Fatigue-independent alterations in muscle activation and effort perception during forearm exercise: role of local oxygen delivery

Abstract

The oxygen-conforming response (OCR) of skeletal muscle refers to a downregulation of muscle force for a given muscle activation when oxygen delivery (O₂D) is reduced, which is rapidly reversed when O₂D is restored. We tested the hypothesis that the OCR exists in voluntary human exercise and results in compensatory changes in muscle activation to maintain force output, thereby altering perception of effort. In eight men and eight women, electromyography (EMG), oxyhemoglobin (O₂Hb) and deoxyhemoglobin (HHb), forearm blood flow (FBF), and task effort awareness (TEA) were measured. Participants completed two nonfatiguing rhythmic handgrip tests consisting of 5-min steady state (SS) followed by two bouts of 2-min brachial artery compression to reduce FBF by ~50% of SS (C1 and C2), separated by 2 min of no compression (NC1) and ending with 2 min of no compression (NC2). When FBF was compromised during C1, EMG/Force (1.58 ± 0.39) increased compared with SS (1.31 ± 0.33, P = 0.001). However, EMG/Force was not restored upon FBF restoration at NC1 (1.48 ± 0.38, P = 0.479), consistent with C1 evoking skeletal muscle fatigue. When FBF was compromised during C2, EMG/Force increased (1.73 ± 0.50) compared with NC1 (1.48 ± 0.38, P = 0.013). EMG/Force returned to NC1 levels during NC2 (1.50 ± 0.39, P = 0.016), consistent with an OCR in C2. TEA (SS 2.2 ± 2.3, C1 3.9 ± 2.5, NC1 3.4 ± 2.7, C2 4.6 ± 2.7, NC2 3.9 ± 2.8) mirrored changes in EMG. It is noteworthy that during the second compromise and then restoration of muscle oxygenation EMG and TEA were rapidly restored to precompromise levels. We interpreted these findings to support the existence of an OCR and its ability to rapidly modify perception of effort during voluntary exercise.

NEW & NOTEWORTHY In healthy individuals, when force output is maintained during rhythmic handgrip exercise, muscle activation and perception of effort rapidly increase with compromised muscle oxygen delivery (O₂D) and then return to precompromised levels when muscle O₂D is restored. These findings suggest that an oxygen-conforming response (OCR) exists and is able to modify perception of effort during voluntary exercise. Therefore, similar to fatigue, an OCR may have implications for exercise tolerance.

INTRODUCTION

Skeletal muscle fatigue, resulting from the buildup of fatigue-related metabolites, can result in impaired excitation-contraction coupling processes, resulting in reduced force per muscle activation [Force/electromyography (EMG)] (1, 2, 18, 57). This necessitates increased motor neuron activation to maintain force output when skeletal muscle is fatigued. Importantly, perception of effort is thought to originate from efferent neural activity arising parallel to and in proportion to central motor drive (CMD) determining motor neuron recruitment (16, 22, 35, 41, 42, 46, 49). Therefore, the increase in CMD required to maintain force output when the skeletal muscle is fatigued is sensed and perceived as an increase in the subjective perception of effort (5, 16, 45, 46). Thus, by increasing the perceived effort during exercise, skeletal muscle fatigue can play an important role in an individual’s exercise tolerance (26, 47, 55).

Another mechanism that could increase CMD at a given skeletal muscle force output is the downregulation of Force/EMG that can accompany reduced muscle oxygen delivery (O₂D) (24, 29, 34, 40). This phenomenon is known as the oxygen-conforming response (OCR) (30, 31, 56). As opposed to the sustained compromise in the Force/EMG relationship characteristic of skeletal muscle fatigue (46, 57), the OCR is characterized by a rapid (within 40–120 s) restoration of the Force/EMG relationship with restoration of muscle oxygenation (24, 29, 40). This OCR phenomenon has been observed in electrically stimulated skeletal muscle contraction models in humans (24, 40) and cats (29). An important limitation of the electrical stimulation models cited (24, 40) is that supramaximal stimulation is being used; therefore, as opposed to typical exercise that follows the size principle of motor recruitment, all motor units are recruited for every contraction in their stimulated exercise designs (21, 28). Considering this limitation, supramaximal stimulation models do not replicate voluntary exercise where a subpopulation of motor units are recruited to produce the required force in submaximal exercise. Because of this, a plausible response to reduced oxygenation in voluntary exercise could be to rotate activated motor units, thereby reducing the O₂ demand of a given motor unit (8, 9). This would not necessarily result in increased muscle activation and therefore not increase perception of effort. Considering these limitations, the existence of an OCR in voluntary human exercise and its role in modifying perception of effort remain unknown.

Therefore, the objectives of this study were to determine 1) whether an OCR effect can occur in voluntary human exercise and 2) whether perception of effort follows OCR-mediated alterations in EMG. To complete these objectives, forearm muscle activation, blood flow, oxygenation, and perception of effort were measured during rhythmic nonfatiguing voluntary handgrip exercise with intermittent compromise to O₂D. We hypothesized that for the same voluntary force output 1) muscle activation would increase with decreased muscle oxygenation, 2) muscle activation would rapidly decrease upon muscle reoxygenation, and 3) perception of effort would follow changes in muscle activation.

METHODS

Participants

Sixteen healthy participants (8 men, 8 women; see Table 1 for anthropometric data) were recruited for this study. Participants were instructed to adhere to the following before testing: to forego strength training exercise for 24 h before and to abstain from alcohol, smoking, and caffeine consumption for 12 h before and eating 4 h before testing. All experimental procedures were reviewed and approved by the Queen’s University Health Sciences and Affiliated Teaching Hospitals Research Ethics Board and conformed to the standards set by the Declaration of Helsinki. On the first of two visits, participants provided verbal and written informed consent as well as completing the Canadian Society for Exercise Physiology Get Active Questionnaire before voluntary participation in this study.

Table 1.

Age, anthropometric measures, and nonfatiguing work rate

Variable	Group Average (n = 16)
Age, yr	25 ± 3.12
Height, cm	174 ± 9.03
Weight, kg	71.22 ± 11.06
BMI, kg/m2	23.44 ± 2.14
%MVC	26 ± 6

Values are means ± SD for n participants. BMI, body mass index; MVC, maximal voluntary contraction.

