Interhemispheric interactions between trunk muscle representations of the primary motor cortex

The mechanisms involved in the bilateral coordination of axial muscles during unilateral arm movement are poorly understood. We thus investigated the nature of interhemispheric interactions in axial muscles during arm motor tasks in healthy subjects. By combining different methodologies, we showed that trunk muscles receive both inhibitory and facilitatory cortical outputs during activation of arm muscles. We propose that inhibition may be conveyed mainly through interhemispheric mechanisms and facilitation by subcortical mechanisms or ipsilateral pathways.

Abstract

Unilateral arm movements require trunk stabilization through bilateral contraction of axial muscles. Interhemispheric interactions between primary motor cortices (M1) could enable such coordinated contractions, but these mechanisms are largely unknown. Using transcranial magnetic stimulation (TMS), we characterized interhemispheric interactions between M1 representations of the trunk-stabilizing muscles erector spinae at the first lumbar vertebra (ES L1) during a right isometric shoulder flexion. These interactions were compared with those of the anterior deltoid (AD), the main agonist in this task, and the first dorsal interosseous (FDI). TMS over the right M1 elicited ipsilateral silent periods (iSP) in all three muscles on the right side. In ES L1, but not in AD or FDI, ipsilateral motor evoked potential (iMEP) could precede the iSP or replace it. iMEP amplitude was not significantly different whether ES L1 was used to stabilize the trunk or was voluntarily contracted. TMS at the cervicomedullary junction showed that the size of cervicomedullary evoked potential was unchanged during the iSP but increased during iMEP, suggesting that the iSP, but not the iMEP, is due to intracortical mechanisms. Using a dual-coil paradigm with two coils over the left and right M1, interhemispheric inhibition could be evoked at interstimulus intervals of 6 ms in ES L1 and 8 ms in AD and FDI. Together, these results suggest that interhemispheric inhibition is dominant when axial muscles are involved in a stabilizing task. The ipsilateral facilitation could be evoked by ipsilateral or subcortical pathways and could be used depending on the role axial muscles play in the task.

NEW & NOTEWORTHY The mechanisms involved in the bilateral coordination of axial muscles during unilateral arm movement are poorly understood. We thus investigated the nature of interhemispheric interactions in axial muscles during arm motor tasks in healthy subjects. By combining different methodologies, we showed that trunk muscles receive both inhibitory and facilitatory cortical outputs during activation of arm muscles. We propose that inhibition may be conveyed mainly through interhemispheric mechanisms and facilitation by subcortical mechanisms or ipsilateral pathways.

axial muscles are mainly used to stabilize the trunk to perform goal-oriented tasks with the limb. Notably, during a shoulder flexion task, the axial muscles are activated to enable postural adjustments, whereas shoulder muscles are the agonists of the task (Hodges and Richardson 1997; Massion 1992; Zattara and Bouisset 1988). The cortical mechanisms underlying simultaneous activation of the agonist muscles and those responsible for postural adjustments during the task have been well researched in animal models (Schepens and Drew 2003) but are still not fully understood in humans. Notably, during reaching movements, axial muscles such as erector spinae (ES) are activated bilaterally before and during the arm movement to maintain trunk stability (Belen’kii et al. 1967; Davey et al. 2002; Lee et al. 1987). Furthermore corticospinal excitability is increased in ES muscles during dynamic shoulder flexion and, to a lesser extent, during static shoulder flexion (Chiou et al. 2016) and shoulder abduction (Davey et al. 2002; Kuppuswamy et al. 2008), suggesting a corticospinal involvement in maintaining trunk stability. This bilateral activation of axial muscles during arm movement could also be enabled by interactions between their respective cortical representations in the two hemispheres, yet no study has investigated this issue so far.

Interhemispheric connections link cortical areas through the corpus callosum (Jenny 1979; Rouiller et al. 1994; van der Knaap and van der Ham 2011) and have been well investigated by transcranial magnetic stimulation (TMS) studies in humans. Prominent interhemispheric inhibition (IHI) has been repeatedly demonstrated between motor areas of hand muscles (Ferbert et al. 1992; Salerno and Georgesco 1996). In axial and more proximal arm muscles, contribution of IHI is not well established and has mainly been studied during voluntary contraction while the targeted muscle was the agonist of the task in biceps and triceps (Harris-Love et al. 2007) and in serratus anterior and trapezius (Matthews et al. 2013). Because previous reports have suggested IHI could be dependent on the role of the muscle during a task (Harris-Love et al. 2007), it can be expected that axial muscles might exhibit a different pattern of IHI while stabilizing the trunk for an arm movement than when voluntarily contracted during a trunk extension task, for example.

These previous TMS studies on both axial and arm muscles underline the differential cortical control of these muscles that may be related to their respective function for the control of movements. Whereas distal arm muscles such as hand muscles receive inputs mainly from the contralateral hemisphere, proximal and axial muscles are innervated by inputs from both contralateral and ipsilateral corticospinal tracts (Brinkman and Kuypers 1973; Kuypers 1962; Palmer and Ashby 1992). TMS studies have also shown that ipsilateral outputs to proximal muscles are more powerful than those to distal muscles (Alexander et al. 2007; Bawa et al. 2004; Chen et al. 2003; Ziemann et al. 1999).

Hence, the goal of this study was to investigate interhemispheric interactions between axial muscle representations during an isometric shoulder flexion task. Our hypothesis was that interhemispheric interactions between M1 representations of axial muscle erector spinae at the level of the first lumbar vertebra (ES L1) are facilitatory, reflecting their stabilizing role. Furthermore, anterior deltoid (AD) was also assessed to determine interactions that are simultaneously taking place in the cortical representation of the agonist, which has not been studied before. We hypothesized that interhemispheric interactions will be inhibitory for AD, similarly to what has been shown for other upper limb muscles, such as the first dorsal interosseous (FDI), an intrinsic hand muscle that we also assessed in this study for comparison.

METHODS

Subjects

Twenty-five subjects (13 women, 12 men; age: 28.04 ± 6.7 yr, mean ± SD) participated in this study that spanned over 6 different sessions, each separated by at least 1 day. The number of participants in each experimental session is detailed in Table 1. Interhemispheric interactions were assessed using two different TMS paradigms: the ipsilateral cortical stimulation and dual-coil stimulation.

