Intensity matters: effects of cadence and power output on corticospinal excitability during arm cycling are phase and muscle dependent

Abstract

The present study investigated the effects of cadence and power output on corticospinal excitability to the biceps (BB) and triceps brachii (TB) during arm cycling. Supraspinal and spinal excitability were assessed using transcranial magnetic stimulation (TMS) of the motor cortex and transmastoid electrical stimulation (TMES) of the corticospinal tract, respectively. Motor-evoked potentials (MEPs) elicited by TMS and cervicomedullary motor-evoked potentials (CMEPs) elicited by TMES were recorded at two positions during arm cycling corresponding to mid-elbow flexion and mid-elbow extension (i.e., 6 and 12 o’clock made relative to a clock face, respectively). Arm cycling was performed at combinations of two cadences (60 and 90 rpm) at three relative power outputs (20, 40, and 60% peak power output). At the 6 o’clock position, BB MEPs increased ~11.5% as cadence increased and up to ~57.2% as power output increased (P < 0.05). In the TB, MEPs increased ~15.2% with cadence (P = 0.013) but were not affected by power output, while CMEPs increased with cadence (~16.3%) and power output (up to ~19.1%, P < 0.05). At the 12 o’clock position, BB MEPs increased ~26.8% as cadence increased and up to ~96.1% as power output increased (P < 0.05), while CMEPs decreased ~29.7% with cadence (P = 0.013) and did not change with power output (P = 0.851). In contrast, TB MEPs were not different with cadence or power output, while CMEPs increased ~12.8% with cadence and up to ~23.1% with power output (P < 0.05). These data suggest that the “type” of intensity differentially modulates supraspinal and spinal excitability in a manner that is phase- and muscle dependent.

NEW & NOTEWORTHY There is currently little information available on how changes in locomotor intensity influence excitability within the corticospinal pathway. This study investigated the effects of arm cycling intensity (i.e., alterations in cadence and power output) on corticospinal excitability projecting to the biceps and triceps brachii during arm cycling. We demonstrate that corticospinal excitability is modulated differentially by cadence and power output and that these modulations are dependent on the phase and the muscle examined.

INTRODUCTION

An early investigation using the adult decerebrate cat found that independent of locomotor speed, locomotor force production (i.e., greater limb displacement) could be enhanced with increasing electrical stimulation intensity of the mesencephalic locomotor region (Shik et al. 1966). This suggested that locomotor intensity, specifically speed and force production, may be controlled differently by subcortical areas involved in locomotion. There is also evidence that cortical neurons may not be responsible for the coding for muscle force output during locomotion of increased intensity. Although pyramidal neurons discharged rhythmically during locomotion in the cat, the mean and peak discharge of the neurons was not related to the speed or the intensity (walking on a flat vs. incline surface) (Armstrong and Drew 1984). It is important to note that although the discharge rate did not increase, the electromyography (EMG) recorded from the limb muscles did, indicating that additional motoneuron output was present in the absence of changes in pyramidal discharge rates and likely produced by subcortical mechanisms, including those at the spinal level.

In humans, the gain of spinal reflex pathways and ascending sensory pathways has been shown to be differentially modulated by changes in locomotor intensity (Hundza et al. 2012; Hundza and Zehr 2009; Larsen et al. 2006; Pyndt et al. 2003; Sakamoto et al. 2004). Most of these studies report velocity- or cadence-dependent modulation, although load-dependent modulation in supraspinal and spinal reflex excitability has also been reported. For example, the suppression of somatosensory-evoked potentials (SEPs), short-latency stretch reflexes (SLRs), and H reflexes have all been reported during leg cycling with increased cadence, while changes in cycling load did not affect SEP or SLR amplitudes, and increased H-reflex amplitudes (Hundza et al. 2012; Larsen et al. 2006; Larsen and Voigt 2004; Sakamoto et al. 2004). Thus the results from these studies suggest that supraspinal and spinal reflex excitability is modulated differently during locomotor outputs and is dependent on the manner in which the intensity is altered.

Substantially less information is currently available regarding the modulation of excitability in the corticospinal pathway during locomotor output of varying intensities (Forman et al. 2015; Spence et al. 2016; Weavil et al. 2015). The excitability of the corticospinal pathway can be assessed by measuring the amplitude of responses evoked by transcranial magnetic stimulation (TMS) known as motor-evoked potentials (MEPs). However, MEPs provide information regarding the excitability of the entire corticospinal pathway and therefore can be influenced by changes at the supraspinal and/or spinal level. As a means to understand the spinal contributions to changes in MEPs, transmastoid electrical stimulation (TMES) of corticospinal axons can be applied to elicit cervicomedullary motor-evoked potentials (CMEPs), which are a measure of spinal excitability (Taylor 2006). Therefore, when TMS and TMES are both used, it can be logically deduced whether a change in MEP is likely due to either supraspinal or spinal factors. Using these techniques, we recently demonstrated cadence (Forman et al. 2015)- and power output-dependent (Spence et al. 2016) changes in supraspinal and spinal excitability to muscles of the upper limb during arm cycling. More specifically, we reported that corticospinal excitability to the biceps brachii was enhanced throughout arm cycling with increased cadence, while spinal excitability was phase dependent (Forman et al. 2015). When cycling power output was manipulated, overall corticospinal excitability to the biceps and triceps brachii increased during both phases of arm cycling, whereas the pattern of spinal excitability to the biceps and triceps brachii tended to increase with power output (Spence et al. 2016). Thus, during arm cycling, alterations in intensity (i.e., changes in cadence or power output) have differential effects and may be important in determining corticospinal excitability. Additionally, these findings suggest that intensity-dependent modulations in corticospinal excitability are also muscle dependent. In these previous studies, however, we did not compare the effects of cadence and power output on corticospinal excitability nor did we examine the effects of cadence on triceps brachii excitability.

The purpose of the present study was to assess the influence of locomotor intensity on phase-dependent modulation of corticospinal excitability projecting to both the biceps and triceps brachii during arm cycling. For this study, we define intensity as an increase in either cadence or power output. We assessed corticospinal excitability to the biceps and triceps brachii at two positions during arm cycling (mid-elbow flexion and mid-elbow extension) at combinations of two different cadences (60 and 90 rpm) at three relative power outputs (20, 40 and 60% of peak power output). It was hypothesized, based on our previous work (Forman et al. 2015; Spence et al. 2016) that 1) corticospinal excitability to both the biceps and triceps brachii would increase in each position examined (flexion and extension; see methods) as cycling intensity (both cadence and power output) increased; 2) supraspinal excitability would account for increases in corticospinal excitability during the less active phase of each muscle (i.e., elbow extension for the biceps and elbow flexion for the triceps brachii), while spinal excitability would be largely responsible for the increase in excitability during the most active phase of each muscle (i.e., elbow flexion for the biceps and elbow extension for the triceps brachii); and 3) supraspinal and spinal excitability would be modulated differently with changes in cadence vs. changes in power output. Thus we hypothesized that supraspinal and spinal excitability would be differentially modulated depending on the muscle, phase, and type of intensity assessed.