Instrumentation

Surface electromyography.

Surface EMG was measured via four mini wireless bipolar electrodes (25 mm × 12 mm × 7 mm) with a 10-mm interelectrode distance (Delsys Trigno Wireless EMG system, Natick, MA). Before sensor placement, the approximate location of EMG sensor placement was cleaned of any hair and thoroughly scrubbed with rubbing alcohol. The electrodes were then placed over the four muscle bellies of the right flexor digitorum superficialis (FDS). The four muscle belly sites were identified by following the guide designed by Bickerton et al. (11) for injection of botulinum toxin into the four muscle bellies of the FDS. Briefly, a landmarking line (LL) was drawn between the right medial epicondyle and the pisiform. The length of the LL was recorded, and the approximate distance from the medial epicondyle to the four muscle bellies was marked (FDS4 49%, FDS3 54%, FDS2 72%, FDS5 76%) (11). Now, from the marks made along the LL, measurements were made laterally to the exact location of each muscle belly and were then marked (FDS4 0.7 cm, FDS3 1.7 cm, FDS2 1.4 cm, FDS5 0.6 cm) (11). EMG electrical activity was measured in volts and upsampled to an even 2,000 Hz/channel.

Force.

Force of handgrip contraction was recorded via a handgrip force transducer connected to PowerLab/8_SP (ADInstruments, Colorado Springs, CO), collected at a frequency of 200 Hz/channel.

Forearm blood flow.

During exercise, brachial artery blood velocity and diameter were measured continuously. Brachial artery blood velocity was measured with a 4-MHz pulsed flat Doppler probe (model 500V 131 Transcranial Doppler; Multigon Industries, Mt. Vernon, NY) attached to the skin over the brachial artery proximal to the antecubital fossa of the exercising arm. Brachial artery diameter was measured with a linear echo ultrasound probe, positioned over the brachial artery ~5 cm proximal to the Doppler probe, operating at 13 MHz in 2D mode (Vivid i; GE Medical Systems, London, ON, Canada).

Muscle oxygenation.

Muscle oxygenation was measured with a commercially available near-infrared spectroscopy (NIRS) system (OxyMon MkIII; Artinis Medical Systems, Einsteinweg, The Netherlands). The NIRS device provided measurements of change in oxyhemoglobin (O₂Hb) and deoxyhemoglobin (HHb). This study used a single channel consisting of one optical fiber operating at three wavelengths (874, 761, and 803 nm), where the light was transmitted into the muscle and received by an avalanche photodiode. NIRS data were collected at a sample rate of 10 Hz with an interoptode distance of 25 mm and a differential path length factor of 3.59, with participant-specific power and gain settings, ensuring optimal signal performance. The NIRS probes were placed longitudinally on the participants’ right forearm between the two distal and two proximal mini-EMG sensors, where the sites were shaved and cleaned before the probe was secured in place.

Perception of effort.

The task effort awareness (TEA) scale, a 15-point scale developed by Swart et al. (54), was used to measure perception of effort: the amount of attention, mental effort, and difficulty experienced while maintaining the prescribed physical task. Participants were able to see the scale continuously throughout exercise and were asked to report their TEA 45 and 90 s into each segment [i.e., a new period of either brachial artery compression (BAC) or no compression] starting at steady state. Participants were asked to verbalize the number that represented their current perception of effort.

To ensure proper reporting of perception of effort in isolation from peripheral afferent feedback associated with discomfort (3, 47), all participants were required to complete a 10-point questionnaire [developed by Swart et al. (54)] used to establish the difference between sense of effort and physical sensations. This questionnaire was completed on day 1 during screening and then repeated on day 2 before beginning exercise. Clarification of the answers was provided when needed. As per Christian et al. (16), the example that “a brief maximal effort requires a maximal conscious effort despite only inducing a small amount of peripheral discomfort” was explained to every participant (48).

Central hemodynamic measures.

Heart rate was monitored with a three-lead electrocardiogram with electrodes attached to the skin in standard CS₅ placement (17). Arterial oxygen saturation was monitored with a pulse oximeter (Nellcor N-395; Covidien-Nellcor, Boulder, CO) placed over the index finger of the participants’ nonexercising hand. A finger photoplethysmograph (Finometer MIDI; Finapres Medical Systems) was used to measure heart rate and mean arterial blood pressure and to provide estimates of stroke volume, cardiac output, and total peripheral resistance via ModelFlow (Finapres Medical Systems).

Experimental Design

This was a within-participant repeated-measure design in which all participants completed a minimum of six exercise tests during a 2-day collection period. Screening and fatigue threshold identification occurred on one day, and the control and BAC protocols occurred on another day. All data collections were completed a minimum of 24 h apart. Each participant completed both data collection sessions at the same time of day, although time of day differed between participants. All data collections were completed with the participants lying supine on a table with their arms resting on tables at heart level, abducted ~70° to their respective sides. All testing sessions were completed in a temperature-controlled room (19–21°C), and isometric handgrip exercise was completed after a contraction:relaxation duty cycle of 2 s:2 s as guided by visual force output and metronome cues.

Reducing muscle oxygenation.

To reduce muscle oxygenation, local forearm blood flow (FBF) was reduced via manual BAC proximal to the antecubital fossa. A single researcher was tasked with reducing mean blood velocity (MBV) by ~50% of the steady-state MBV measured during a control trial completed on day 2. The researcher completing BAC used live feedback from the MBV measured via a 4-MHz pulsed flat Doppler probe and displayed on a monitor. The researcher was provided with a visual target between 40% and 60% of steady-state MBV on LabChart 7 (ADInstruments) for simplified MBV tracking.

TEA familiarization.

Before completing exercise in data collection session 1 or 2, participants were asked to read a description of both the physical rating of perceived exertion (P-RPE) and the TEA scale (54). As per Swart et al. (54), participants were asked to complete a 10-point true/false questionnaire, testing their understanding of the TEA scale. Any incorrect answers were clarified before continuing to exercise testing. Additionally, participants were asked their TEA during all exercise sessions on day 1 in order to orient them to the concept before the control and BAC protocols completed on day 2.