Table 1.

Number of subjects participating in each experimental session

	Ipsilateral Cortical Stimulation			Dual-Coil Stimulation
	Stimulation of ES L1/AD/FDI	ES L1 During Shoulder and Trunk Task	CMEP in ES L1	ES L1	AD	FDI
No. of subjects	12	8	10	10	8	7

The muscles erector spinae at the level of the L1 vertebrae (ES L1), anterior deltoid (AD), and first dorsal interosseous (FDI) were assessed with ipsilateral cortical stimulation and dual-coil stimulation paradigms. Cervicomedullary motor evoked potential (CMEP) were also assessed in ES L1.

In ipsilateral cortical stimulation, TMS was applied on the right cortex ipsilateral to the targeted arm. Twelve participants took part in a single 3-h session where all three muscles (AD, FDI, and ES L1) were tested (see Table 1). To determine whether the responses obtained in ES L1 were task dependent, 8 subjects took part in an additional 2-h session. In that session, TMS was applied on the right ipsilateral cortex to test responses in right ES L1 when the muscle was activated either as the agonist of the task (see Experimental Tasks, Trunk task) or as a trunk stabilizer during isometric contraction of the AD (see Experimental Tasks, Shoulder task). Two subjects participated in both the session where ipsilateral cortical stimulation was applied and the session where task dependency was assessed. Finally, to better understand the mechanisms involved in the responses obtained in ES L1 (i.e., cortical or subcortical), 10 subjects participated in a 2-h experiment where a cervicomedullary motor evoked potential (CMEP) was evoked alone (test CMEP) or conditioned by an ipsilateral cortical stimulation (conditioned CMEP). Three subjects participated in both the ipsilateral cortical stimulation and CMEP sessions.

In the second part of this study, a dual-coil paradigm was used in which a test stimulus (TS) was applied with a coil over the cortex contralateral to the targeted arm and a conditioning stimulus (CS) was applied with a coil over the cortex ipsilateral to the targeted arm at different interstimulus intervals (ISI). This paradigm between the representations of ES L1 was tested in 10 subjects (5 subjects took part in both ipsilateral cortical stimulation and dual-coil ES L1 sessions), between representations of AD in 8 subjects (6 subjects took part in both ipsilateral cortical stimulation and dual-coil AD sessions) and between representations of FDI in 7 subjects (3 subjects took part in both ipsilateral cortical stimulation and dual-coil FDI sessions). Dual-coil assessment of these 3 muscles was made in 3 different sessions (1 muscle/session) as each session lasted 3 h. Three subjects took part in the ipsilateral cortical stimulation session as well as the dual-coil sessions in ES L1, AD, and FDI.

Handedness was determined by the Edinburgh inventory (Oldfield 1971), and only right-handed subjects were included (mean Edinburgh inventory index = 98.5; range 80–100). Subjects had no contraindication to TMS, namely, epileptic history, pregnancy, cardiac pacemaker, and presence of surgical metallic fragment or clip in the cranium. The experimental protocol was approved by the local ethics research board of Centre for Interdisciplinary Research in Rehabilitation (CRIR) and was in accordance with the Declaration of Helsinki. Participants received oral and written information about the study and then gave their written consent.

Electromyographic Recordings

Electromyographic (EMG) activity was recorded bilaterally from ES L1, AD, and FDI using surface electrodes (Ag-AgCl, Blue sensor; interelectrode distance of 1.5–2 cm), applied over the belly of each muscle in accordance with SENIAM recommendations (Hermens et al. 2000). To place the electrodes on ES L1, the L1 spinal process was located by palpation by an experienced physical therapist and the electrodes were placed on either side, at a distance of 2 cm from the spinal process. EMG signals were filtered (bandpass 10 Hz to 1 kHz), amplified (×500–1,000), sampled at 2 kHz, and recorded on a computer using Signal 4.07 software for online and offline analyses (CED Micro1401 interface; Cambridge Electronic Design, Cambridge, UK).

Experimental Tasks

We evaluated interhemispheric interactions with TMS during the performance of three different motor tasks, all involving isometric contractions: 1) a unilateral flexion of the right shoulder, 2) unilateral extension combined to a right inclination of the trunk, and 3) unilateral abduction of the right index finger.

Shoulder task.

This task was performed with the right upper limb while the left upper limb remained at rest. In this task, ES L1 is activated to stabilize the trunk. Subjects were seated on an instrumented chair with the back unsupported and were instructed to maintain their back as straight as possible (Fig. 1A). Both arms were suspended vertically at the side of the body, with the elbow fully extended and the forearm in a neutral position. A fixation cuff was placed around the distal third of the right arm (above the elbow), mounted on a force transducer, and fixed to a rigid frame. Subjects were instructed to keep the contraction of AD at 10% of the maximal voluntary contraction (MVC) and were given visual feedback to maintain the appropriate level of contraction. Feedback was provided using a screen displaying the rectified and smoothed (2nd-order 1-Hz low-pass Butterworth filter) EMG activity of the AD. To determine the MVC for AD, subjects performed a maximal voluntary isometric shoulder flexion after first performing brief submaximal contractions at 25% and 50% of their subjective maximum. MVC was assessed a second time, and if its value exceeded the first MVC value by 10%, a third measurement of MVC was taken. Each MVC assessment was separated by a period of 2 min of rest, and the MVC was determined as the highest value obtained between the 2 or 3 trials. The MVC was first assessed for the right AD and subsequently for the left AD.

Fig. 1.

Experimental setup. A: shoulder task. Subjects performed an isometric shoulder flexion with the right AD at 10% MVC. Visual display of online rectified and smoothed AD EMG was presented to all subjects during testing so they could maintain the desired level of contraction throughout the experiment. The left upper limb was at rest. B: trunk task. Subjects performed a combined trunk extension/right inclination against a resistance (isometric contraction). Online visual feedback of rectified and smooth right ES L1 EMG was given to the subject, who had to maintain a target level of contraction of 30% MVC.

Trunk task.