METHODS

Ethical Approval

Before data collection, all participants received verbal explanation of the experimental protocol. Once all questions were answered, written informed consent was obtained. This study was conducted in accordance to the Helsinki declaration, and all protocols were approved by the Interdisciplinary Committee on Ethics in Human Research (ICEHR No. 20170217-HK). Additionally, the protocols were carried out in accordance with the Tri-Council Guidelines in Canada, with the potential risks being fully disclosed to all participants.

Participants

Thirteen healthy, recreationally active (>10 h of physical activity per week), male volunteers (24.8 ± 4.4 yr of age, height = 179 ± 7.2 cm, weight = 86 ± 9.5 kg, 10 right-hand dominant, and 3 left-hand dominant), with no known neurological impairments participated in part 1 of the study. Ten of those participants also participated in part 2 (see below). Following consent, all participants completed a safety checklist to screen for contraindications to magnetic stimulation (Rossi et al. 2009) and a Physical Activity Readiness Questionnaire to screen for any contraindications to physical activity (Canadian Society for Exercise Physiology 2002). Hand dominance was determined using the Edinburgh handedness inventory (Veale 2014). This was done to ensure that the evoked potentials (see Stimulation Conditions) were measured from the dominant arm, given that differences in neural control may depend on limb dominance (Daligadu et al. 2013).

General Setup

This study was conducted over 3 separate days and consisted of two parts: part 1 assessed corticospinal excitability and part 2 assessed spinal excitability. Parts 1 and 2 were performed during arm cycling at different cadences and power outputs. All arm cycling was performed using an arm cycle ergometer (SCIFIT ergometer, model PRO2 Total Body, Tulsa, OK) with the forearms fixed in a pronated position and the pedals locked 180° out of phase (i.e., asynchronous cranking pattern; see Fig. 1A). With the use of this ergometer, power output could be held constant, while the cadence could be altered. Participants were seated in an upright position at a comfortable distance from the hand pedals to ensure there was no reaching or trunk variation during cycling. The height of the ergometer seat was adjusted so that participants’ shoulders were approximately the same height as the axis of rotation of the arm cranks. Participants wore wrist braces to limit the amount of wrist flexion and extension during cycling to reduce the influence of heteronymous reflex connections that exist between the wrist flexors and the biceps brachii (Manning and Bawa 2011).

Fig. 1.

A: example of the experimental setup. Black arrows label each of the stimulation techniques utilized and the location of EMG electrodes. Measurements were taken at the 6 o’clock (shown here) and 12 o’clock positions from the dominant arm. B: raw EMG trace for the biceps (black, solid trace) and triceps brachii (gray, solid trace) of a single participant while arm cycling at 60 rpm and 20% maximum aerobic power output (W_max). No stimulations were given in this example, and the black arrows denote the 12 and 6 o’clock positions, accordingly. TMS, transcranial magnetic stimulation; TMES, transmastoid electrical stimulation.

Responses were evoked at two positions during arm cycling: 6 and 12 o’clock, defined relative to a clock face. Similar to our previous arm cycling studies, 6 o’clock was specified as “bottom dead center” and 12 o’clock was specified as “top dead center” (Forman et al. 2014,2015,2016; Spence et al. 2016). These two positions were examined because they occur during mid-elbow flexion (6 o’clock) and extension (12 o’clock) during arm cycling (Fig. 1B), which was important given our interest in both the biceps and triceps brachii. Stimuli were triggered automatically when the right hand passed a magnetic sensor at one of the predetermined positions (6 and 12 o’clock). For a left-handed participant, stimulations at the 6 o’clock position occurred when the right hand passed the sensor at the 12 o’clock position, while stimulations at the 12 o’clock position occurred when the right hand passed the sensor at the 6 o’clock position. The movement from 3 o’clock (full elbow extension) to 9 o’clock (full elbow flexion) was defined as elbow flexion, and the movement from 9 to 3 o’clock was defined as elbow extension.

The study required participants to cycle at combinations of two different cadences and three relative power outputs. Measurements were taken separately at 6 and 12 o’clock for a total of 12 trials. The order of the two positions as well as the TMS and TMES testing days were randomized across all participants.

Electromyography Recordings

EMG signals were recorded from the biceps and triceps brachii of the dominant arm using pairs of disposable Ag-AgCl surface electrodes (MediTrace 130 Foam Electrodes with conductive adhesive hydrogel; Covidien IIC). With the use of a bipolar configuration, electrodes were positioned ~2 cm apart (center to center) over the midline of the biceps brachii and on the lateral head of the triceps brachii. A ground electrode was positioned on the lateral epicondyle of the dominant arm. Preceding electrode placement, the skin was thoroughly prepared by removal of the hair (via a handheld razor) and dead epithelial cells (via abrasive paper), followed by sanitization using isopropyl alcohol swabs. This was done to reduce the impedance for EMG recordings. The EMG was sampled at a rate of 5 KHz using CED 1401 interface and the associated Signal 5 (Cambridge Electronic Desi, Cambridge, UK) software. All signals were amplified (×300) and bandpass filtered using a 3-Pole Butterworth filter with cut-off frequencies ranging from 10 to 1,000 Hz.

Stimulation Conditions

Evoked potentials were elicited via 1) electrical stimulation at Erb’s point, 2) TMS, and 3) TMES. All participants had prior experience with these stimulation procedures before participating. Initially, stimulation intensities were determined with participants seated comfortably in the chair of the SCIFIT ergometer with their hands in their lap. This position was defined as “rest.” Stimulation intensities were set at rest to observe a measurable response during elbow extension in the biceps brachii, as this muscle is approximately quiescent during elbow extension (Forman et al. 2015).