Experimental Protocol and Measurement Details

Identification of nonfatiguing exercise work rate.

On day 1, participants were asked to complete three 3-s maximal voluntary contractions (MVCs), each separated by 1 min. EMG and force output were recorded during MVCs to provide an estimate of maximal EMG activity and force output. Subsequently, forearm fatigue threshold identification was accomplished with a noncontinuous incremental ramp protocol. All participants were asked to remain still for 1 min while silent EMG values were taken; then they were asked to complete 13 min of rhythmic forearm handgrip exercise targeting 10% of their highest MVC. Throughout this protocol EMG data were recorded continuously on the exercising arm from the four muscle bellies of the FDS. Participants were given a 5-min break in which a regression analysis of EMG amplitude was completed for each muscle. The regression analysis was used to identify fatigue onset. Specifically, fatigue was evident if the EMG amplitude regression slope was significant positive. This fatigue onset or fatigue threshold has been suggested to represent the aerobic-anaerobic transition point, indicating the point where adenosine triphosphate demand exceeds the rate of supply from primarily aerobic sources (23). See Fig. 1 for a full description of this protocol.

Fig. 1.

Fatigue threshold identification on day 1. Participants began forearm handgrip exercise at 10% of their maximal voluntary contraction (MVC). After 13 min at this intensity, participants were asked to relax for 5 min while electromyography (EMG) data were analyzed to check the EMG amplitude regression slope for this %MVC. If the EMG amplitude slope was significant (YES), participants decreased %MVC by 5% and the 13-min test was repeated. If the EMG amplitude slope was not significant (NO), participants increased %MVC by 10% and the 13-min test was repeated. The protocol was terminated once a YES-NO had been identified (in that order), confirming the highest level %MVC sustainable for 13 min, the length of the experimental protocol on day 2. If a NO-YES-YES occurs, %MVC that is sustainable is assumed to be the first NO. Fatigue threshold = YES-NO or NO-YES-YES. This %MVC was confirmed on day 2.

Control.

On day 2, participants completed three 3-s MVCs, each separated by 1 min. Participants then completed a control exercise trial. This control was used to confirm that the %MVC identified on day 1 did not cause fatigue over the duration of the 13-min test. To ensure that no fatigue was present, the same fatigue identification analysis as used in the identification of nonfatiguing exercise work rate was used. See Fig. 2 for a summary of this protocol. A 10-min break was provided before continuing to the reduced muscle oxygenation protocol.

Fig. 2.

Control protocol completed on day 2. The % maximal voluntary contraction identified on day 1 was completed for 13 min with no compression (NC) following a contraction:relaxation cycle of 2 s:2 s. Abbreviations outside of parentheses represent time points for comparison with the brachial artery compression protocol. C, compression; SS, steady state; T, task effort awareness recording.

Reduced muscle oxygenation protocol.

Once confirmation of the fatigue threshold was completed and successfully passed, FBF was manipulated during the 13-min rhythmic forearm handgrip exercise at the confirmed nonfatiguing exercise work rate (see Fig. 3). During this exercise, participants had their right arm resting at heart level on a cushioned table. Before beginning the experimental manipulation the designated researcher manually located the brachial artery proximal to the antecubital fossa.

Fig. 3.

Brachial artery compression (BAC) protocol. Steady-state forearm handgrip exercise was completed at each participant’s individually determined % maximal voluntary contraction work rate. Exercise began with 5 min of no-compression (NC) handgrip exercise following a contraction:relaxation cycle of 2 s:2 s. Starting at minutes 5 and 9, BAC (C) was maintained for 2 min. This protocol was completed twice for all participants. SS, steady state; T, task effort awareness recording.

As Fig. 3 depicts, participants achieved a steady-state (SS) condition by completing 5 min of rhythmic handgrip exercise at the confirmed nonfatiguing exercise work rate. At the beginning of the 5th and 9th minutes of exercise BAC was completed for 2-min periods (C1 and C2). Participants were asked to maintain force of contraction throughout, where the only change during exercise was BAC. After completion of the first bout of the BAC protocol, a 10-min break was provided and then the BAC protocol was repeated. See Fig. 3 for a summary.

Data Acquisition

Identification of nonfatiguing exercise work rate.

EMG data were recorded continuously, and each muscle contraction was used to complete the regression analysis required for identification of an individual’s maximal nonfatiguing exercise work rate.

Control.

Confirmation was completed with the same method as that described during the identification of nonfatiguing exercise work rate protocol. Additionally, all recorded data were averaged into 1.5-min time bins. Specifically, data were averaged going from 15 s to 1 min 45 s into a new segment. Two measures of TEA (45 s apart) were averaged to provide one value for each segment.

Reduced muscle oxygenation protocol.

Data acquisition techniques were identical between the control and reduced muscle oxygenation protocols.

Data Analysis

Identification of nonfatiguing exercise work rate.

With the EMG data collected through the incremental exercise test, a linear regression of the EMG root mean square (RMS) was created for each exercise intensity. The fatigue threshold was identified by finding the lowest work rate to have a significant positive EMG RMS slope coefficient and then reducing work rate by 5% to identify the highest work rate that did not have a significant positive EMG RMS slope (13, 19, 20). This work rate was used for all tests completed on day 2.

EMG analysis was completed with MATLAB 2017a (MathWorks, Natick, MA). Raw EMG data were processed by detrending the raw EMG data, converting it into absolute values, and then subtracting the baseline noise from the signal. Once cleaned, the signal was filtered with a Butterworth filter (2 Hz cutoff). The EMG activity from the MVC was then processed to identify the peak EMG activity. Subsequently, the filtered EMG exercise data were processed to identify the on/off cycles for every muscle contraction. On/off cycles were identified with an on/off EMG threshold equal to 5% of the peak EMG amplitude seen during the participant’s maximal MVC. Once the on/off cycles were identified and grouped into vectors, these vectors were used to quantify the amplitude and median frequency (mF) of each on/off cycle. Amplitude was quantified by measuring the RMS of each on/off cycle from the raw EMG data. Alternatively, mF was measured by completing a fast Fourier transform taking raw EMG signal from the time domain into the frequency domain and then calculating the mF. Any unacceptable contractions (i.e., time of contraction being <1.5 s or >2.5 s) were removed. With function “regstats” a linear regression of amplitude was tested to be different from 0, and further visual analysis was used to confirm fatigue onset.

