The present work was approved by the Research Ethics Committee of the Federal University of Santa Catarina and was conducted in accordance with the Declaration of Helsinki. After being fully informed of the risks and stresses associated with the study, the participants gave their written informed consent to participate.

Twenty healthy participants (10 females: age 28 ± 6 years; mass 58 ± 7 kg; height 161 ± 5 cm; 10 males: age 26 ± 5 years; mass 75 ± 7 kg; height 177 ± 5 cm) volunteered to participate in the study. Subjects undertook exercise at a recreational level (3–4 sessions per week; 150–300 min per week), and were familiar with laboratory exercise testing procedures. Women performed the constant work-rate tests in the follicular phase being performed 2–4 days after the menses.

Subjects were required to visit the laboratory on 5 occasions. On the first visit, each subject performed a maximal ramp test for the determination of the GET, VO_2peak and peak power output (P_peak). On subsequent visits, subjects performed bouts of very heavy-intensity exercise immediately following unloaded (20 W – U-VH) or moderate (95% GET – M-VH) baseline to verify the effects of work-to-work transitions on VO₂ kinetics and muscle force behavior (Figure 1). A maximal isokinetic effort (MIE; constant pedal cadence at 60 or 120 rpm) was performed following a standardized warm-up and immediately following the constant work-rate cycling bout to quantify reductions in peak torque. Subjects were instructed to avoid any intake of caffeine for 3 h, or alcohol and strenuous exercise in the 24 h preceding the test sessions and to arrive at the laboratory in a rested and fully hydrated state, at least 2 h post-prandial. All tests were performed at the same time of day in a controlled environmental laboratory condition (19–22°C; 50–60% RH) to minimize the effects of diurnal biological variation. Subjects performed only one test on any given day, and each test was separated by at least 48 h but completed within a period of 2 weeks.

All tests were performed on an electromagnetically braked cycle ergometer (Excalibur Sport PFM, Lode BV, Groningen, Netherlands). Respiratory and pulmonary gas exchange variables were measured using a breath-by-breath analyzer (Quark PFTergo, Cosmed, Rome, Italy). Before each test, the O₂ and CO₂ analysis systems were calibrated using ambient air (20.94% O₂ and 0.03% CO₂) and a gas of a known O₂ and CO₂ concentration (16.00% O₂ and 5.00% CO₂) according to the manufacturer’s instructions. Likewise, the turbine flow meter was calibrated before each test using a 3 L syringe (Quark PFTergo, Cosmed, Rome, Italy). A monitor connected to the gas analyzer was used to measure heart rate (HR). Capillary blood samples (25 μl) were obtained from the earlobe of each subject and the blood lactate concentration ([La]) was measured using an electrochemical analyzer (YSI 2700 STAT, Yellow Springs, OH, United States). The cycle ergometer, the breath-by-breath analyzer and the electrochemical analyzer were calibrated in accordance with specific manufacturer’s recommended procedures.

On the first laboratory visit, 15 min after the isokinetic sprint familiarization subjects performed an incremental ramp test for the determination of the GET, VO_2peak, and P_peak. After a 4 min period of cycling at 20 W (baseline), an incremental ramp test to exhaustion was undertaken with power output increasing by a rate of 30 W.min^-1 from the baseline. Subjects were instructed to maintain their preferred cadence (female 77 ± 6 rpm; male 82 ± 6 rpm) throughout the test. The preferred cadence along with saddle and handle bar height and configuration was recorded and replicated in subsequent tests. Each subject was verbally encouraged to undertake maximal effort. The test was terminated when the cadence fell by more than 10 rpm below the preferred cadence for more than 5 s despite strong verbal encouragement (Black et al., 2015). Breath-by-breath pulmonary gas exchange and HR data were measured continuously during the test and averaged over 15 s periods. VO_2peak was defined as the highest value obtained in a 15 s interval, or if a VO₂ plateau observed, it was considered as the average of the final minute of exercise (Day et al., 2003). The attainment of VO_2peak was defined using the criteria proposed by Bassett and Howley (2000). The P_peak was considered as the highest power output attained during the test. The GET was determined using a cluster of measurements as the V-slope method and the ventilatory equivalent method (Beaver et al., 1986). The data from the ramp test were used to calculate the work rate corresponding to 60% Δ (i.e., GET plus 60% of the difference between the work-rate at the GET and VO_2peak). The lag in VO₂ during incremental exercise taken into account by a deduction of two-thirds of the ramp rate from the work-rate at the GET (Burnley et al., 2011).

