrasts with the ~35% measured at optimal or longer muscle lengths. Given that most studies typically prescribed submaximal exercise intensities between 40% and 50% of maximal torque, our findings suggest that participants exercising at shorter muscle length were likely operating below critical torque, whereas those exercising at optimal or longer muscle length were likely above critical torque. Such a shift in critical torque relative to maximal torque dramatically changed the physiological responses to exercise and consequently exercise tolerance [5, 9, 41].

The differential effects of maximal torque reduction on exercise capacity were partly explained by changes in critical torque. Despite major reductions in maximal torque at short muscle lengths, absolute critical torque remained unchanged relative to the optimal muscle length (Figure 2). Conversely, critical torque decreased length‐dependently with muscle lengthening. The differential effects of muscle shortening and lengthening on critical torque partly explain the smaller reduction in exercise capacity observed at short compared with long muscle lengths. While maximal torque surely set an upper limit for absolute critical torque, the fact that a reduction in maximal torque had different outcomes on critical torque suggests that a mechanistic relationship between the variables is unlikely. For example, a moderate association was found between the change in maximal torque and the change in critical torque but careful consideration of this relationship indicates that critical torque decreased only at maximal torque reduction greater than 15% (Figure 2).

Several studies have established a mechanistic link between critical torque and oxygen delivery and metabolism [3, 20, 42, 43]. The present findings provide additional support for this mecanistic link by showing that V̇O₂ gain measured at critical torque (i.e., during the final 30 s of exercise) was reduced at short muscle lengths but increased at long muscle lengths relative to the optimal length. More importantly, changes in V̇O₂ gain were closely associated with changes in critical torque (Figure 5). Considering that no difference in V̇O₂ gain was observed at exercise onset between −30°, −15° and optimal length, our findings suggest that the differences in V̇O₂ gain observed later during exercise did not result from alterations in muscle metabolic function. Indeed, we found that V̇O₂, oxy[heme], deoxy[Heme], StO₂(Figure 6), blood lactate, and V̇CO₂/O₂ were not different between conditions, suggesting that systemic and local metabolic demand was not different between conditions. In addition, several studies in human using 31P‐MRS or on skinned muscle fibers showed that ATP consumption was mainly unchanged by muscle lengths, particularly at short lengths [44, 45, 46]. Taken together, if muscle metabolic function was not altered by muscle length, the differences in V̇O₂ gain found toward exercise end were likely attributable to differences in neuromuscular function, and particularly to the participant's ability to sustain force production and resist the development of fatigue.

4.2. Muscle Length Differentially Affects Neuromuscular Fatigue

By design, critical torque is determined by participants' fatigue resistance (Figures 2 and 4). Our data show that neuromuscular fatigue was preserved in a length‐dependent manner at short muscle length but was unaffected at long muscle length compared with the optimal length. Because no difference in voluntary activation was found between conditions, the preserved neuromuscular fatigue at short muscle length was not the result of difference in muscle activation failure [10, 47]. In contrast, the smaller reduction in postexercise ∆QT_single, ∆QT₁₀, ∆QT₁₀₀ and ∆QT_10:100 at −15° and −30° relative to optimal muscle length suggest that the greater fatigue resistance at short muscle length was the result of reduced contractile impairment (Figure 4). The lack of pre‐ to postexercise difference in VL, VM, and RF M_max, indicates that these contractile impairments were not caused by membrane excitation failure but rather by dysfunction at the excitation‐contraction coupling level, likely involving decreased Ca²⁺ sensitivity or handling [34, 48]. Among others, the mechanisms responsible for the limited contractile fatigue found at short muscle lengths were likely the result of a reduced recruitment of high‐threshold fatigue‐sensitive motor units [49, 50, 51]. At rest, voluntary activation and VL, VM, and RF muscle activation decreased length‐dependently with muscle shortening and increased length‐dependently with muscle lengthening (Figure 1). These results suggest that a significant portion of the quadriceps muscles was not recruited and fatigued at short muscle lengths [12, 13, 50]. This is further supported by our correlative data showing that the degrees of ∆QT_single, ∆QT₁₀ and ∆QT₁₀₀ reduction were moderately associated with the maximal levels of VL, VM, and RF muscle activation reached during exercise (Figure 3). Although several mechanisms were likely at play, the impaired muscle activation at short muscle length was likely the result of decreased Ia afferent input originating from muscle spindles reducing their facilitating effects on motor‐units recruitment [52].