Brachial Plexus Stimulation

For parts 1 and 2, electrical stimulation of the brachial plexus at Erb’s point was delivered to elicit compound muscle action potentials (M waves). The cathode was placed in the supraclavicular fossa, and the anode was placed on the acromion process (see Fig. 1A). A constant current stimulator (model DS7AH; Digitimer, Welwyn Garden City, Hertfordshire, UK) with a pulse width of 200 μs was used for all participants. The stimulator intensity was initially set at 25 mA and was gradually increased until the size of the M-wave plateaued (i.e., maximal M wave (M_max)]. At this point, the stimulation intensity was then increased by 20% to ensure that M_max was elicited throughout the study since the M_max can change throughout the course of an experiment (Crone et al. 1999). The stimulation intensity used to elicit M_max was 186 ± 55.8 mA (means ± SD) and ranged from 100 to 300 mA.

Transcranial Magnetic Stimulation

TMS to the motor cortex was applied using a Magstim 200 stimulator (Magstim, Whitland, Dyfed, UK). A circular coil (13.5 cm outside diameter) was positioned over the vertex of each participant’s skull, with the direction of current flow in the coil preferentially activating the left or right motor cortex, depending on hand dominance. The vertex was determined by measuring the midpoint between the participant’s nasion and inion and the midpoint between the participant’s tragi (Forman et al. 2014, 2015; Pearcey et al. 2014; Taylor et al. 1997). The location at which these two points intersected was marked with a marker and defined as the vertex. The coil was held tangentially and firmly against the participant’s skull by an investigator who ensured careful and consistent placement of the coil over the vertex throughout the experiment. Stimulation intensity started at ~30% maximal stimulator output (MSO) and was increased gradually until resting motor threshold (RMT) was found. RMT was defined as a clearly discernable MEP in the biceps brachii with an amplitude ≥50 μV in four out of eight trials. Once RMT was determined, the %MSO was increased by 20% (i.e., 120% RMT), and an average of eight MEPs were recorded at this new intensity in the rest position. The 20% increase in stimulator output was performed to ensure that MEPs could be measured during both flexion and extension phases of arm cycling, as was done by Forman et al. (2015). This percent intensity of MSO was then used for the remainder of the experiment (Forman et al. 2015).

Transmastoid Electrical Stimulation

TMES was delivered using Ag-AgCl surface electrodes placed slightly inferior to the mastoid processes (Taylor 2006; Taylor and Gandevia 2004). A second Digitimer constant current stimulator (model DS7AH; Digitimer) with a pulse width of 200 μs was used to pass the current between the electrodes. Stimulation intensity began at 25 mA and was gradually increased until a clearly discernable CMEP was found. This was defined as CMEP threshold (CMEP_Thres). This intensity was then increased by 20% to ensure that a CMEP could be recorded and could either increase or decrease in amplitude to avoid potential “flooring” or ceiling effects of the CMEP (Taylor and Gandevia 2004; Weavil et al. 2015). This new intensity was then used for the remainder of the experiment. The stimulation intensity that was used to evoke CMEPs throughout the experiment was 137 ± 37.0 (means ± SD) and ranged from 75 to 198 mA. The latency of the CMEP was monitored to ensure that the corticospinal tract and not the ventral root was being stimulated (Taylor 2006). Latencies were visually monitored by recording the time between two vertical cursors, one of which placed at the stimulation artifact and the other placed at the initial deflection of the CMEP. Latencies were ~4 ms longer than the M_max latencies.

Experimental Protocol

On the initial visit, participants were familiarized with the various stimulation techniques and then performed an incremental aerobic arm cycling test (described below in Incremental Test). Part 1 (TMS day) was conducted on a separate day and required participants to cycle at two different cadences (60 and 90 rpm) at three different power outputs (20, 40, and 60% of their maximum power output determined from the Incremental Test). Part 2 (TMES day) was also conducted on a separate day and involved the same protocol as part 1 except participants received TMES instead of TMS. Parts 1 and 2 were randomized as was the order of the different conditions (cadences and power outputs) was randomized for each participant within each part of the study.

Incremental Test

Each participant completed a continuous, incremental arm cycling test on the SCIFIT ergometer to determine their maximum aerobic power output (W_max; group means ± SE: 126 ± 4.4 W). The initial work rate was set at 40 W and increased by 20 W every 2 min (Price et al. 2007). Participants were asked to maintain a constant cadence of 60 rpm throughout the entire incremental test. The W_max was considered to be the power recorded at the last completed stage. All participants exercised until volitional exhaustion or until the cadence dropped below 60 rpm for a period of 5 s. Following the test, participants completed a 2-min self-selected pace cool-down at 25 W. This test was completed to set the power outputs used during parts 1 and 2 of the study relative to each participant’s W_max. The relative power outputs at 20, 40, and 60% W_max were 25.2 ± 3.54, 50.5 ± 7.08, and 75.7 ± 10.62 W, respectively.

Part 1: Corticospinal Excitability During Arm Cycling at Different Cadences and Power Outputs

Once the stimulation intensities for brachial plexus stimulation and TMS were determined, the 12 different trials were performed (20, 40, and 60% W_max at 60 and 90 rpm at the 6 and 12 o’clock position) in a randomized order. The arm cycle ergometer was set at the predetermined power output, and participants were required to maintain the specified cadence. While cycling, a trial consisting of 10 MEP frames, 4 blank frames, and 2 M_max frames was completed at 1 of the 2 predetermined positions. Blank frames were given to reduce the effect of participants anticipating the stimulations. The order of the stimulations throughout each trial was randomized and were separated by ~7–8 s. The total length of each cycling trial was ~2.5 min. To reduce the potential influence of fatigue, 30-s rest periods were given half-way through the 20 and 40% W_max trials, and a 60-s rest period was given half-way through the 60% W_max trials. Additionally, following the completion of all six trials at one position, a 5-min rest period was provided. Following this rest period, these steps were repeated for the remaining position.

Part 2: Spinal Excitability During Arm Cycling at Different Cadences and Power Outputs

Part 2 of the study was conducted the exact same way as part 1 with the exception that participants (n = 10) now received TMES instead of TMS while cycling. These trials consisted of eight CMEP frames, two blank frames, and two M_max frames, which were randomized for each trial, as in part 1. Fewer TMESs were given than TMS since TMES is transiently painful. Rest periods were given in the same manner as part 1, and trials were completed in a randomized order. By using TMES in conjunction with TMS and by assessing the patterns of change in both MEPs and CMEPs, it allowed us to deduce differences in overall corticospinal excitability between supraspinal and spinal factors.