The cycle ergometer was instrumented with pedal force measurements (Lode PFM, Groningen, Netherlands) to quantify muscle fatigue during the MIE. A switch from the hyperbolic mode to the isokinetic mode happened instantaneously when required. This protocol was similar to previously used protocols (Cannon et al., 2011; de Souza et al., 2016; Hopker et al., 2016; do Nascimento Salvador et al., 2018) and considered as reliable (do Nascimento Salvador et al., 2018) (no differences were observed for both peak torque – intraclass correlation coefficient = 0.99, typical error = 3.7% and, peak power output – intraclass correlation coefficient = 0.99; typical error = 4.2% obtained during the pre-exercise assessments). Peak torque was assessed for each subject by a 5 s cycling sprint test in the isokinetic mode at 60 or 120 rpm. In the pre-exercise muscle function assessment, subjects performed a 5 min warm-up at 95% GET immediately followed by the 5 s MIE. After this, subjects performed 5 min of active recovery at 95% GET and a period of 15 min rest before the main exercise bouts. The MIE was repeated immediately after the exercise in U-VH and M-VH (see Figure 1). Subjects were given an auditory cue to begin the all-out effort in the seated position and strong verbal encouragement was given throughout the 5 s. The torque and power data were recorded continuously during the MIE. As described by Altenburg et al. (2007), the peak torque in each crank arm was determined as the average of the four consecutive highest torque values (2 s). Thus, the peak torque during the MIE was then considered as the average of the peak values of both left and right crank arms.

Breath-by-breath data for each test were initially examined to exclude outlier values caused by sighs, swallowing and coughs (Lamarra et al., 1987). After that, for each exercise transition, the breath-by-breath data was processed and analyzed as described previously (do Nascimento et al., 2015; de Souza et al., 2016; do Nascimento Salvador et al., 2018). Non-linear regression techniques were used to fit the data after the onset of a fundamental phase with an exponential function (OriginPro 8; OriginLab). An iterative process ensured that the sum of squared errors was minimized. Due to large inter-individual differences in the duration of the exponential region (Murgatroyd et al., 2011), the identification of the VO₂sc during the VH intensity exercise was performed individually. The fundamental VO₂ kinetics (phase II) was isolated following the iterative method to identify the exponential region (Rossiter et al., 2001, 2002; Murgatroyd et al., 2011). The identification at the end of the fundamental phase (i.e., TD_s) was performed by fitting a window from the start of the fundamental phase (i.e., after 20 s cardio-dynamic phase) initially set at 60 s. The window was lengthened iteratively until the exponential model fit demonstrated a discernible and consistent departure from the measured VO₂-values by considering the criteria proposed in literature (Rossiter et al., 2001, 2002; Murgatroyd et al., 2011). Thus, the fitting window was constrained to this time point and a single-exponential fitting was performed only on the fundamental phase to identify the kinetics parameters. The model was constrained in VO_2baseline to aid in the identification of the key parameters according to the following equation:

where: VO₂(t) represents the value of VO₂ at a given time (t); VO_2baseline is the average value over the last minute of baseline cycling; A is the asymptotic amplitude for the exponential term describing changes in VO₂ from baseline to its asymptote; τ is the time constant; and the TD is the time delay. The VO_2SC was calculated according to the Equation (2):

Where: VO_2end is the average VO₂ value over the last 20 s of a 6 min exercise bout.

Descriptive statistics are expressed as mean ± standard deviation. The Shapiro–Wilk test was applied to ensure a Gaussian distribution of the data (n < 50). A two-way mixed-model ANOVA was used to analyze the interaction over time and condition. Assumptions of sphericity were assessed using the Mauchly test, and any violation was corrected using the Greenhouse-Geisser correction factor. The Shapiro–Wilk test was used to verify the normality of residuals. When significant effects were observed the Bonferroni post hoc test was used for pairwise comparisons. Analyzes were performed using the Statistical Package for Social Sciences Windows (SPSS Inc. version 17.0; Chicago, IL, United States). The level of significance adopted was set at p < 0.05.

During M-VH, the VO_2baseline was significantly higher, A, TD, and VO_2sc lower and τ and TDs slower compared to U-VH for both sexes (p < 0.05). There were no significant differences for Atotal (i.e., A + VO_2baseline) and VO_2end between conditions for both sexes (p > 0.05; Table 2). Female subjects presented lower amplitudes, VO_2sc and VO_2end than male counterparts (p > 0.05; Figure 2). The [La] post-exercise was not significantly different between conditions or sexes (p > 0.05; Figure 3). There was no difference between conditions (p = 0.23) in both sexes (p = 0.31) for rate of perceived exertion (RPE), thus, male and female subjects perceived the effort in a similar way after the exercise in both conditions (Figure 4).