In this task, we assessed interhemispheric interactions while ES L1 was the agonist of a unilateral movement. In a setup similar to that in the shoulder task, subjects did an isometric extension combined with a right inclination of the trunk to activate right ES L1 (Fig. 1B). Both right and left arms remained at rest in this task. The level of EMG recorded in ES L1 during the shoulder task was displayed as a target line on the screen for the subject to reach during the trunk task. In this way, we ensured that background EMG level for ES L1 was matched between the shoulder and the trunk task. The value for ES L1 MVC was determined with the same protocol as the one described for AD but during isometric extension combined with a right inclination for right ES L1.

Index finger task.

Subjects were seated on a chair with their hands and forearms resting on a table and the forearms pronated. In the right hand, the thumb was slightly abducted from the index finger so that the finger and the thumb formed an angle of ~45°–50° with each other. The radial side of the right index finger was in contact with a custom-made resistance to enable isometric abduction with the right index finger. The left hand remained at rest on the table. Subjects were instructed to maintain 10% MVC contractions of the right FDI and were provided with visual feedback.

TMS Protocols

In the present study, we chose to investigate how the nondominant hemisphere (right) affects the outputs of the dominant hemisphere (left) to muscles of contralateral body (hand, arm, and trunk) during movements. We thus assessed muscles on the right side of the body in right-handed individuals.

Ipsilateral cortical stimulation.

Single suprathreshold TMS pulses evoke the suppression of ongoing voluntary EMG activity in ipsilateral limb muscles (Ferbert et al. 1992; Meyer et al. 1995). This ipsilateral silent period (iSP) is thought to reflect inhibition from the stimulated to the nonstimulated hemisphere. Cortical stimulation over the right hemisphere was used to study the iSP in the right ES L1, AD, and FDI. Twelve subjects were recruited for this part, and assessment of all 3 muscles took place in a single 3-h session. TMS (single pulse, monophasic stimulator Magstim 200²; Magstim, Dyfed, UK) was applied over the right hemisphere, ipsilateral to the movement performed with the right side of the body. To determine the optimal coil positions for the stimulation (hotspot), the cortical areas enabling activation of left ES L1, left AD, and left FDI muscles were targeted, respectively. The hotspot was defined as the scalp position where the stimulus produces the largest motor evoked potential (MEP) amplitude at a given intensity in each of the muscles tested (Rossini et al. 1994). The location of the hotspot for each muscle was visualized and recorded on a computer through a neuronavigation system (Brainsight; Rogue Research, Montreal, QC, Canada) by placing trackers both on the coil and on the subject, which allowed us to maintain the coil at the same position throughout the experiment. The stimulation was performed with a figure-of-eight coil (70-mm diameter) positioned tangentially over the scalp with the handle directing posterolaterally and at 45° away from the midline to evoke current flow in an anteromedial direction (Di Lazzaro et al. 2004).

The active motor threshold (aMT) was established for left ES L1, AD, and FDI. For AD and FDI, aMT is defined as the lowest TMS stimulation output that produces MEPs of ≥100 µV in 5 of 10 trials (Chen et al. 2003) with a constant background EMG contraction at 10% of MVC of the targeted muscle. Evoking a MEP of 100 µV in ES L1 during its contraction as a stabilizer in the shoulder task was not possible in all subjects. Thus the aMT for ES L1 was defined as the lowest stimulation output that produced recognizable MEPs of ≥50 µV in 5 of 10 trials during contraction of ipsilateral AD at 10% MVC. Cortical stimulation of the right motor cortex (ipsilateral to the recorded muscles) was applied at an intensity of 130% of aMT. This intensity was chosen according to several studies highlighting that high intensity of stimulation was required to evoke iSP. Notably, Harris-Love et al. (2007) showed that significant IHI in distal (FDI) and proximal muscles (biceps and triceps brachii) could be evoked at 120% of resting motor threshold. Matthews et al. (2013) have tested different stimulation intensities, and motor responses could be evoked by stimulation of 120% of aMT for axial muscles (but not by stimulation of 80% of aMT). Because our experiment was conducted during contraction, we chose to identify aMT (instead of resting motor threshold), and a value of 130% was chosen after testing different intensities in pilot experiments and observing clearer responses than with 120%.

EMG responses to ipsilateral cortical stimulation were recorded in right AD and right ES L1 during the shoulder task and in right FDI during the index finger task. As a control, we compared responses obtained in ES L1 during the shoulder and trunk tasks in a separate session (n = 8). The data acquisition consisted of 30 control (no stimulation) and 30 stimulated trials applied in a randomized manner with at least a 7-s interval between stimuli. Pauses were taken halfway through the acquisition or more frequently if required by the subject to prevent and limit muscle fatigue.

Magnetic stimulation at the cervicomedullary junction.

To examine possible involvement of subcortical pathways to the responses evoked by ipsilateral stimulation on right ES L1, magnetic stimulation was applied at the cervicomedullary junction in 10 subjects in a single session (Taylor 2006). Stimulation at the cervicomedullary junction directly activates corticospinal axons to generate CMEPs in the arm muscles (Ugawa et al. 1994), and in some subjects, in the leg muscles (Butler et al. 2003; Ugawa et al. 1996). It thus could also evoke CMEP in the lower back, without activating the ventral roots of ES L1 located at the thoracolumbar junction. The responses evoked by such stimulation do not reflect changes due to intracortical activity and rather reflect changes occurring at spinal levels. Magnetic stimulation at the cervicomedullary junction can be used as a control to examine if responses evoked are solely due to changes in intracortical interhemispheric excitability (at a cortical level) or if there is a spinal contribution (Compta et al. 2006; Gerloff et al. 1998; Taylor 2006).

In the present study, stimulation at the cervicomedullary junction was used as a TS and ipsilateral stimulation over the right motor cortex was used as a CS (Perez et al. 2014). To evoke the CMEP, we used the technique described by Taylor (2006), with a circular coil (90-mm diameter) instead of a double-cone coil, because the subjects better tolerated the stimulation. The coil was first placed over the inion to have the current flow downward over the targeted area. It was then moved caudal and to the right along the right cervical spine until a clear evoked potential was observed in the right ES L1. This constituted the hotspot to elicit CMEPs in right ES L1. The aMT for CMEPs was assessed while subjects had their head slightly bent forward and performed a right isometric shoulder flexion at 10% of MVC. An intensity of 120% of the aMT was used to elicit CMEPs in right ES L1 (test response). Ipsilateral cortical stimulation was used as a CS and was applied on right motor cortex at 130% of the aMT found to elicit MEP in left ES L1 (see Ipsilateral cortical stimulation). Ipsilateral stimulation over the right motor cortex evoked a response composed of a facilitation followed by an inhibition in ES L1 (detailed in results). On the basis of these responses, stimulation at the cervicomedullary junction was applied 10 ms after the conditioned TMS pulse to elicit the CMEP in the middle part of the facilitation observed, and at 36 ms after the conditioned TMS pulse to elicit the CMEP in the middle part of the iSP (Perez et al. 2014).