Data Analysis

All data were stored and analyzed offline using Signal 5.08 data collection software (CED). The averaged peak-to-peak amplitudes of MEPs, CMEPs, and M_max were measured from the biceps and triceps brachii of the dominant arm, from the initial deflection of the voltage trace to the return of the trace to baseline background EMG levels. Additionally, latencies of all evoked responses were examined and monitored. To account for changes in peripheral neuromuscular propagation during exercise (Taylor, 2006), averaged MEP and CMEP amplitudes were normalized to the averaged M_max evoked during the same trial. Prestimulus EMG, defined as the mean rectified EMG immediately before the stimulation artifact, was measured from the rectified virtual channel created for each muscle. The window used for mean EMG calculation was determined by the cadence at which the participant was cycling. For trials at 60 rpm, mean EMG was calculated by taking the mean of a 50-ms window, while at 90 rpm, it was calculated by taking the mean of a 33.3-ms window. The timeframes were chosen to represent 5% of one complete revolution (Forman et al. 2015).

Statistical Analysis

All statistics were performed using IBM’s SPSS Statistics version 23. Assumptions of sphericity were tested using Mauchly's test, and if violated, the appropriate correction was made to the degrees of freedom (Field 2013; Park et al. 2009). To assess whether there were statistical differences in MEP and CMEP amplitudes (normalized to M_max), and average prestimulus EMG between the two cadences and three power outputs, separate two-way repeated-measures ANOVAs (cadence x power output) were used for each position. Where significant results were found, repeated pairwise comparisons using Sidak post hoc tests were used to determine where the significance existed within conditions. All statistics were performed on group data, and a significance level of P < 0.05 was used.

RESULTS

For the following section, it is important to distinguish between periods of high activity and periods of much less activity for both the biceps and triceps brachii during arm cycling. At the 6 o’clock position, the biceps brachii is most active and is functioning as an agonist for elbow flexion, while the triceps is active but not the most active it becomes during cycling. The triceps is most active at the 12 o’clock position, when it is acting as an agonist to produce elbow extension and the biceps is much less active (almost quiescent) at this position (refer to Fig. 1). All data are reported in text as means ± SD and are illustrated in figures as means ± SE.

Biceps Brachii

Corticospinal excitability to the biceps brachii during elbow flexion.

mep amplitude.

Figure 2, A and B, shows representative and grouped data for MEP amplitudes from the biceps brachii at the 6 o’clock position, respectively. There were significant main effects for cadence [F_(1,12) = 16.78, P = 0.001] and power output [F_{(1.156,13.869)} = 28.13, P < 0.001], as well as the interaction between cadence and power output [F_(2,24) = 5.20, P = 0.013]. Pairwise comparisons revealed that MEPs increased by ~11.5% with cadence (90 > 60 rpm, P = 0.001) and up to ~57.2% with power output (60 > 40 > 20% W_max, P < 0.05 for all comparisons). MEPs increased at 20% (P = 0.004) and 40% (P = 0.003) W_max with increased cadence (90 > 60 rpm). There was no significant change in MEP amplitudes with increased cadence at 60% W_max (P = 0.220).

Fig. 2.

A: representative example for biceps brachii motor-evoked potentials (MEPs) at the 6 o’clock position (n = 1). In this example, MEP amplitudes at 60 rpm were 37.7, 58.1, and 71.6% maximal M wave (M_max) during the 20, 40, and 60% maximum aerobic power output (W_max) trials, respectively. At 90 rpm, MEP amplitudes were 68.7, 69.4, and 77.8% M_max during the 20, 40, and 60% W_max trials, respectively. B: group data (means ± SE; n = 13) for biceps brachii MEP amplitudes at the 6 o’clock position. C: average prestimulus EMG before transcranial magnetic stimulation. *P < 0.05, significant difference for power output; †P < 0.05, significant difference for cadence.

prestimulus emg for meps.

As a group, prestimulus EMG at 60 rpm at the 6 o’clock position was 56.37, 108.59, and 152.84 μV at 20, 40, and 60% W_max, respectively, while at 90 rpm prestimulus EMG was 86.82, 164.58, and 206.59 μV at 20, 40, and 60% W_max, respectively (Fig. 2C). There were significant main effects for cadence [F_(1,12) = 15.27, P = 0.002] and power output [F_(2,24) = 51.70, P < 0.001]; however, no significant interaction effect was observed [F_(2,24) = 1.69, P = 0.205]. Pairwise comparisons indicated that prestimulus EMG before MEPs increased on average ~44.1% as cadence (90 > 60 rpm, P = 0.002) increased and up to ~151.0% as power output (60 > 40 > 20% W_max, P < 0.05 for all comparisons) increased.

Spinal excitability to the biceps brachii during elbow flexion.

cmep amplitude.

Figure 3, A and B, shows representative and grouped data for CMEP amplitudes from the biceps brachii at the 6 o’clock position, respectively. There were significant main effects for cadence [F_(1,9) = 6.92, P = 0.027] and power output [F_{(1.111,10.001)} = 24.40, P < 0.001], but there was no interaction between cadence and power output [F_(2,18) = 1.63, P = 0.224]. Pairwise comparisons indicated that CMEP amplitudes increased by ~18.3% as cadence increased (90 > 60 rpm, P = 0.027) and increased by up to ~63.6% as power output (60 > 40 > 20% W_max, P < 0.05 for all comparisons) increased.

Fig. 3.

A: representative example for biceps brachii cervicomedullary motor-evoked potentials (CMEPs) at the 6 o’clock position (n = 1). In this example, CMEP amplitudes at 60 rpm were 31.7, 35.6, and 40.4% maximal M wave (M_max) during the 20, 40, and 60% maximum aerobic power output (W_max) trials, respectively. At 90 rpm, CMEP amplitudes were 43.8, 45.9, and 53.2% M_max during the 20, 40, and 60% W_max trials, respectively. B: group data (means ± SE; n = 10) for CMEP amplitudes at the 6 o’clock position. C: average prestimulus EMG before transmastoid electrical stimulation. *P < 0.05, significant difference for power output; †P < 0.05, significant difference for cadence.

prestimulus emg for cmeps.

As a group, prestimulus EMG for the 60 rpm condition was 60.28, 85.73, and 125.28 μV at 20, 40, and 60% W_max, respectively, while at 90 rpm prestimulus EMG was 76.52, 128.19, and 181.48 μV at 20, 40, and 60% W_max, respectively (Fig. 3C). Significant main effects for cadence [F_(1,9) = 7.67, P = 0.022] and power output [F_{(1.204,10.835)} = 29.30, P < 0.001] were observed. However, no significant interaction effects were observed [F_{(1.201,10.813)} = 3.07, P = 0.104]. Pairwise comparisons indicated that prestimulus EMG before CMEPs increased on average by ~42.4% as cadence (90 > 60 rpm, P = 0.022) increased and up to ~124.2% as power output (60 > 40 > 20% W_max, P < 0.05 for all comparisons) increased.