The putative mechanism likely explaining the alterations in VO₂ dynamics preceded by elevated baseline VO₂ may be represented by the balance between the parasympathetic and sympathetic control of the HR. Elevated work rates seem to alter parasympathetic withdrawal leaving the slower sympathetic control to mediate increases in HR (Hughson and Morrissey, 1982; Bearden and Moffatt, 2001; DiMenna et al., 2010). It has been suggested that a slowing of the HR kinetics may limit the O₂ delivery to cellular respiration (Hughson and Morrissey, 1982; DiMenna et al., 2010). Alternatively, cellular respiration might adjust more slowly in muscle fibers that are already active and/or recruitment of motor units that are believed to possess slower VO₂ kinetics and a higher VO₂ cost of tension (Poole et al., 2008). Likewise, the fundamental PCr τ is lengthened and the fall in PCr is greater compared to U-VH transitions (Jones et al., 2008). These mechanisms are consistent with a greater proportional involvement of muscle fibers that are positioned higher in the recruitment hierarchy (e.g., type II muscle fibers) (Henneman et al., 1965; Jones et al., 2008; Dimenna et al., 2010). Although the present study does not allow to distinguish between the effect of elevated VO₂ and elevated work-rate (Please, see the Supplementary Material about this discussion), our results extend the effects of elevated baseline work rate from previous literature for female subjects (Hughson and Morrissey, 1982; Wilkerson and Jones, 2006; Jones et al., 2008; Dimenna et al., 2010; Da Boit et al., 2014; Wüst et al., 2014).

It is important to acknowledge that training status largely influences O₂ kinetics, presenting higher phase II τ-values in untrained vs. trained subjects (Koppo et al., 2004). Our phase II τ-values are similar compared with participants from similar training status (recreationally active) at similar intensities (60% Δ) (DiMenna et al., 2010; Cannon et al., 2011; Da Boit et al., 2014; Keir et al., 2016). Despite the differences in the amplitudes of the fundamental and slow phases between sexes in both U-VH and M-VH, τ was not significantly different between sexes despite an elevated baseline. These results are in contrast to previous findings reported for τ in adolescents (Fawkner and Armstrong, 2003; Lai et al., 2016), but in agreement in relation to VO_2SC. Besides, we extend these results to a healthy population during cycling compared to the results of Reis et al. (2017) which showed no differences in τ between women and men trained swimmers during heavy-intensity swimming. Moreover, our results are in accordance to findings reported for middle aged subjects during moderate cycling (O’Connor et al., 2012). To date, no study has investigated the effect of work-to-work transition on the VO₂ kinetics and MFP comparing sexes. Although males present higher fundamental amplitudes and similar τ-values compared to females, the gross rate of increase of oxygen uptake per second is higher in men, suggesting a quicker onset (Reis et al., 2017). According to Reis et al. (2017) this could be due to the higher VO_2peak and anatomic differences (e.g., muscle mass) presented by males. It has been reported that females could present lower O₂ delivery, O₂ extraction and blood flow due to smaller hearts, smaller lung volumes and lower diffusion capacities and cardiac outputs (Harms, 2006; Dominelli et al., 2015). The present study found significant lower values for VO_2peak, power at GET and VE for females compared to their male counterparts. This could be a result of the anatomic differences in the cardiorespiratory system. According to Murias et al. (2013) females were less effective in matching O₂ delivery and O₂ utilization, which may be due to a lower tonic sympathetic activity in women leading to systemic differences in blood flow distribution. However, a lower oxygen delivery to the muscles seems not to influence the VO₂ kinetics in high-intensity exercise (Reis et al., 2017). Olfert et al. (2004) affirmed that female subjects did not experience greater O₂ diffusion limitations during exercise, which may emphasize the importance of absolute lung size or aerobic fitness in determining susceptibility to gas exchange impairment rather than sex per se.