Assessment of interhemispheric interactions with dual-coil paradigm.

A dual-coil paradigm was used to further assess interhemispheric interactions between cortical representations of ES L1, AD, and FDI. In this technique, a CS is applied to the cortex ipsilateral to the moving limb before a TS is delivered onto the contralateral hemisphere. Each muscle was evaluated in a separate session of 2 h (3 sessions in total). Interhemispheric interactions were assessed in right ES L1 (10 subjects), in right AD (8 subjects), and in right FDI (7 subjects). For right ES L1, a test MEP could only be evoked in 7 of 10 subjects. Therefore, the experiment was carried out in only 7 subjects. Interhemispheric interactions in AD and ES L1 were tested during the shoulder task, and interhemispheric interactions targeting FDI were assessed during the index finger task. The CS was applied to the right motor cortex at 130% aMT (as described in Ipsilateral cortical stimulation). The TS was applied to the left motor cortex at 120% aMT, found to elicit MEP in right ES L1, AD, and FDI. Because the aMT for ES L1 was very high, it was not possible to match the amplitudes of test MEP in ES L1 (in mV) with the test MEP in AD or FDI. This corroborates previous studies showing that it is more difficult to elicit MEPs in paraspinal muscles compared with limb muscles (Ferbert et al. 1992; Kuppuswamy et al. 2008; Nowicky et al. 2001; Taniguchi and Tani 1999). Hence, the intensity of stimulation, rather than amplitude of responses in millivolts, was kept identical in all muscles.

During assessment of either AD or FDI, both coils (figure-of-eight coils; 70-mm diameter each) were positioned tangentially over the scalp at ~45° to the midline with the handle directed posterolaterally. For ES L1 muscles, because their respective hotspots were close to each other, the coil used for the TS was positioned at 45° and the one used for the CS, at 90° to the midline. This enabled the active part of both TMS coils to be over their respective hotspots. For each muscle, 7 different ISIs were tested: 2, 4, 6, 8, 10, 12, and 40 ms. These ISIs were chosen according to previous reports suggesting that IHI at 8–10 ms and IHI at 40 ms may be mediated by different mechanisms (Chen et al. 2003). For each ISI, 10 control trials (no TMS pulse), 10 TS pulses alone, 10 CS pulses alone, and 10 TS preceded by the CS pulses were applied randomly.

Data Analysis

Ipsilateral silent period and ipsilateral motor evoked potential.

EMG responses in control and stimulated trials (30 of each) were recorded, rectified, and averaged separately for each condition with a custom-made MATLAB program (The MathWorks, Natick, MA). The mean and SD of the background EMG were calculated from a 200-ms window before the onset of the TMS stimulation. When an inhibition of the EMG was observed (iSP), three measures were taken: onset (latency), duration, and area (percentage of inhibition). By using data from the stimulated trials, the iSP onset was determined as the time point when the EMG dropped below the mean − 1SD for at least 10 ms, and the offset of iSP was the time point when the EMG rebounded above the mean − 1SD. The iSP duration was defined as the time window from the onset to the offset. The iSP area was calculated by integrating the EMG signal within the time window of the iSP on the control and stimulated trials. The percentage of inhibition was obtained by dividing the area of the stimulated trial by the corresponding area on the nonstimulated trial and multiplying by 100:

$$\text{\% Inhibition}=\frac{\mathit{\text{mean}}\hspace{0.17em}\mathit{\text{area}}\hspace{0.17em}\mathit{\text{of stimulated trial}}}{\mathit{\text{mean}}\hspace{0.17em}\mathit{\text{area}}\hspace{0.17em}\mathit{\text{of}}\hspace{0.17em}\mathit{\text{unstimulated trial}}}\times 100$$

When a facilitation (i.e., iMEP) was observed, its onset, offset, and area (percentage of facilitation) were also calculated in a manner similar to the iSP analysis described above, but with the signal rising above the mean baseline +1SD rather than going below. Because the characteristics of the tasks are different between ES L1 (stabilizing role) and AD or FDI (agonist role), the characteristics of iSP could only be compared between AD and FDI. Three parameters were compared between iSP elicited in AD and FDI: the latency, duration, and area (percentage of inhibition compared with background EMG).

Comparison of iMEP in ES L1 obtained in shoulder (stabilizing) vs. trunk tasks (agonist).

The percentage of facilitation calculated for iMEP evoked in ES L1 in the shoulder task was compared with the percentage of facilitation obtained in ES L1 during the trunk task, to determine changes in the amplitude of the facilitation (see Statistics). For this calculation, we controlled for background EMG in two ways: 1) the level of EMG in ES L1 recorded during the shoulder task was displayed as a target for the subject during the trunk task (rectified and filtered low-pass 1 Hz), and 2) offline, we compared the level of background EMG of right ES L1 during both tasks, to confirm assessment was made at matched EMG level. The mean background EMG level was measured for 200 ms before stimulation during each task and normalized to the MVC of ES L1.

Dual-coil paradigm and CMEP.

Peak-to-peak amplitudes of unrectified conditioned and test MEP were measured for each trial. The average amplitude of the 10 MEPs from conditioned trials was then expressed relative to the average amplitude of the 10 MEPs from test trials to determine the percentage of facilitation or inhibition. The following formula was used:

$$\begin{array}{c}\mathrm{\%}\mathrm{I}\mathrm{n}\mathrm{h}\mathrm{i}\mathrm{b}\mathrm{i}\mathrm{t}\mathrm{i}\mathrm{o}\mathrm{n}\mathrm{\hspace{0.17em}}\mathrm{o}\mathrm{r}\\ \mathrm{f}\mathrm{a}\mathrm{c}\mathrm{i}\mathrm{l}\mathrm{i}\mathrm{t}\mathrm{a}\mathrm{t}\mathrm{i}\mathrm{o}\mathrm{n}\end{array}=\frac{\mathit{\text{mean amplitude of conditioned MEP}}}{\mathit{\text{mean amplitude of test MEP}}}\times 100.$$

The same formula was used to calculate the percentage of facilitation or inhibition for the CMEP.