Control and BAC protocols.

During the control and BAC tests, EMG, NIRS, echo, and Doppler ultrasound were measured continuously. The control analysis was the same for EMG RMS as that completed during the identification of nonfatiguing work rate protocol. Data from both trials of the BAC protocol were averaged together to provide one representative data set for the BAC protocol.

forearm blood flow.

Brachial artery diameters were quantified with Measurements from Arterial Ultrasound Imaging (MAUI, Hedgehog Medical). FBF was calculated as MBV (cm/s)·60 s/min·π [brachial artery diameter (cm)/2]².

muscle activation.

EMG was quantified by completing the same analysis for EMG completed during the identification of nonfatiguing exercise work rate described above. After completion of the steps mentioned above, the EMG RMS of each muscle belly were averaged together to get one value for each BAC protocol. These EMG RMS data were then divided by the average force output over the same period. This was done to account for any involuntary changes in force output between perfusion pressure conditions.

force output.

Force output data were analyzed to confirm that contractions were consistently completed throughout FBF conditions. Specifically, code was written to first clean the force data by zeroing the baseline voltage. Next, the on/off cycles for every contraction were identified by setting a threshold of 1 kg. With on/off vectors identified, the average kilogram output was calculated over each cycle. The length of each contraction was calculated and unacceptable contractions removed from the analysis, as per EMG analysis above. Force was measured in kilograms, which was then converted to newtons and then presented as a percentage of the individual’s MVC.

muscle oxygenation.

NIRS data were measured in arbitrary units. To provide a meaningful relative comparison, all arbitrary units were compared to estimated maximal and minimal level of both O₂Hb and HHb. To obtain estimates of maximal and minimal O₂Hb and HHb, arm cuff occlusion was completed for 5 min, followed by 2 min of rest. Measurements of maximal HHb and minimal O₂Hb were calculated as an average of the highest and lowest values measured (respectively) before cuff release. Measurements of minimal HHb and maximal O₂Hb were calculated as an average of the lowest and highest values measured (respectively) during the 2-min period following cuff release. When negative values were present, the lowest measured value was added to all measured values. Addition of the lowest value to all values provided a measure of 0–100% HHb and O₂Hb. All values were then calculated as a percentage of the maximal measured value, providing estimates of change in muscle HHb and O₂Hb.

perception of effort.

TEA was recorded 45 and 90 s into each segment starting at steady state (see Fig. 3). These two values were averaged together to provide a single value for TEA in each segment.

Statistical Analysis

Identification of nonfatiguing work rate.

A linear regression of EMG RMS was completed for all exercise intensities completed during the maximal nonfatiguing exercise work rate identification. A P < 0.01 was required to indicate a significant slope.

Control and BAC protocols.

To test the main effects of condition (control and BAC) and time on EMG RMS/Force, TEA, FBF, O₂Hb, HHb, and mF, two-way repeated-measures ANOVAs with sex as a between-subject factor were completed. When assumptions of sphericity were violated, a Greenhouse-Geisser correction was used. In the case of a significant F statistic, pairwise comparisons were assessed with Bonferroni correction. When assumptions of normality were not met, a Friedman’s nonparametric test was run to identify a main effect of time within condition. In the case of a significant χ²-statistic, a Wilcoxon signed-rank test was run for pairwise comparisons. P < 0.05, P < 0.01, and P < 0.005 were required for parametric tests, pairwise comparisons between conditions, and nonparametric post hoc tests, respectively. Statistical analyses were performed with SPSS version 24.0 software (SPSS, Chicago, IL). Data are presented as means ± SD.

RESULTS

Participants

Twenty-one participants began this study; two were excluded because of inadequate EMG signal; one was excluded because of bifurcation of the brachial artery proximal to the site of compression, thereby making complete BAC and FBF measurements impossible; one dropped out after day 1; and finally one was excluded from analysis because the manipulation was unsuccessful in reducing muscle oxygenation. Therefore, 16 (8 women, 8 men) participants completed this study (see Table 1 for anthropometric data).

Identification of Nonfatiguing Work Rate

Completion of the incremental exercise test to identify a participant’s specific maximal nonfatiguing exercise work rate identified an average nonfatiguing work rate of 26 ± 6% MVC.

Control

Did FBF change over time during the control trial?

Control FBF was significantly different over time [F(2.040, 26.523) = 7.556, P = 0.002] (see Table 3 and Fig. 5 for pairwise comparisons).

Table 3.

Measures of hemoglobin, blood flow, muscle activation, and effort sensation over time within BAC and control protocols