Previous literature demonstrated a relation between VO₂ kinetics parameters (i.e., VO_2SC, τ) and the loss in torque/force production during high-intensity exercise (Cannon et al., 2011; Keir et al., 2016; Temesi et al., 2017). Cannon et al. (2011) proposed that greater levels of muscle fatigue are reflected by a larger amplitude of the VO_2SC or vice-versa. The mechanisms contributing to peripheral muscle fatigue has been suggested to contribute to an increased O₂ cost of exercise (Keir et al., 2016). Moreover, Temesi et al. (2017) suggested that subjects with slower VO₂ kinetics (higher τ-values) experience a greater level of peripheral muscle fatigue, specifically, more excitation-contraction coupling failure. Thus, it was expected that alterations of MFP behavior would be displayed in the VO₂ kinetics. In a case of cause-effect relationship, it was expected that higher τ-values would be related with a greater muscle force loss. Or in other hand, lower VO_2SC would be related with a lower muscle force decrement. However, we could not confirm this hypothesis. The decrease in torque production in both conditions and for both sexes were not related or linked with the VO_2SC or τ. Our results are in accordance with Hopker et al. (2016) who showed that exercise-induced muscle damage led to significant locomotor muscle fatigue, but did not alter the VO_2SC or τ during subsequent high-intensity cycling. Hopker et al. (2016) also affirmed that the results from Cannon et al. (2011) could be misleading because the relationship presented by these authors was considering different intensity domains. Deley et al. (2006) demonstrated an inverse relation between the level of muscle fatigue in type II muscle fibers induced by an electromyostimulation protocol and the VO_2SC during subsequent high-intensity cycling. Additionally, τ was not altered by this intervention. Moreover, Thistlethwaite et al. (2008) investigated two types of prior exercise (knee extension vs. cycling) which caused different activation patterns/levels of additional motor units (∼38% in knee extension vs. 21% in cycling) and found similar VO₂ responses (VO_2SC or τ) during the subsequent bout of heavy cycling exercise. The authors concluded that muscle fatigue is neither the primary determinant of the VO_2SC nor does it affect the τ of the fundamental phase. Recently, do Nascimento Salvador et al. (2018) showed that prior VH cycling changed the VO_2SC and the trajectory of the VO_2SC in a subsequent VH cycling exercise, but did not alter MFP behavior. The present study is in line with these findings and challenges a “cause-effect” relationship between VO_2SC and muscle fatigue.

The present study observed no differences in Δtorque at 60 and 120 rpm following M-VH and U-VH for males and at 120 rpm (Figure 5) for females despite the differences in the VO₂ kinetics. The differences in VO₂ fundamental and slow amplitudes as well as in muscle torque production in absolute values could indicate a possible link between muscle force and VO₂ kinetics and this hypothesis was not discarded. However, no correlation between Δtorque and VO_2SC or τ was found in any condition or velocity in both sexes. These results do not support previous literature proposing a causative relation between muscle fatigue and VO₂ kinetics (Cannon et al., 2011; Keir et al., 2016; Temesi et al., 2017). However, female participants showed a smaller decrement in torque production in M-VH at MIE 60 rpm. This may be explained by a smaller percentage of type II fibers and a greater fatigue resistance in females compared to males (Maher et al., 2009; Liu et al., 2010; Hunter, 2014). Further, a reduced proportion of type II muscle fibers may explain the smaller decrements in torque production by females observed at 120 rpm at which the reliance on type II muscle fibers increases. Considering the higher percentage of type II muscle fibers in men (Maher et al., 2009), it may be suggested that proportionally more type II fibers were activated at 120 rpm (Sargeant, 2007) and consequently, led to a greater level of muscle fatigue compared to 60 rpm and compared to women. When exercise is preceded by an elevated baseline work rate, it may be suggested that the recruitment of additional motor units which are characterized by a smaller mitochondrial content and a higher VO₂ cost per unit of force (i.e., type II fibers) becomes inevitable in order to maintain the exercise intensity (Wüst et al., 2014). Thus, female subjects may have shown a lower Δtorque in M-VH at 60 rpm because of a lesser reliance on type II fibers.

It is noteworthy that Temesi et al. (2017) did not find a relationship between changes in maximal voluntary activation and τ-values. They stated a lack of relationship between τ-values and central fatigue. According to Hopker et al. (2016), the RPE measured indirectly supports the assumption that a level of central motor command was required in order to produce the power output. This study found that RPE immediately post-exercise was neither different between conditions (U-VH vs. M-VH) nor sexes. Based on these considerations, a dissociation between the changes in VO₂ kinetics (i.e., slower τ and the lower VO_2SC) and a decrease in central motor drive may be suggested. The performance of a maximal isokinetic cycling effort taken immediately post-exercise is a “global” but ecologically valid protocol to quantify the decrease in MFP in cycling. However, the identification of the origin of fatigue (i.e., central or peripheral) is not possible. Beelen et al. (1995) suggest that fatigue in this type of dynamic isokinetic exercise may be due to changes in the muscle itself and not due to failure of central drive. Considering that muscle fatigue measured by MIE recovers back to baseline values within 1–3 min (Beelen et al., 1995; Coelho et al., 2015; do Nascimento Salvador et al., 2018), it represents a tendency to indicate more the peripheral fatigue.