Statistics

Analyses assumptions were tested statistically. When a test's normality distribution assumption was not met according to Shapiro-Wilk, a transformation was used to attempt and make the data normally distributed or nonparametric statistics were used.

Ipsilateral silent period.

To compare iSP latency, duration, and percentage across AD and FDI, Student's paired t-test analyses comparing AD and FDI were used separately for each iSP aspect. For this analysis, only participants for which iSP was elicited in both AD and FDI were included.

Mixed model analyses were also performed to compare the same parameters of iSP between AD and FDI including all participants’ data (FDI: 9 participants; AD: 7 participants). Cohen's d effect size on the difference between AD and FDI was measured for each analysis.

iMEP in ES L1.

To determine if the amplitude of iMEP evoked by ipsilateral cortical stimulation and the background EMG were different when ES L1 was acting as an agonist or a stabilizing muscle, a Student’s paired t-test was performed. Cohen's d effect size on the difference between agonist and stabilizing was computed.

CMEP.

One-sample t-tests were conducted for each of the two ISIs (10 and 36 ms) to determine if the percentage of facilitation differed significantly from 100% on average, i.e., to test if the conditioned CMEP was significantly inhibited or facilitated compared with the unconditioned CMEP. Cohen's d effect sizes for the difference to 100% were computed.

Dual-coil assessment.

For each muscle separately, one-sample t-tests were conducted on each ISI to determine if they differed significantly from 100% on average (i.e., to determine if the CS delivered with a given ISI evoked inhibition or facilitation). Because seven different ISIs were assessed, the P values for each muscle were corrected for multiple testing using the Bonferroni procedure. Cohen's d effect sizes on the difference to 100% were computed for each delay and each muscle.

All values are means ± SE. Statistical significance was set at α = 0.05. All analyses were conducted using SPSS (version 20.0; SPSS, Chicago, IL) for Windows (Microsoft).

RESULTS

Parameters of the Ipsilateral Stimulation

For each subject and tested muscle, lateromedial (LM; positive values to the right) and anteroposterior (AP; negative values in front of Cz) locations of the hotspot in relation to the vertex were determined. On average, values for ES L1 were 2.5 ± 0.4 cm LM and −1.3 ± 0.3 cm AP, values for AD were 4.1 ± 0.3 cm LM and −1.9 ± 0.4 cm AP, and values for FDI were 6.5 ± 0.4 cm LM and −2 ± 0.3 cm AP. Over the 12 tested subjects, aMT was found to be 79 ± 3% of maximum stimulator output for ES L1, 48 ± 2% for AD, and 38 ± 2% for FDI.

Generation of iSP and iMEP

TMS was applied over the right motor cortex at 130% aMT. In a representative subject (Fig. 2A), ipsilateral stimulation of the ES L1 representation during the shoulder task evoked a short-latency facilitation (iMEP; EMG between the 2 solid lines), which was followed by an iSP (between the 2 dashed lines). In the same subject, ipsilateral stimulation over the cortical representations of AD during shoulder task (Fig. 2B) and FDI during index finger task (Fig. 2C) evoked an iSP with no prior facilitation. Group data show that iSP could be evoked in all muscles tested. Of the 12 subjects tested, iSPs were clearly elicited in 5 subjects in right ES L1 (41.6%), 7 subjects in right AD (58.3%), and 9 subjects in right FDI (75%) (Fig. 2E). The five participants that exhibited an iSP in ES L1 also exhibited an iSP in FDI when that muscle was targeted by the TMS, and three of those five participants also exhibited an iSP in AD when that muscle was tested. Interestingly, in four of five participants that exhibited an iSP in ES L1, an iSP in the right AD was simultaneously evoked.

Fig. 2.

Ipsilateral responses evoked by right motor cortex stimulation of the ES L1, AD, and FDI representations. A–C: raw EMG data of a representative subject show responses to TMS (black arrow indicates stimulation) of right motor cortex, recorded in ipsilateral ES L1 (A) and AD (B) during the shoulder task and in FDI (C) during the index finger task. Solid cursors 1 and 2 represent onset and offset of iMEP. Dashed cursors 3 and 4 represent the onset and offset of iSP. D: in another subject, raw EMG data show that only iMEP, without iSP, can be evoked in ES L1 by ipsilateral TMS. E: group data (n = 12) showing the probability of evoking an iSP or iMEP in each muscle, after ipsilateral TMS. In A–D, each trace is an average of 30 trials. R, right.

As stated in methods, the characteristics of iSP were only compared between AD and FDI, because they were involved in similar tasks (agonists and not stabilizer). iSPs were elicited in both right AD and right FDI (50%) in 6 of the 12 subjects tested. Statistical analysis on the assumptions for all parameters indicated no deviation from normality for the difference between AD and FDI (P = 0.295 to 0.995). Paired t-tests showed that the mean latency of iSP in AD (26.6 ± 2 ms) was significantly shorter than the latency of iSP in FDI (41.7 ± 3 ms; Cohen’s d = 2.07, P = 0.004; Fig. 3A). No significant difference was found between iSP mean area or duration between the two muscles (area: d = 0.19, P = 0.668; duration: d = 0.003, P = 0.995; Fig. 3, B and C).

Fig. 3.

Comparison of iSP between right AD and FDI in subjects that showed iSPs in both muscles (n = 6). Histograms compare iSP latency (A), duration (B), and area (C) between the muscles tested. iSP latency was significantly shorter in AD than in FDI. Values are means; error bars are SE. *P < 0.05.

When performed on all participants’ data (FDI: 9 subjects; AD: 7 subjects), mixed model analysis also showed a significant difference in the mean latency of iSP between FDI and AD (d = 0.61, P = 0.029). There was no significant difference in iSP mean area (d = 0.12, P = 0.628) or duration (d = 0.20, P = 0.446) between the two muscles.