BAC protocol
FBF, ml/min	C1 (102 ± 62)	NC1 (277 ± 172)	C2 (114 ± 68)	NC2 (294 ± 189)
SS (226 ± 144)	P < 0.001†	P = 0.001†	P < 0.001†	P = 0.001†
C1 (102 ± 62)		P < 0.001†	P = 0.005	P < 0.001†
NC1 (277 ± 172)			P < 0.001†	P = 0.013
C2 (114 ± 68)				P < 0.001†
O2Hb, %	C1 (34 ± 15)	NC1 (61 ± 16)	C2 (37 ± 14)	NC2 (63 ± 18)
SS (57 ± 17)	P < 0.001*	P = 0.061	P < 0.001*	P = 0.029*
C1 (34 ± 15)		P < 0.001*	P = 0.336	P < 0.001*
NC1 (61 ± 16)			P < 0.001*	P = 0.890
C2 (37 ± 14)				P < 0.001*
HHb, %	C1 (68 ± 18)	NC1 (48 ± 11)	C2 (69 ± 17)	NC2 (47 ± 12)
SS (47 ± 12)	P = 0.001†	P = 0.836	P < 0.001†	P = 0.0836
C1 (68 ± 18)		P < 0.001†	P = 0.918	P < 0.001†
NC1 (48 ± 11)			P < 0.001†	P = 0.379
C2 (69 ± 17)				P < 0.001†
EMG RMS/Force	C1 (1.58 ± 0.39)	NC1 (1.48 ± 0.38)	C2 (1.73 ± 0.50)	NC2 (1.50 ± 0.39)
SS (1.31 ± 0.33)	P = 0.001*	P < 0.001*	P = 0.001*	P = 0.012*
C1 (1.58 ± 0.39)		P = 0.479	P = 0.124	P = 1.000
NC1 (1.48 ± 0.38)			P = 0.017*	P = 1.000
C2 (1.73 ± 0.50)				P = 0.020*
TEA	C1 (3.9 ± 2.5)	NC1 (3.4 ± 2.7)	C2 (4.6 ± 2.7)	NC2 (3.9 ± 2.8)
SS (2.2 ± 2.3)	P < 0.001*	P = 0.142	P = 0.001*	P = 0.040*
C1 (3.9 ± 2.5)		P = 1.000	P = 0.079	P = 1.000
NC1 (3.4 ± 2.7)			P < 0.001*	P = 0.022*
C2 (4.6 ± 2.7)				P = 0.007*
Control protocol
FBF, ml/min	C1 (187 ± 107)	NC1 (196 ± 116)	C2 (201 ± 114)	NC2 (201 ± 125)
SS (178 ± 104)	P = 0.237	P = 0.021*	P = 0.004*	P = 0.043*
C1 (187 ± 107)		P = 1.000	P = 0.240	P = 0.780
NC1 (196 ± 116)			P = 0.300	P = 1.000
C2 (201 ± 114)				P = 1.000
TEA	C1 (2.3 ± 2.4)	NC1 (2.7 ± 2.7)	C2 (3.2 ± 2.8)	NC2 (3.6 ± 2.8)
SS (1.8 ± 2.3)	P = 0.092	P < 0.048*	P = 0.017*	P < 0.005*
C1 (2.3 ± 2.4)		P = 0.304	P = 0.063	P = 0.016*
NC1 (2.7 ± 2.7)			P = 0.011*	P = 0.005*
C2 (3.2 ± 2.8)				P = 0.028*

Values are means ± SD. BAC, brachial artery compression; C, compression; EMG, electromyography; FBF, forearm blood flow; HHb, deoxyhemoglobin; NC, no compression; O₂Hb, oxyhemoglobin; RMS, root mean squared; SS, steady state; TEA, task effort awareness.

^*Significant difference for parametric test.

^†Significant difference for nonparametric tests.

Fig. 5.

Percent oxyhemoglobin (%O₂Hb; A) and percent deoxyhemoglobin (%HHb; B) in response to no manipulation (control) and to brachial artery compression (BAC) targeting a 50% reduction in forearm blood flow (FBF; C) for control and BAC conditions during nonfatiguing handgrip exercise. SS, steady state; C, compression; NC, no compression. *Significantly different from control at a given measurement time. ^#Significantly different from SS within a condition. †Significantly different from C1 within a condition. ‡Significantly different from NC1 within a condition. ^&Significantly different from C2 within a condition. Significance: P < 0.05 for all comparisons except BAC FBF, where P < 0.005.

Did HHb or O₂Hb change over time during the control trial?

Control HHb was not significantly different between time points [F(1.977, 27.672) = 1.348, P = 0.276]. Control O₂Hb was not significantly different between time points [F(1.938, 27.130) = 2.323, P = 0.119].

Did EMG/Force remain constant throughout the control trial?

Control EMG/Force was not significantly different between time points [F(4, 56) = 0.817, P = 0.519]. There was no time × sex interaction [F(4, 56) = 1.294, P = 0.283].

Did TEA remain constant throughout the control trial?

There was a significant [F(1.273, 17.816) = 13.825, P = 0.001] increase in TEA over time during the control protocol (see Table 3 and Fig. 6 for pairwise comparisons). There was no time × sex interaction [F(1.273, 17.816) = 0.396, P = 0.587].

Fig. 6.

Task effort awareness (TEA; A) and electromyography root mean square (EMG RMS) per unit of force (B) for control and brachial artery compression (BAC) conditions during nonfatiguing handgrip exercise. SS, steady state; C, compression; NC, no compression. *Significantly different from control at a given measurement time, P < 0.01. ^#Significantly different from SS within a condition, P < 0.05. ‡Significantly different from NC1 within a condition, P < 0.05. ^&Significantly different from C2 within a condition, P < 0.05.

Reduced Muscle Oxygenation Protocol

Was force output consistent throughout BAC exercise?

Force output was not effectively maintained throughout the length of the BAC test [χ²(4) = 16.400, P = 0.003]. Post hoc tests determined that force output was significantly higher in SS (21.24 ± 4.6%) compared with C2 (20.56 ± 4.3%, P = 0.001) and no compression (NC)2 (20.87 ± 4.5%, P = 0.002) (see Table 2). See Fig. 4 for a representation of raw hand grip data during BAC.

Table 2.

Quantification of muscle activation and force output during BAC protocol

Variable	SS	C1	NC1	C2	NC2
EMG RMS, %MVC	28.3 ± 11	33.8 ± 12*	31.9 ± 12*	36.5 ± 15*†	32.1 ± 12*
Force, %MVC	21.2 ± 4.6	20.9 ± 4.5	21.0 ± 4.7	20.6 ± 4.3*	20.9 ± 4.5*
mF, Hz	157 ± 18	156 ± 21	154 ± 19	157 ± 21	154 ± 18

Values are means ± SD. BAC, brachial artery compression; C, compression; EMG, electromyography; mF, median frequency; MVC, maximal voluntary contraction; NC, no compression; RMS, root mean squared; SS, steady state.

^*Significantly different from SS;

^†significantly different from NC1. All P < 0.05.

Fig. 4.

Individual response demonstrating the increase in electromyography (EMG) upon brachial artery compression-evoked compromise to forearm blood flow [represented by brachial artery blood velocity (BABV)] and the rapid reduction of EMG when oxygen delivery was restored. Top: EMG. Middle: BABV. Bottom: hand grip force output.

Was FBF reduced during BAC?