4.3. Maximal Torque Plays a Key Role in Determining W′

Beyond critical torque, the reduced exercise capacity found at all conditions except −15° was also the result of a smaller W′, which decreased by ~16% to ~45% relative to optimal muscle length (Figure 2). In line with previous experiments [53, 54], our correlative data showed that the change in maximal torque relative to the optimal angle was moderately associated with the change in W′ (Figure 2). Reducing maximal torque before exercise therefore appeared to constrain the amount of impulse that could be generated above critical torque. Two components seem to determine this relationship: first, the available torque range that can be produced above critical power (i.e., the instantaneous W′), since a narrower torque reserve would limit the rate at which W′ can be produced; and second, the rate at which neuromuscular fatigue is accumulated, since delaying neuromuscular fatigue would allow a greater total W′ to be produced. This interpretation is consistent with our recent findings showing that slowing or accelerating the rate of neuromuscular fatigue through eccentric or concentric contractions was associated with larger and smaller W′, respectively [20]. In the present experiment, muscle shortening narrowed the torque reserve above critical torque but was accompanied by a slower rate of neuromuscular fatigue than at optimal muscle length (Figure 4). By contrast, muscle lengthening did not alter the torque reserve above critical torque, but was associated with a faster rate of neuromuscular fatigue. Overall, these findings challenge the notion that W′ is mechanistically related to the degree of contractile fatigue, such that a larger W′ is necessarily resulting in greater end‐exercise contractile fatigue [17, 55, 56]. Indeed, neither the present data nor previous studies observed meaningful association between total W′ and postexercise degree of contractile muscle fatigue [13, 16, 20]. Instead, our findings suggest that W′ is more closely related to the ability to delay the rate of fatigue during exercise (i.e., fatigue resistance), thereby enabling a greater impulse to be generated above critical torque.

4.4. Methodological Consideration

Before concluding, several methodological considerations should be acknowledged to give perspective to our dataset. By manipulating maximal torque through several levels of muscle shortening or lengthening, we were able to investigate the role of maximal torque in determining exercise capacity and neuromuscular fatigue while by‐passing the usual limits of comparing different muscle groups or participants. Indeed, differences in muscle size, muscle mass, fiber type distribution, and/or metabolic function could independently influence maximal torque, neuromuscular fatigue and exercise capacity [12, 23, 57, 58]. While this approach is not without limitations and has been adopted previously [18], our experimental design brought novelty in the fact that we investigated two shortened and two lengthened muscle lengths individualized around optimal length to dose‐dependently modulate maximal torque.

Our results demonstrated that changing muscle length modifies critical torque and W′ in distinct ways depending on whether muscle length was shortened or lengthened. These observations reinforce the need for future studies to prescribe exercise intensity according to relevant physiological threshold—such as critical torque or critical power—rather than a fixed percentage of maximal torque or peak power output as previously recommended [21, 22, 59]. Prescribing exercise based solely on maximal capacity may inadvertently place participants in different exercise‐intensity domains despite identical relative intensities, thereby eliciting markedly different physiological responses and confounding comparisons across conditions or populations. This issue was first highlighted by Ansdell et al. [22] who showed that apparent sex differences in exercise tolerance were substantially reduced or even eliminated when exercise intensity was normalized to critical power rather than peak power output.

As also found by other laboratories [35, 60, 61], we reported QT10:100 on average greater than 1 at some knee angles (Table 1). While it was not observed in all of our participants, some subjects had QT10 larger than QT100. This is not something we systematically investigated, but we noticed that this phenomenon was not present when QT was unpotentiated. We thus hypothesized that the increase in QT rate of force development induced by the potentiating effect of contraction had greater effects on QT10 than QT100.

5. Conclusion

This study investigated the effects of muscle length on maximal torque, exercise capacity, neuromuscular fatigue and the parameters of the torque–duration relationship. Compared to optimal muscle length, exercising at long muscle lengths decreased length‐dependently maximal torque, exercise capacity, critical torque and W′ while accelerating neuromuscular fatigue. Likewise, shortening the muscle also produced a length‐dependent reduction in maximal torque and W′, but these effects were offset by a slower rate and lower magnitude of muscle contractile fatigue. These effects preserved critical torque, producing a comparatively smaller reduction in exercise capacity. Collectively, these findings indicate that maximal torque is an important determinant of W′ but plays a more limited role in determining critical torque and the rate of neuromuscular fatigue. More broadly, they highlight that the impact of a reduction in maximal torque on exercise capacity critically depends on the mechanisms through which that reduction is induced. These results provide novel insights on the determinants of exercise capacity and should be a valuable help for practitioners and clinicians to optimize exercise prescription and target specific neuromuscular and metabolic responses.

6. Perspective

Exercise intolerance is a common feature of many neuromuscular and cardiovascular disorders. The results of the present study may help clinicians and practitioners better prescribe exercise intensity and duration to target specific neuromuscular stresses. If greater stress on muscle function is targeted, our data suggest that athletes should exercise at longer‐than‐optimal muscle lengths, which results in greater muscle activation and marked muscle contractile fatigue within short exercise durations. Conversely, if maintaining high mechanical output while limiting disruption of muscle homeostasis is targeted, exercising at shorter‐than‐optimal muscle lengths may be preferable. Further experiments should be conducted to determine whether these stresses effectively produce the targeted adaptations in both healthy and different clinical populations. Finally, critical torque and W′ were determined using the 5‐min all‐out method. Further experiments should determine whether similar effects of muscle length can be found using the gold‐standard, albeit time‐consuming, method of performing multiple trials to exhaustion.

Author Contributions

G. P. D. conceived the study, analyzed, interpreted the data and prepared the manuscript. All authors executed the study and edited, revised and approved the final version of the manuscript. All authors agree to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part are appropriately investigated and resolved. All persons designated as authors qualify for authorship, and all those who qualify for authorship are listed.

Funding

This work was supported by Université de Strasbourg (2021‐022‐10).

Disclosure

The results of the study are presented clearly, honestly, and without fabrication, falsification, or inappropriate data manipulation.

Conflicts of Interest

The authors declare no conflicts of interest.

Data Availability Statement

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