Group data also show that ipsilateral TMS evoked a short-latency facilitation in ES L1 (75%) in 9 subjects but did not elicit iMEP in any of the other muscles tested (Fig. 2E). More precisely, four different patterns of responses combining iSP and iMEP were observed in ES L1: 1) in 5 subjects, the stimulation evoked only an iMEP (42%; Fig. 2, D and E); 2) in 4 subjects, the iMEP was followed by an iSP (33%; Fig. 2A); 3) in 1 subject, only an iSP was evoked (8%); and 4) 2 subjects did not show any response (16%). On average, the latency of the ipsilateral MEP evoked was 14.8 ± 1.4 ms, the mean duration was 18.1 ± 3.1 ms, and the mean percentage of facilitation compared with background EMG was 201.8 ± 17.8%. Thus ipsilateral cortical stimulation only evoked an iSP in AD and FDI muscles and both iMEP and iSP in axial ES L1 muscle.

Amplitude of iMEP in ES L1 Is Unchanged During the Trunk Task

As indicated in the previous paragraph, iMEPs were only evoked in ES L1. This might be due to the role of ES L1 in stabilizing the trunk during the shoulder task, whereas AD and FDI were agonists of their respective tasks. To determine whether the iMEPs could also be evoked if ES L1 were the agonist of the task, ipsilateral stimulation of right motor cortex (130% aMT) was applied on ES L1 representation while subjects performed both the shoulder task and the trunk task. The amplitudes of iMEP were then compared between both conditions. The assumption of normality was met for the shoulder task (P = 0.166). However, for the trunk task, the difference was nonnormal (P < 0.001), and as such, a log-transformation was applied, after which the assumption was met (P = 0.285). The effect size and P value for the second test are based on the transformed data, but the transformed and nontransformed data had similar results. Figure 4A shows the group data, with the comparisons performed with matched background EMG level in ES L1. The facilitation observed during shoulder task (217 ± 23.3%) was not significantly different from that evoked during the trunk task (187.74 ± 21.5%; d = 0.37, P = 0.33). Background EMG in ES L1 was equivalent during shoulder task (10.8 ± 3.5% MVC) and trunk task (8.3 ± 1.4% MVC; d = 0.17, P = 0.662; Fig. 4B). Thus the level of facilitation evoked by ipsilateral cortical stimulation does not appear to be dependent on the role of ES L1 muscle during the task.

Fig. 4.

Comparison of ipsilateral responses evoked by TMS over ES L1 representation during two different tasks. A: percentage of facilitation evoked in ES L1. Histogram shows group data (n = 8) of facilitation (iMEP) evoked during shoulder task (dark gray) and trunk task (light gray), respectively, compared with unstimulated background EMG level (control, 100%). There was no significant difference in the percentage of facilitation in ES L1 during shoulder and trunk tasks (P = 0.33). B: level of background EMG in right ES L1 during shoulder task (dark gray) and trunk task (light gray). Background EMG in ES L1 was equivalent during both tasks (no significant difference, P = 0.662).

Potential Pathways Responsible for iMEP and iSP in ES L1

The iMEP and iSP triggered in ES L1 by the ipsilateral cortical stimulation might not be conveyed through intracortical mechanisms (such as interhemispheric inhibition, IHI). They could also be mediated by subcortical changes. To investigate the involvement of subcortical mechanisms, the effect of ipsilateral TMS on the amplitude of CMEPs evoked in right ES L1 was tested in 10 participants. However, CMEP could not be observed in all subjects and data were gathered in only 8/10 subjects. The average location of the coil for inducing CMEP was at 7.85 ± 0.41 cm below (range 6.5–10 cm) and 3.19 ± 0.39 cm (range 1–4.5 cm) right of the inion, at the level of the C5 spinal process (Fig. 5A). This position could activate the corticospinal tract and produce reliable evoked potentials in right ES L1. Furthermore, because the coil was positioned rostrally to the thoracic and lumbar spinal segments, the magnetic stimulation does not directly activate ventral roots of ES L1.

Fig. 5.

Effect of ipsilateral cortical stimulation on CMEP elicited in right ES L1. A: experimental setup (CS, conditioned stimulus; TS, test stimulus; C7, spinous process of C7 vertebra). B and C: raw EMG data from a representative subject showing effect of ipsilateral stimulus on test CMEP response at ISI of 10 (B) and 36 ms (C). Black traces are unconditioned CMEP; gray traces are conditioned CMEP. D: group data showing conditioned CMEP responses at 10 (n = 8) and 36 ms (n = 7). *P < 0.05 compared with 100%.

Mean TMS intensity to evoke CMEP was 49.2 ± 3.8% (range from 32 to 75%), and the mean latency of CMEP was 8.86 ± 0.53 ms. Data reported above in Ipsilateral cortical stimulation showed that TMS over the right motor cortex evoked facilitation at a latency of 14.8 ± 1.4 ms and lasting 18 ± 3.1 ms, and an iSP at a latency of 41.8 ± 4.4 ms and lasting 30.1 ± 2.2 ms. Therefore, in a first experiment, the CMEP was evoked at 10 ms after the CS over the right cortical hemisphere to correspond to the time when iMEP was evoked. In a second experiment, CMEP was evoked at 36 ms after the CS. With these delays, the CMEP was evoked in the middle part of either the iMEP or the iSP. Both experiments were performed in a single session. Data from a single subject show a small facilitation at 10 ms but no change when the conditioned CMEP was applied 36 ms after stimulation (Fig. 5, B and C). The normality assumptions were met for both iMEP (P = 0.284) and iSP (P = 0.127). Group data confirmed these observations where significant facilitation was observed during iMEP (109.4 ± 2.5%; d = 1.34, P = 0.007; Fig. 5D), but changes were not significant during iSP (112.7 ± 10.1%; d = 0.47, P = 0.256). These findings suggest that iSP is evoked by intracortical mechanisms but that subcortical mechanisms might be involved in induction of iMEP.