There was a significant effect of time on FBF during the BAC protocol [χ²(4) = 57.05, P < 0.001]. Within BAC FBF was successfully reduced during C1 and C2. Additionally, blood flow was restored after cessation of BAC at NC1 and NC2 (see Table 3 and Fig. 5C for within-condition pairwise comparisons) (see Table 4 for between-condition within-time pairwise comparisons). See Fig. 4 for a representation of raw FBF data during BAC.

Table 4.

Measures of hemoglobin, blood flow, muscle activation and effort sensation compared between BAC and control within time

	BAC
Control	SS (210 ± 133)	C1 (94 ± 56)	NC1 (258 ± 159)	C2 (107 ± 64)	NC2 (272 ± 174)
FBF, ml/min
SS (178 ± 104)	P = 0.003*
C1 (187 ± 107)		P < 0.001*
NC1 (196 ± 117)			P < 0.001*
C2 (202 ± 125)				P < 0.001*
NC2 (201 ± 126)					P < 0.001*
O2Hb, %	SS (57 ± 17)	C1 (34 ± 15)	NC1 (61 ± 16)	C2 (37 ± 14)	NC2 (63 ± 18)
SS (44 ± 20)	P = 0.019
C1 (45 ± 20)		P = 0.055
NC1 (46 ± 21)			P = 0.008*
C2 (46 ± 21)				P = 0.083
NC2 (47 ± 21)					P = 0.008*
HHb, %	SS (46 ± 11)	C1 (66 ± 13)	NC1 (47 ± 12)	C2 (67 ± 13)	NC2 (46 ± 12)
SS (51 ± 13)	P = 0.035
C1 (52 ± 13)		P < 0.001*
NC1 (52 ± 14)			P = 0.007*
C2 (53 ± 14)				P < 0.001*
NC2 (53 ± 15)					P = 0.003*
EMG RMS/Force	SS (1.31 ± 0.33)	C1 (1.58 ± 0.39)	NC1 (1.49 ± 0.38)	C2 (1.73 ± 0.50)	NC2 (1.50 ± 0.39)
SS (1.3 3 ± 0.34)	P = 0.681
C1 (1.35 ± 0.36)		P = 0.005*
NC1 (1.36 ± 0.35)			P = 0.022
C2 (1.41 ± 0.41)				P = 0.008*
NC2 (1.40 ± 0.38)					P = 0.036
TEA	SS (2.2 ± 2.3)	C1 (3.9 ± 2.5)	NC1 (3.4 ± 2.7)	C2 (4.6 ± 2.7)	NC2 (3.9 ± 2.8)
SS (1.8 ± 2.3)	P = 0.428
C1 (2.3 ± 2.4)		P = 0.018
NC1 (2.7 ± 2.7)			P = 0.385
C2 (3.2 ± 2.8)				P = 0.084
NC2 (3.6 ± 2.8)					P = 0.717

Values are means ± SD. BAC, brachial artery compression; C, compression; EMG, electromyography; FBF, forearm blood flow; HHb, deoxyhemoglobin; NC, no compression; O₂Hb, oxyhemoglobin; RMS, root mean squared; SS, steady state; TEA, task effort awareness.

^*Significant difference.

Was muscle oxygenation decreased during BAC?

oxyhemoglobin.

There was a significant effect of time on O₂Hb during the BAC protocol [F(1.278, 19.168) = 42.2, P < 0.001]. Within BAC, O₂Hb was successfully reduced during C1 and C2. After cessation of BAC, O₂Hb was restored at NC1 and NC2 (see Table 3 and Fig. 5A for within-condition pairwise comparisons; see Table 4 for between-condition within-time pairwise comparisons).

deoxyhemoglobin.

There was a significant effect of time on HHb during the BAC protocol [χ²(4) = 44.5, P < 0.001]. Within BAC, HHb was successfully increased during C1 and C2. After cessation of BAC, HHb was restored at NC1 and NC2 (see Table 3 and Fig. 5B for within-condition pairwise comparisons; see Table 4 for between-condition within-time pairwise comparisons).

Did EMG RMS/Force increase with BAC?

There was a significant effect of time on EMG RMS/Force during the BAC protocol [F(2.442, 34.194) = 17.135, P < 0.001]. Within BAC, EMG RMS/Force was increased at C1 and C2; however, EMG RMS/Force was not restored after C1 at NC1, although EMG RMS/Force was restored after C2 at NC2 (see Table 3 and Fig. 6B). There was no main effect of sex [F(1,14) = 0.868, P = 0.367] or time × sex interaction [F(2.442, 34.194) = 0.426, P = 0.696] on EMG RMS/Force (see Table 4 for between-condition within-time pairwise comparisons). See Fig. 4 for a representation of raw EMG data during BAC.

Did TEA follow changes in EMG/Force?

There was a significant effect of time on TEA during the BAC protocol [F(1.620, 22.684) = 15.299, P < 0.001]. BAC led to an increase in TEA at C1 and C2; however, TEA was not restored after C1 at NC1, although TEA was restored after C2 at NC2 (see Table 3 and Fig. 6A). There was no main effect of sex [F(1,14) = 0.009, P = 0.927] or time × sex interaction [F(1.620, 45.637) = 1.178, P = 0.316] on TEA.

Did EMG median frequency change with BAC?

There was not a statistically significant effect of time on mF [F(4, 60) = 1.394, P = 0.247; see Table 2].

Was EMG RMS/Force different between control and BAC?

Compared with control, BAC led to an increase in EMG RMS/Force during C1 and C2. Compared with control, EMG RMS/Force was not different during NC1 and NC2 in the BAC protocol (see Table 4 and Fig. 6B).

Was TEA different between control and BAC?

BAC led to an increase in TEA during C1 and approached significance in C2. TEA was not different during NC1 and NC2 in the BAC protocol compared with control (see Table 4 and Fig. 6A).

DISCUSSION

The purpose of this study was to determine whether an OCR can occur during voluntary exercise in humans and whether the OCR can play a role in modifying perception of effort. The key findings from the present study were as follows. 1) EMG/Force increased during both bouts of compromised forearm muscle O₂D. 2) EMG/Force rapidly returned to precompromise values after the second period of compromised forearm muscle O₂D (i.e., measurement NC2 following C2), although this was not the case after the initial compromise to forearm muscle O₂D (i.e., measurement NC1 following C1). 3) Perception of effort reflected changes in EMG/Force, such that TEA increased and decreased with EMG/Force. 4) Perception of effort increased over time during a control trial without an increase in EMG/Force.