Corticospinal excitability to the biceps brachii during elbow extension.

mep amplitude.

Figure 4, A and B, shows representative and grouped data for MEP amplitudes from the biceps brachii at the 12 o’clock position, respectively. There were significant main effects for cadence [F_(1,12) = 4.92, P = 0.047] and power output [F_{(1.271,15.247)} = 12.65, P = 0.002], but no significant interaction effect existed between cadence and power output [F_(2,24) = 0.56, P = 0.580). Pairwise comparisons revealed that MEP amplitudes increased by ~26.8% as cadence (90 rpm > 60 rpm, P = 0.047] increased and increased up to 96.1% as power output (60 > 40 > 20% W_max, P < 0.05 for all comparisons) increased.

Fig. 4.

A: representative example for biceps brachii motor-evoked potentials (MEPs) at the 12 o’clock position (n = 1). In this example, MEP amplitudes at 60 rpm were 7.5, 7.8, and 7.7% maximal M wave (M_max) during the 20, 40, and 60% maximum aerobic power output (W_max) trials, respectively. At 90 rpm, MEP amplitudes were 6.9, 10.3, and 14.2% M_max during the 20, 40, and 60% W_max trials, respectively. B: group data (means ± SE; n = 13) for biceps brachii MEP amplitudes at the 12 o’clock position. C: average prestimulus EMG before transcranial magnetic stimulation. *P < 0.05, significant difference for power output; †P < 0.05, significant difference for cadence.

prestimulus emg for meps.

As a group, prestimulus EMG at 60 rpm was 33.02, 35.57, and 36.70 μV at 20, 40, and 60% W_max, respectively, while prestimulus EMG at 90 rpm was 38.95, 36.55, and 38.85 μV at 20, 40, and 60% W_max, respectively (Fig. 4C). There were no significant main effects for cadence [F_(1,12) = 1.76, P = 0.209], power output [F_{(1.220,14.645)} = 1.48, P = 0.250], or interaction effects [F_{(1.098,13.182)} = 0.98, P = 0.390] observed.

Spinal excitability to the biceps brachii during elbow extension.

cmep amplitude.

Figure 5, A and B, shows representative and grouped data for CMEP amplitudes from the biceps brachii at the 12 o’clock position, respectively. There was a significant main effect for cadence [F_(1,9) = 9.51, P = 0.013], but no significant effects were found for power output [F_(2,18) = 0.16, P = 0.851], or the interaction between cadence and power output [F_(2,18) = 0.16, P = 0.857]. Pairwise comparisons revealed that CMEP amplitudes decreased by ~29.7% as cadence increased (60 > 90 rpm, P = 0.013).

Fig. 5.

A: representative example for biceps brachii cervicomedullary motor-evoked potentials (CMEPs) at the 12 o’clock position (n = 1). In this example, CMEP amplitudes at 60 rpm were 2.1, 4.5, and 5.4% maximal M wave (M_max) during the 20, 40, and 60% maximum aerobic power output (W_max) trials, respectively, while at 90 rpm, CMEP amplitudes were 2.1%, 2.9%, and 2.1% M_max during the 20, 40, and 60% W_max trials, respectively. B: group data (means ± SE; n = 10) for biceps brachii CMEP amplitudes at 12 o’clock. C: average prestimulus EMG before transmastoid electrical stimulation. *P < 0.05, significant difference for power output; †P < 0.05, significant difference for cadence.

prestimulus emg for cmeps.

As a group, prestimulus EMG at 60 rpm was 31.04, 33.21, and 35.96 μV at 20, 40, and 60% W_max, respectively, while prestimulus EMG at 90 rpm was 31.36, 31.56, and 37.70 μV at 20, 40, and 60% W_max, respectively (Fig. 5C). There was no significant main effect for cadence [F_(1,9) = 0.011, P = 0.919], but there was a significant main effect for power output [F_(2,18) = 6.57, P = 0.007]. There was no interaction effect between cadence and power output [F_(2,18) = 1.62, P = 0.226]. Pairwise comparisons revealed that prestimulus EMG increased by ~13.7% only when power output increased from 40 to 60% W_max (P = 0.008).

Triceps Brachii

Corticospinal excitability to the triceps brachii during elbow flexion.

mep amplitude.

Figure 6, A and B, shows representative and grouped data for MEP amplitudes from the triceps brachii at the 6 o’clock position, respectively. There was a significant main effect of cadence [F_(1,12) = 8.43, P = 0.013] but no effect of power output [F_(2,24) = 2.81, P = 0.080] and no significant interaction effect [F_(2,24) = 1.18, P = 0.325]. Pairwise comparisons revealed that MEPs at 90 rpm were ~15.2% higher than MEPs at 60 rpm (P = 0.013).

Fig. 6.

A: representative example for triceps brachii motor-evoked potentials (MEPs) at the 6 o’clock position (n = 1). In this example, MEP amplitudes at 60 rpm were 14.2, 18.4, and 19.5% maximal M wave (M_max) during the 20, 40, and 60% maximum aerobic power output (W_max) trials, respectively. At 90 rpm, MEP amplitudes were 20.4, 18.5, and 22.9% M_max during the 20, 40, and 60% W_max trials, respectively. B: group data (means ± SE; n = 13) for triceps brachii MEP amplitudes at the 6 o’clock position. C: average prestimulus EMG before transcranial magnetic stimulation. *P < 0.05, significant difference for power output; †P < 0.05, significant difference for cadence.

prestimulus emg for meps.

As a group, prestimulus EMG was 24.60, 35.66, and 41.04 μV at 20, 40, and 60% W_max, respectively, for the 60 rpm condition and 30.14, 47.83, and 52.95 μV at 20, 40, and 60% W_max, respectively, for the 90 rpm condition (Fig. 6C). There were significant main effects for cadence [F_(1,12) = 14.37, P = 0.003] and power output [F_(2,24) = 32.92, P < 0.001], but there was no significant interaction effect [F_(2,24) = 1.72, P = 0.204]. Pairwise comparisons revealed that prestimulus EMG at 90 rpm was ~29.2% higher than prestimulus EMG at 60 rpm (P = 0.003) and that prestimulus EMG at 20% W_max was ~62.1% smaller than prestimulus EMG at 40% W_max (P < 0.001) and 60% W_max (P < 0.001), but 40% W_max was not different from 60% W_max (P = 0.176).

Spinal excitability to the triceps brachii during elbow flexion.

cmep amplitude.