Assessment of Transcallosal Inhibition Using the Dual-Coil Paradigm

To further characterize the nature of the interhemispheric interactions between representations of ES L1, AD, and FDI, a dual-coil paradigm was used. All the normality assumptions were met in ES L1 (P = 0.073–0.943), AD (P = 0.101–0.959), and FDI (P = 0.059–0.950). Transformations were still attempted for marginally significant variables, but the results were similar to the nontransformed data. Significant IHI was observed in ES L1 at ISI of 6 ms in a single subject (Fig. 6A) and in 6 of the 7 subjects where MEP could be evoked in the ES L1 (Fig. 6B). The significant inhibition detected at ISI of 6 ms became nonsignificant after multiple testing correction (76.4 ± 9.2%; P = 0.043, corrected P = 0.298, d = 0.97). No other significant facilitation or inhibition was detected for this muscle (d = 0.06–0.61, P = 0.154–0.877). In AD, the strongest inhibition was obtained at ISI of 8 ms and observed in 7 of 8 subjects (78.2 ± 4.7%; P = 0.002, corrected P = 0.017, d = 1,63; Fig. 6, C and D). No other significant facilitation or inhibition was detected for this muscle (d = 0.02–0.75, P = 0.073–0.949). As expected from previous literature (Chen et al. 2003; Harris-Love et al. 2007), inhibition was also observed in FDI in 6 of the 7 subjects tested, reaching significance only at ISI of 8 ms, but becoming nonsignificant after correction for multiple comparisons (76.8 ± 8.1%; P = 0.029, corrected P = 0.201, d = 1.01; Fig. 6, E and F). No other significant facilitation or inhibition was detected for this muscle (d = 0.05–0.8, P = 0.078–0.904). No significant facilitation was observed in any of the three muscles at any of the ISI tested. Thus the dual-coil paradigm showed predominant IHI between cortical representations of ES L1, AD, and FDI, albeit at different ISI.

Fig. 6.

IHI revealed by the dual-coil paradigm at different ISIs. A, C, E: superposition of test and conditioned MEPs in a representative single subject in ES L1 at 6 ms (A), in AD at 8 ms (C), and in FDI at 10 ms (E). B, D, F: group data showing the modulation of interhemispheric interactions with the different ISIs used in ES L1 (n = 7), AD (n = 8), and FDI (n = 7). *P < 0.05 (uncorrected).

DISCUSSION

The present study examined interhemispheric mechanisms in the control of trunk muscles when they stabilize the movement in a shoulder flexion task in healthy humans. Because previous studies suggested that iSP and dual-coil paradigm should be considered complementary measures of interhemispheric interaction (Boddington and Reynolds 2017; Chen et al. 2003), we used both ipsilateral cortical stimulation and dual-coil paradigms. During an isometric shoulder flexion task, interhemispheric inhibition (IHI) could be observed in ES L1. Ipsilateral cortical stimulation could also evoke facilitation in ES L1. This facilitation remained whether ES L1 was involved as a stabilizer (shoulder task) or agonist (trunk task). We then investigated the pathways underlying such responses in ES L1 by assessing CMEPs, which are not affected by changes in cortical excitability but are sensitive to changes in spinal excitability (Taylor 2006). The CMEPs tested during the time at which the iMEP was present were significantly facilitated.

These results suggest a contribution of subcortical mechanisms and ipsilateral corticospinal tract in the facilitation observed. However, the facilitation of CMEP was <10% (109.4 ± 2.5%), whereas facilitation evoked in ES L1 by ipsilateral cortical stimulation was 217% and 188% during shoulder and trunk tasks, respectively. This discrepancy indicates that the contribution of subcortical pathways in generation of iMEP, although present, is likely limited. On the contrary, the CMEPs elicited during iSP remained unchanged, suggesting that the iSP is mainly due to intracortical inhibition.

Together, our data support that the ipsilateral cortex does not have a different modulatory effect on paraspinal muscles depending on the goal of the task (i.e., stabilizing the trunk to support arm isometric contraction vs. trunk isometric contraction). We propose that cortical representations of paraspinal muscles have stronger ipsilateral facilitatory output to favor bilateral contractions and that interhemispheric interactions might not contribute primarily to the facilitation.

In this study, interhemispheric interactions were also assessed in AD, the agonist of the task, and were found to be inhibitory, similarly to the inhibition described thoroughly in the literature for FDI and which was also demonstrated in the current study.

Ipsilateral Cortical Outputs Vary in Function of the Role of the Muscles

Our results showed that iMEPs were evoked only in the axial muscle (ES L1), in 75% of subjects. No iMEP could be evoked in proximal (AD) and distal (FDI) arm muscles in any of the subjects with the protocol we used. This difference between axial and arm muscles is consistent with findings of other studies. When axial muscles have been assessed, previous studies have shown that unilateral stimulation of M1 produced ipsilateral excitatory responses in right and left axial muscles (trapezius, pectoralis) during tonic co-contraction (Bawa et al. 2004). Kuppuswamy et al. (2008) also reported generation of MEPs in both right and left ES L4 when applying TMS at the vertex, during unilateral arm abduction. Moreover, comparison of ipsilateral responses evoked by TMS of left M1 in a trunk muscle (internal oblique), deltoid, and FDI also showed that iMEPs were less frequent in arm muscles compared with trunk muscles (Strutton et al. 2004). Also, many studies reported difficulty in evoking iMEP in muscles of the arm (Chen et al. 2003; Ziemann et al. 1999), which often required bilateral contraction (Bawa et al. 2004).

Axial muscles are rarely contracted unilaterally even when the arm produces a unilateral movement. Trunk muscles, including ES, are involved in postural control during unilateral arm movement, and they are activated bilaterally to stabilize the trunk (Davey et al. 2002; Hodges et al. 1999, 2001). In the present study, facilitation was easily observed in right ES L1 following stimulation of the ipsilateral cortex during unilateral shoulder flexion, and the facilitation remained unchanged during unilateral contraction of ES L1. Thus the ipsilateral cortex would contribute to inhibit contralateral AD while it facilitates the ipsilateral ES L1 during unilateral arm movement. Recently, an inhibition of iMEP was reported during voluntary co-contraction of homologous upper limb muscles (Tazoe and Perez 2014). This discrepancy may be explained by the targeted muscles and the specificity and characteristics of the task. Notably, the load is greater during a shoulder movement than during an elbow movement with the arm supported. Although not observed in the current study, interhemispheric facilitation (or disinhibition) was observed by Perez et al. (2014) during bilateral activation of antagonistic proximal muscles, but not during bilateral activation of homologous proximal muscles, which led to pronounced inhibition. Other studies also reported possible interhemispheric facilitation in the FDI (Bäumer et al. 2006; Hanajima et al. 2001). In this context, the excitability of interhemispheric interactions might be task dependent.