We interpret the above findings to support the existence of an OCR in voluntary human exercise and its ability to rapidly modify perception of effort. The increased EMG/Force required to maintain voluntary force of contraction under conditions of reduced muscle oxygenation and the rapid restoration of EMG/Force upon muscle reoxygenation (measurement NC2 following C2; see Fig. 6) are consistent with the OCR seen in stimulated human (24, 40) and animal (29–32) exercise. In contrast, the lack of restoration in EMG/Force with muscle oxygenation restoration after the first compromise to muscle O₂D (measurement NC1 following C1; see Fig. 6) is not consistent with an OCR. Rather, this response is consistent with mechanisms of skeletal muscle fatigue (36, 57). Importantly, perception of effort followed changes in EMG/Force whether EMG/Force increased because of OCR or because of hypothesized skeletal muscle fatigue mechanisms. Therefore, similar to skeletal muscle fatigue, we interpret these findings to support the potential for an OCR to impact exercise tolerance. Finally, the increase in perception of effort over time during control, not associated with increased EMG/Force, suggests there can be an increase in perception of effort over time during exercise that may be independent of CMD.

Rapid Increase and Decrease in EMG/Force with Compromise and Restoration of Oxygenation

Skeletal muscle fatigue is characterized by a sustained increase in the EMG/Force relationship, i.e., muscle activation (i.e., EMG) must increase to maintain force (5, 46), or if muscle activation stays constant force production is reduced (25, 57). With skeletal muscle fatigue, a sustained compromise to the EMG/Force relationship is still evident for at least 2–3 min after exercise cessation, and often for considerably longer (15). Importantly, this fatigue-related compromise is independent of O₂D (37, 50, 57). In contrast, the OCR compromise to EMG/Force when muscle oxygenation is reduced is rapid, but so is its restoration (within 40–120 s) upon muscle reoxygenation, even while the muscle continues to exercise (24, 29, 40). Therefore, the OCR can be distinguished from skeletal muscle fatigue by identifying whether there is a rapid restoration of EMG/Force after restoration of oxygenation.

The OCR is well established in electrically stimulated exercise models (24, 29–32, 40). In contrast, only one study has quantified changes in EMG during voluntary human exercise (29). Importantly, the immediacy of the restoration of EMG/Force relationship was not identified. Hobbs and McCloskey (29) had participants perform plantar flexion exercise, during which the exercising limb was moved from a heart-level (i.e., uncompromised O₂D) position to 30° and 60° above heart level (i.e., compromised O₂D) and then back to heart level. An important limitation of Hobbs and McCloskey’s (29) design was that neither muscle blood flow nor muscle oxygenation was measured; therefore, the interpretation that muscle activation increased as a result of a decrease in muscle O₂D was never directly tested. Furthermore, considering the recent identification of a vasodilator compensator phenotype (10), it may be that Hobbs and McCloskey’s (29) design was not effective in reducing exercising limb blood flow and therefore O₂D.

To circumvent the issues associated with the vasodilator compensator phenotype, the present study used alternating 2-min periods of free flow and BAC to reduce FBF by ~50% of that measured at steady state. Consistent with the characteristics of the OCR, EMG/Force increased during the second bout of BAC and was rapidly restored to precompromise levels (see Fig. 6B) upon restoration of muscle O₂D (Fig. 5). We interpret these findings to support the existence of an OCR in voluntary human exercise. In contrast, Luu and Fitzpatrick (40) hypothesized that this response is dependent on reduction and restoration of exercising muscle blood flow affecting the rate of interstitial K⁺ clearing. However, this hypothesis proposed by Luu and Fitzpatrick (40) has yet to be tested. Instead, what has been confirmed is that restoration of EMG/Force (consistent with the OCR) is evident regardless of whether force depression and restoration at a given EMG were achieved by manipulation of blood flow (24, 29, 34, 40) or arterial oxygen content (33, 34). Considering that the OCR effect occurs with arterial oxygen content manipulation, the OCR cannot be explained by a K⁺ washout phenomenon. Furthermore, given that the OCR response is the same with blood flow or arterial oxygen content manipulation of O₂D, it is most likely that it is the oxygenation per se that is determining the EMG/Force in both conditions.

In contrast to the second O₂D compromise, the first O₂D compromise led to an increase in EMG/Force that was not restored upon restoration of O₂D (Fig. 6B; time of measurement: SS → C1 → NC1), indicating skeletal muscle fatigue (37, 50, 57). In fact, this lack of EMG/Force restoration reinforces the interpretation that the increase in EMG/Force during the second compromise is an OCR and not skeletal muscle fatigue. But it also begs the question: “Why wasn’t there additional fatigue during the second compromise to O₂D?”

Recent research investigating the effect of local and remote ischemic preconditioning (IPC) (7, 27, 38, 39) as well as metabolic preconditioning (MP) (6, 12, 14) suggests that a bout of ischemia or high-intensity exercise, respectively, before exercise testing may have an ergogenic effect. In the case of IPC, a recent meta-analysis identified a small effect of IPC on overall exercise performance, the largest effect being on aerobic exercise (51). Importantly, in some cases certain individuals may require a combination of both MP and IPC to see a performance-enhancing effect (53). In the present study the condition of reduced O₂D while maintaining exercise is the equivalent of a combined IPC and MP condition. Although the mechanisms of MP and IPC are not fully understood, the IPC and MP work leads us to propose the following testable hypothesis: It may be that the initial compromise to O₂D acted as a type of metabolic and low-oxygen preconditioning. In doing so, the second bout of compromised O₂D was protected against further fatigue development.

In summary, the rapid compromise and then restoration of EMG/Force (Fig. 6B; time of measurement: NC1 → C2 → NC2) is consistent with an O₂D dependence, namely, an OCR. It may be that the first bout of compromised O₂D acted to protect against an additional fatigue effect, thereby bringing the OCR to the forefront during the second compromise to O₂D.