Figure 7, A and B, shows representative and grouped data for CMEP amplitudes from the triceps brachii at the 6 o’clock position, respectively. There were significant main effects for cadence [F_(1,9) = 9.73, P = 0.012] and power output [F_(2,18) = 6.03, P = 0.010], but there was no significant interaction effect [F_(2,18) = 1.85, P = 0.194]. Pairwise comparisons revealed that CMEP amplitudes increased by ~16.3% as cadence increased (90 > 60 rpm, P = 0.012) and increased by ~19.1% only between the largest and the smallest power output (60 > 20% W_max, P = 0.038).

Fig. 7.

A: representative example for triceps brachii cervicomedullary motor-evoked potentials (CMEPs) at the 6 o’clock position (n = 1). In this example, CMEP amplitudes at 60 rpm were 18.3, 19.9, and 24.5% M_max during the 20, 40, and 60% maximum aerobic power output (W_max) trials, respectively, while at 90 rpm, CMEP amplitudes were 28.4, 32.5, and 34.8% M_max during the 20, 40, and 60% W_max trials, respectively. B: group data (means ± SE; n = 10) for triceps brachii CMEP amplitudes at 6 o’clock. C: average prestimulus EMG before transmastoid electrical stimulation. *P < 0.05, significant difference for power output; †P < 0.05, significant difference for cadence.

prestimulus emg for cmeps.

As a group, prestimulus EMG was 22.46, 29.87, and 36.23 μV at 20, 40, and 60% W_max, respectively, for the 60 rpm condition and 26.53, 39.11, and 44.10 μV at 20, 40, and 60% W_max, respectively, for the 90 rpm condition (Fig. 7C). There were significant main effects for cadence [F_(1,9) = 12.39, P = 0.008] and power output [F_(2,18) = 27.21, P < 0.001], but there was no significant interaction effect [F_(2,18) = 1.28, P = 0.315]. Pairwise comparisons revealed that prestimulus EMG at 90 rpm was ~23.9% larger than prestimulus EMG at 60 rpm (P = 0.018) and that with increased power output prestimulus EMG increased up to ~64.0% (60 > 40 > 20% W_max, P < 0.05 for all comparisons).

Corticospinal excitability to the triceps brachii during elbow extension.

mep amplitude.

Figure 8, A and B, shows representative and grouped data for MEP amplitudes from the triceps brachii at the 12 o’clock position, respectively. There was no significant main effect for cadence [F_(1,12) = 0.410, P = 0.534], but there was a significant main effect for power output [F_(2,24) = 4.79, P = 0.030]. Additionally, there was no significant interaction effect [F_(2,24) = 0.20, P = 0.993]. Despite the significant main effect of power output, pairwise comparisons revealed that no significant differences between MEPs existed among the power outputs (P > 0.05 for all comparisons).

Fig. 8.

A: representative example for triceps brachii motor-evoked potentials (MEPs) at the 12 o’clock position (n = 1). In this example, MEP amplitudes at 60 rpm were 26.5, 27.2, and 40.3% maximal M wave (M_max) during the 20, 40, and 60% W_max trials, respectively. At 90 rpm, MEP amplitudes were 35.2, 41.9, and 40.9% M_max during the 20, 40, and 60% W_max trials, respectively. B: group data (means ± SE; n = 13) for triceps brachii MEP amplitudes at the 12 o’clock position. C: average prestimulus EMG before transcranial magnetic stimulation. *P < 0.05, significant difference for power output.

prestimulus emg for meps.

As a group, prestimulus EMG during elbow extension at 60 rpm was 48.00, 96.66, and 139.16 μV at 20, 40, and 60% W_max, respectively, while prestimulus EMG at 90 rpm was 54.16, 87.83, and 125.08 μV at 20, 40, and 60% W_max, respectively (Fig. 8C). There was no significant main effect for cadence [F_(1,12) = 0.26, P = 0.621], but there was a significant main effect for power output [F_{(1.095,13.145)} = 11.92 P = 0.004]. Additionally, there was no significant interaction effect [F_{(1.269,13.689)} = 1.12, P = 0.328]. Pairwise comparisons revealed that prestimulus EMG increased up to ~158.7% as power output increased (60 > 40 > 20%, all comparisons P < 0.05).

Spinal excitability to the triceps brachii during elbow extension.

cmep amplitude.

Figure 9, A and B, shows representative and grouped data for CMEP amplitudes from the triceps brachii at the 12 o’clock position, respectively. There were significant main effects for cadence [F_(1,9) = 5.30, P = 0.047] and power output [F_(2,18) = 5.89, P = 0.011], but there was no significant interaction effect [F_(2,18) = 1.85, P = 0.194]. Pairwise comparisons revealed that CMEPs at 90 rpm were ~12.8% larger than CMEPs at 60 rpm (P = 0.047), and that CMEPs were ~13.5% greater at 40% W_max than at 20% W_max (P = 0.032).

Fig. 9.

A: representative example for triceps brachii cervicomedullary motor-evoked potentials (CMEPs) at the 12 o’clock position (n = 1). In this example, CMEP amplitudes at 60 rpm were 21.5, 21.7, and 19.8% maximal M wave (M_max) during the 20, 40, and 60% maximum aerobic power output (W_max) trials, respectively, while at 90 rpm, CMEP amplitudes were 18.9, 19.6, and 22.6% M_max during the 20, 40, and 60% W_max trials, respectively. B: group data (means ± SE; n = 10) for triceps brachii CMEP amplitudes at 12 o’clock. C: average prestimulus EMG before transmastoid electrical stimulation. *P < 0.05, significant difference for power output; †P < 0.05, significant difference for cadence.

prestimulus emg for cmeps.

As a group, prestimulus EMG at 60 rpm was 49.48, 100.14, and 148.37 μV at 20, 40, and 60% W_max, respectively, while prestimulus EMG at 90 rpm was 61.42, 107.47, and 156.87 μV at 20, 40, and 60% W_max, respectively (Fig. 9C). There was no significant main effect for cadence [F_{(1, 9)} = 0.830, P = 0.389], but there was a significant main effect for power output [F_{(1.084,8.672)} = 10.45, P = 0.010]. Additionally, there was no significant interaction effect [F_(2,18) = 0.16, P = 0.985]. Pairwise comparisons revealed that prestimulus EMG increased up to ~175.2% as power output increased (60 > 40 > 20% W_max, P < 0.05 for all comparisons).

DISCUSSION

In the present study we demonstrate that supraspinal and spinal excitability are differentially modulated depending on the manner in which the intensity of arm cycling is altered (power output or cadence) and that these effects are both muscle and phase dependent. Thus, in general, the findings did confirm our initial hypotheses with one notable exception with respect to the less active phase of the triceps brachii (see below). Perhaps the most interesting findings were with respect to muscle-dependent differences in corticospinal excitability during the less active phase of each muscle examined (12 o’clock for biceps brachii and 6 o’clock for triceps brachii; see Fig. 1B). Increasing power output did not significantly change spinal motoneuron excitability whereas an increased cadence resulted in a decrease in spinal motoneuron excitability to the biceps brachii but an increase in spinal motoneuron excitability to the triceps brachii. Furthermore, the increase in power output during the less active phase of the biceps brachii did not counteract the decrease in spinal motoneuron excitability that occurred with increased cadence.

Muscle-Dependent Changes in Corticospinal Excitability During the Less Active Phases

During the less active phase of the biceps brachii, corticospinal excitability increased with increased cycling intensity (i.e., power output and cadence; see Fig. 4B), while motoneuron excitability did not change with power output and decreased with increased cadence (see Fig. 5B). This suggests that an increase in supraspinal excitability occurred in both cases. The decrease in motoneuron excitability observed with the higher cadence was not offset by higher power outputs. Moreover, these changes in supraspinal and spinal excitability occurred with minimal changes in background EMG activity. In contrast, overall corticospinal excitability to the triceps brachii did not significantly change with power output (though the pattern of change suggests an increase) and increased with cadence (Fig. 6), while the pattern of spinal excitability increased with both power output and cadence (Fig. 7). Thus, with respect to the less active phases of each muscle, the data suggest that supraspinal excitability is enhanced to the biceps brachii as intensity increases, particularly with cadence when spinal motoneuron excitability decreases, whereas enhanced spinal motoneuron excitability to the triceps brachii appears to drive overall increases in corticospinal excitability as power output and cadence are increased. This finding in the triceps brachii was in opposition to our second hypothesis, given that we postulated that supraspinal excitability would be mediating increases in overall corticospinal excitability within the less active phase of each muscle. These findings suggest specific muscle-dependent modulation in supraspinal and spinal excitability that is also dependent on the phase of the muscle.

These muscle-dependent differences in corticospinal excitability found here are adding to a growing body of literature on the topic during locomotor outputs in humans (Carroll et al. 2006; Sidhu et al. 2012; Spence et al. 2016; Weavil et al. 2015). For example, supraspinal excitability is greater in the biceps brachii during arm cycling compared with tonic contraction (Forman et al. 2014) whereas the opposite is true of the flexor carpi radialis (Carroll et al. 2006), likely resulting from differences in the role of each muscle during arm cycling. Recently, we demonstrated a lack of phase dependence (elbow flexion vs extension) in corticospinal excitability to the triceps brachii and higher spinal motoneuron excitability during elbow flexion (less active phase), which was opposite to that in the biceps brachii (strong phase dependence with higher spinal excitability during flexion, the more active phase) (Spence et al. 2016). In the leg musculature, Schubert and colleagues (1997) demonstrated that corticospinal excitability was enhanced to a larger degree in the tibialis anterior than the gastrocnemius during locomotion with similar results demonstrated by Capaday and colleagues (1999). These collective findings raise the question: what factor(s) can account for these intermuscle differences(?). Prior to a discussion of the putative mechanisms it is important to recognize that the pattern of muscle activity is different between the biceps and triceps brachii. Most notably, the biceps brachii is almost quiescent during elbow extension whereas the triceps brachii is active during elbow flexion, albeit to a lesser degree than during extension (see Fig. 1B). The reason(s) behind triceps brachii activity during elbow flexion is currently unclear, but it may be due to maintaining elbow joint position and hand contact with the pedals as the hand pulls downwards during elbow flexion.

Putative Mechanisms for Muscle-Dependent Differences in Corticospinal Excitability During the Less Active Phases

There are multiple descending pathways that can influence, directly or indirectly, excitability along the corticospinal pathway during motor output. Interestingly, the rubrospinal and corticospinal tracts mainly excite flexors and inhibit extensors while the vestibulospinal tract provides strong excitation to extensors during locomotion in the cat (Orlovsky 1972). This is in agreement with findings in humans, namely greater supraspinal input to flexors than extensors (Dietz et al. 1992) and may play a role in the findings presented herein as a function of intensity, specifically the effect of cadence on supraspinal excitability during the inactive phase for the biceps brachii. Higher supraspinal control of flexors is indirectly supported by findings of a greater amount of cortico-motoneuronal monosynaptic connections for flexor motoneuron pools compared with extensors (Brouwer and Ashby 1990). Thus higher cortical control of the biceps brachii compared with the triceps brachii may be a contributing factor in the current study. In addition, the increase in supraspinal excitability to the biceps brachii as cadence increased may be related to cortical spread from homologous and heterologous cortical connections, keeping in mind that as the ipsilateral arm is extending (arm from which recordings were made) the contralateral arm is flexing.

Provided that the triceps brachii are under less supraspinal control than the biceps brachii, what mechanisms may account for the increase in spinal excitability during its’ inactive phase (i.e., elbow flexion)? First, spinal motoneurons possess persistent inward currents that amplify synaptic input and have a greater effect in extensor motoneurons (Cotel et al. 2009; Hounsgaard et al. 1988), a finding recently confirmed in the triceps as compared with the biceps brachii in humans, albeit during tonic contractions (Wilson et al. 2015). When the elbow is flexed at the 6 o’clock position, the triceps brachii is stretched, thereby increasing input from Ia afferents to the motoneuron pool. This afferent information could be increased with cadence as the rate of stretch of Ia afferents would be increased, thus enhancing persistent inward current activation and amplifying synaptic input to the motoneuron pool evident as an increase in TMES-evoked CMEP amplitudes (see Fig. 7). A second possibility is that the lateral head of the triceps brachii may not represent the activity or excitability to the other three elbow extensors (i.e., long and medial head of the triceps and anconeus). Previous work has shown that the lateral head (muscle from which recordings were made) has a lower recruitment threshold than the long head at a shoulder and elbow position similar to the 6 o’clock position (Davidson and Rice 2010) and that increased movement speed lowers recruitment thresholds (Desmedt and Godaux 1977). Finally, during elbow flexion the activity of the triceps brachii increases as the cadence and/or power output increases (see background EMG in Figs. 7C and 8C), likely activating tension-sensitive Ib afferents located at the musculotendinous unit (Jami 1992). During locomotion, the inhibitory action of Ib afferents onto extensor motoneurons during stance is replaced with excitation, a process referred to as a reflex reversal (Conway et al. 1987; McCrea et al. 1995). Altough a reflex reversal has not been confirmed during human locomotion, Ib facilitation has been demonstrated (Faist et al. 2006). This increase in excitatory synaptic input to the motoneuron pool during the flexion phase of arm cycling as cycling intensity increased would thus contribute to enhanced CMEP amplitudes in the triceps brachii during arm cycling (see Fig. 7B). Additionally, during elbow flexion the triceps brachii is acting as an antagonist and as such would be under less tension than when the muscle is acting as the agonist (i.e., elbow extension). Thus perhaps there is reduced Ib activity in the triceps brachii during elbow flexion and enhanced Ib inhibitory feedback during elbow extension, which could potentially explain the increased CMEP at the 6 o’clock position compared with 12 o’clock. This would not occur in the biceps brachii during the less active phase because there is little to no motoneuron output, suggesting little tension is produced by the muscle, and thus Ib afferents would likely not be not activated, at least not to the same degree as in the triceps brachii.