Oxygen-Conforming Effects Contribute to Perception of Effort

Skeletal muscle fatigue has important implications for exercise tolerance. As skeletal muscle fatigue becomes more severe, the amount of CMD required to maintain a given force output increases (4, 5, 43, 46). Importantly, when CMD increases, efferent sensations (conscious awareness of CMD activating motor neurons) also increase. This increase in efferent sensations is measured as an increase in an individual’s perception of effort (4, 5, 44, 46). Although it is well established that skeletal muscle fatigue can act to increase the effort/force relationship, it was not known whether changes in muscle oxygenation can act rapidly to modify the effort/force relationship.

During the BAC protocol, perception of effort followed changes in EMG/Force (Fig. 6). Specifically, when O₂D was compromised (see Fig. 5) both EMG/Force and perception of effort increased (see Fig. 6). As mentioned above, with the first restoration of O₂D the EMG/Force levels remained elevated. Consistent with TEA reflecting efferent perception of effort, the TEA score also remained elevated. In response to the second O₂D compromise and restoration sequence, TEA followed the compromise in EMG/Force with O₂D compromise and the restoration of EMG/Force when O₂D was restored (see Fig. 6). These observations that TEA mirrored EMG/Force when it remained compromised during the first O₂D restoration, and also mirrored EMG/Force when it was rapidly restored after the second O₂D restoration, eliminate the possibility that TEA was simply a function of the physical act of BAC and strongly support that TEA reflects changes in muscle activation. We therefore conclude that whether muscle activation increases because of skeletal muscle fatigue or an OCR, perception of effort follows changes in muscle activation. These are to our knowledge the first findings to suggest that, similar to skeletal muscle fatigue, an OCR can also increase the effort/force relationship when O₂D is compromised.

During the control protocol, perception of effort increased over time despite the lack of changes in EMG/Force relationship (see Fig. 5). Interestingly, perception of effort and EMG/Force were not different between the BAC and control conditions at either of the NC time points during exercise. We interpret these findings to suggest that there is an inherent increase in the amount of attention, mental effort, and difficulty experienced while continuing to exercise that is not associated with changes in muscle activation.

Methodological Considerations

There are several methodological considerations that warrant discussion. First, we were not able to conduct potentiated twitch force assessment of fatigue, and we did not measure peripheral fatigue-related metabolites during exercise. Therefore, our interpretation that there are fatigue mechanisms involved in the first increase in EMG/Force is based on the response of EMG/Force alone. However, the consistent experimental support documenting compromised EMG/Force at a given muscle oxygenation (5, 25, 57) as characteristic of skeletal muscle fatigue is highly consistent with our interpretation that the maintained compromise in EMG/Force following the first BAC reflects skeletal muscle fatigue.

Second, we did not ask participants to rate their P-RPE; therefore, we are unable to report whether participants would have rated their P-RPE to change any differently from TEA during exercise or whether participants are even able to distinguish between these hypothesized afferent (P-RPE) and efferent (TEA) sensations during voluntary handgrip exercise (54). Measurement of P-RPE was left out in order to ensure that participants’ attention was exclusively directed to interpreting sensations of effort while completing voluntary handgrip exercise (52). Third, participants were not blinded to the BAC. Therefore, it may have been possible for participants to modify their subjective perceptions in response to the act of brachial artery palpation. However, because we see a lack of TEA decrease following the first compression, when BAC is stopped (Fig. 6; time of measurement: C1 → NC1), we interpret these findings to argue strongly against the sensation of palpation leading to increased TEA.

Fourth, by design, the experimental protocol started with control in order to confirm that the work rate identified on day 1 was in fact nonfatiguing before continuing to the BAC protocol. Thus the issue of a possible order effect impacting assessment of differences between BAC and control must be considered. However, given that steady-state measures of EMG RMS/Force were not different between control and BAC protocols, it is unlikely that there were any residual effects of control on BAC (see Fig. 6).

Finally, the magnitude of reduction in perfusion and oxygenation employed was substantial in order to ensure an adequate deoxygenation effect on EMG/Force and was only performed under nonfatiguing forearm exercise. Whether less of a change in oxygenation is required to evoke an OCR response at higher exercise intensities remains to be determined. Importantly, an exercise intensity dependence of the OCR has been demonstrated in stimulated exercise (24); therefore, although we chose to use a nonfatiguing work rate to minimize fatigue as a confound, it would seem important to explore this potential OCR exercise intensity dependence in voluntary human exercise. Therefore, our “proof of principle” finding would suggest that pursuing such a question is a worthwhile next step.

Conclusions

This study is the first to simultaneously measure muscle oxygenation, EMG, and force during intermittent controlled alterations in oxygenation to determine the existence of an OCR in voluntary human exercise. Specifically, we have demonstrated that for a given skeletal muscle force output, muscle activation increases when muscle oxygenation is decreased and that this response is rapidly reversible upon muscle reoxygenation. Furthermore, we have demonstrated that perception of effort, as measured by the TEA scale, follows changes in muscle activation associated with changes in muscle oxygenation during voluntary small-muscle mass exercise. In conclusion, similar to skeletal muscle fatigue, we interpret these findings to support the ability of an OCR to modify perception of effort during voluntary rhythmic small-muscle mass exercise. This may have implications for the OCR as a contributor to exercise tolerance.

GRANTS

Support for this project was provided by Natural Sciences and Engineering Research Council (NSERC) of Canada Discovery Grant 250367-11 and Research Tools and Instruments Grant EQPEQ0407690-11 as well as infrastructure grants from the Canadian Foundation for Innovation and the Ontario Innovation Trust (to M. E. Tschakovsky).

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

P.J.D. and M.E.T. conceived and designed research; P.J.D., Z.I.N.K., O.K.M., M.J.T.L., and A.M.F. performed experiments; P.J.D. analyzed data; P.J.D. and M.E.T. interpreted results of experiments; P.J.D. prepared figures; P.J.D. drafted manuscript; P.J.D., Z.I.N.K., O.K.M., M.J.T.L., A.M.F., and M.E.T. edited and revised manuscript; P.J.D., Z.I.N.K., O.K.M., M.J.T.L., A.M.F., and M.E.T. approved final version of manuscript.