Additional differences in the processing of afferent feedback to supraspinal and/or spinal centers may also underlie the observed differences in spinal excitability with differing cadence and power outputs. During leg cycling, soleus H reflexes are significantly suppressed with increases in cadence, indicative of increased presynaptic inhibition, while increases in cycling load have been shown to increase H-reflex amplitudes (Pyndt et al. 2003). In the present study, it is possible that differences in presynaptic inhibition to the spinal motoneuron and/or supraspinal centers may help to explain the difference in the TMES-evoked CMEPs (i.e., suppression of CMEP with cadence versus no change in CMEP with power output). Perhaps as cadence increases, presynaptic inhibition of group Ia afferents to the biceps brachii motoneuron pool at the 12 o’clock position is enhanced more than when cycling at an increased power output, therefore explaining the reduction in CMEP amplitude with increased cadence but no effect of power output.

Intensity-Dependent Changes in Corticospinal and Spinal Excitability During the Active Phase

During the most active phase of the biceps brachii (i.e., 6 o’clock position; see Fig. 1B), supraspinal (Fig. 2B) and spinal excitability (Fig. 3B) increased with increases in cycling cadence and power output, a finding that we have previously shown in two separate reports (Forman et al. 2015; Spence et al. 2016). Similar findings occurred for the triceps brachii during its more active phase (i.e., 12 o’clock; see Figs. 1B, 8, and 9). In our prior work related to the biceps brachii, we suggested that the increase in MEP and CMEP amplitudes reflected an increase in the descending motor drive to the spinal cord necessary to increase the recruitment and/or firing frequency of spinal motoneurons during tonic contraction, thus providing adequate muscle activation for movement (Pearcey et al. 2014). In the present study, the same rationale can be applied to both muscles during their respective active phases such that larger descending drive is needed to increase recruitment and firing frequency of motoneurons of the biceps and triceps in their respective most active phases.

Methodological Considerations

Several factors must be taken into consideration for the interpretation of the current results. First, MEP and CMEP amplitudes were not matched between days, and MEP amplitudes from the biceps brachii were approximately twice as large as CMEP amplitudes at each cycling intensity during elbow flexion (Figs. 2B and 3B). The discrepancy in the amplitudes represents the activation of different portions of the motoneuron pool, which would be an important limitation to the observed results. To address this issue, we collected four more participants (n = 4; data not shown) and set the stimulation intensities used to elicit MEPs, and CMEPs (i.e., 15–20% of M_max) during the flexion phase of arm cycling at 60 rpm and 25 W. Interestingly, we found that MEP amplitudes were still ~50% larger than CMEPs at each cycling intensity during elbow flexion despite having initially matched MEP and CMEP amplitudes. Thus the pattern of modulation in MEP and CMEP amplitudes does not appear to be dependent on the way in which the stimulations were set (i.e., at ‘rest’ vs. during cycling). Instead it suggests that TMS-evoked MEPs may be more sensitive (i.e., larger increase in excitability) than TMES-evoked CMEPs as the intensity of arm cycling increases as we have previously shown during tonic contractions (Pearcey et al. 2014; Philpott et al. 2015).

Second, at the highest power output (60% W_max), increases in cadence did not result in a significant increase in biceps brachii MEP amplitudes at the 6 o’clock position (i.e., mid-elbow flexion; Fig. 2B), despite large increases in voluntary EMG activity (Fig. 2C). Although we cannot be certain, this lack of increase in MEP amplitude as cadence increases from 60 to 90 rpm at 60% W_max may indicate a potential ‘plateau effect’ in MEP amplitudes similar to that which has been reported in isometric contractions (Martin et al. 2006; Pearcey et al. 2014) and more recently during leg cycling (Weavil et al. 2015). This finding could be due to a disconnect between increased central motor drive (i.e., voluntary EMG) and MEP amplitudes, meaning that further increases in central motor drive will have no effect on corticospinal excitability. However, a simpler explanation could be that the stimulation intensity used to elicit MEPs was too high, therefore limiting the increase in MEP size. This explanation seems more realistic since CMEPs did not experience a similar plateau effect, despite showing a similar pattern of modulation (Fig. 3B). We cannot speculate if a plateau effect would be preserved at higher cycling intensities.

Conclusions

The present study demonstrates that changes in supraspinal and spinal excitability during rhythmic arm cycling are phase, muscle, and intensity dependent (i.e., cadence vs power output). It appears that supraspinal factors may contribute more to changes in overall corticospinal excitability to the biceps brachii whereas spinal factors play a larger role for the triceps brachii, especially during the respective less active phase of each muscle. These possibilities and how they are influenced by changes in cycling intensity require further investigation to elucidate potential mechanisms.

GRANTS

This study was supported by Canada Graduate Scholarships-Master's Natural Sciences and Engineering Research Council (NSERC) funding (to E. J. Lockyer) as well as NSERC Discovery Grant (to K. E. Power).

DISCLOSURES

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

AUTHOR CONTRIBUTIONS

E.J.L., D.C.B., and K.E.P conceived and designed research; E.J.L., R.J.B., A.P.H., L.R.A., and A.J.S. performed experiments; E.J.L., D.C.B., and K.E.P. interpreted results of experiments; E.J.L. analyzed data; E.J.L., and D.C.B. prepared figures; E.J.L., and K.E.P. drafted manuscript; E.J.L., R.J.B., A.P.H., L.R.A., A.J.S., and K.E.P. approved final version of manuscript.