Other Possible Neuronal Mechanisms for Generation of iMEP in Axial Muscles

Anatomic studies in primates show that proximal arm and axial muscles receive bilateral corticospinal innervation, whereas distal arm muscles are mostly controlled by the contralateral motor cortex (Brinkman and Kuypers 1973). Given these differences in the pattern of cortical projections, the generation of iMEPs in ES L1, and not in the other muscles, suggests that the ipsilateral corticospinal tract may also mediate, at least in part, the change in excitability evoked by ipsilateral stimulation of the cortical representation of ES L1 in M1.

Other bilaterally organized pathways have been proposed to mediate ipsilateral excitatory responses, such as corticoreticulospinal or propriospinal pathways (Ziemann et al. 1999). Studies in animals have shown that corticospinal fibers send collateral projections to nuclei from which reticulospinal neurons originate. These projections come mainly from the areas of the motor cortex that control movements of proximal rather than distal parts of limbs (Keizer and Kuypers 1984, 1989). Outputs from medullary reticular formation have been demonstrated to have bilateral effects on limb muscles in awake cats (Drew and Rossignol 1990) and monkeys (Davidson and Buford 2004). Thus reticulospinal pathway may be another candidate to mediate ipsilateral facilitatory responses evoked by TMS in ES L1.

Because the locations of ES L1 representations are very close to the vertex, one could question whether the facilitation observed is due to current spread from ipsilateral to contralateral M1, which would cause a nonspecific increase in excitability. Moreover, the intensity was high, which was deemed to be appropriate to evoke responses in ES L1 in pilot experiments performed on three subjects. Because of this small number of subjects, the selected stimulation intensity (130%) may not be optimal to evoke both facilitatory and inhibitory responses. The use of high intensity could also have facilitated current spread. Although we cannot completely exclude this possibility, the dual-coil stimulation data suggest that it is not the case. If the iMEP was generated by current spread, we would expect interhemispheric facilitation with all ISIs tested. This was not the case, because we found significant IHI between ES L1 cortical areas with ISI of 6 ms and no significant facilitation with any of the tested ISIs.

Possible Neuronal Mechanism for Generation of iSP

iSP is considered to reflect transcallosal inhibition between the two M1s given that it is absent or delayed in patients with agenesis or surgical lesions of the corpus callosum (Meyer et al. 1995, 1998). In our study, iSP could be evoked in ES L1, AD, and FDI. The iSP observed in FDI had a similar latency to that reported by Chen et al. (2003). When compared with other muscles, latency of iSP in FDI was longer than that obtained in AD. The distance between this muscle and the cortex could explain the shorter latency we found for AD. With our setup, iSP was clearly observed in 9/12 subjects in FDI, whereas it was observed in all subjects in other studies (Chen et al. 2003). This discrepancy is likely due to methodological differences. Our criteria to determine iSP were that the inhibition had to be below 1 SD from the mean for at least 10 ms, whereas in the study by Chen et al. (2003) the inhibition had to be below 1SD from the mean for at least 5 ms. Another methodological difference is the level of voluntary contraction required (10% MVC in our study, whereas its was 20% for Chen et al. 2003).

Our data also strongly suggest that iSP observed in ES L1 is due to changes in cortical excitability. Indeed, iSP may be mediated by transcallosal pathways given that no change was observed in the amplitude of the CMEP evoked during the period of time corresponding to the iSP. However, contribution of ipsilateral corticospinal tract cannot be ruled out, because Jung and Ziemann (2006) described a second phase of iSP in FDI that might be mediated by ipsilateral polysynaptic projections.

Dual-Coil Stimulation Further Showed the Predominance of Inhibitory Interhemispheric Interactions

In the current study, IHI was observed in AD and FDI at 8 ms, and significant inhibition was also observed in ES L1 at 6 ms. Because several different ISIs were tested, inhibition observed in ES L1 and FDI was no longer statistically significant after the correction for multiple tests. However, these inhibitions were robust given they were observed in 6 of 7 participants in ES L1 and FDI (compared with 7 of 8 participants in AD). The presence of inhibition in FDI at 8 ms is consistent with previous work that showed greater IHI in FDI at 8 and 10 ms (Chen et al. 2003; Harris-Love et al. 2007). Comparison of IHI observed in all three muscles is limited because 1) MEP amplitude could not be matched in all three muscles because MEP in ES L1 required a significantly higher intensity to be evoked compared with MEPs in AD or FDI (see methods); 2) investigation of all three muscles could not be performed in a single session because assessment for each muscle required 3 h to perform; 3) because the hotspots for right and left ES L1 were close together, the alignment of the coil delivering the conditioning pulse had to be at an angle of 90°, rather than 45°, which could change the efficiency of neuronal activation. Nonetheless, the presence of inhibition and absence of significant facilitation could be observed in ES L1 during the shoulder task. IHI has been previously shown in axial muscles while they were activated as agonists. Indeed, using dual-coil paradigm, Matthews et al. (2013) investigated interhemispheric interactions between representations of serratus anterior and trapezius, two muscles that stabilize the scapula during arm movements. Dual-coil stimulation was performed during unilateral or bilateral elevation of the arm through the scapula plane of motion (scaption). A low level of IHI was obtained at 8 ms between upper trapezius representations while the pairs of muscles were directly activated (agonist) during the scaption movement. Thus the current data suggest that interhemispheric facilitation is not a dominant mechanism for trunk stabilization during a shoulder task and that IHI is more prominent.

Clinical Implications

Understanding cortical mechanisms underpinning postural adjustments are important to better understand mechanisms enabling a movement. Previous studies have shown that a different cortical region, the supplementary motor area, is involved in the timing of postural adjustments (Jacobs et al. 2009; Viallet et al. 1992). The current study suggests that bilateral facilitation of axial muscles through interhemispheric interactions is present but might be limited. This could be of particular importance in cases where central nervous system is damaged, such as following a stroke. After a stroke, one hypothesis is that increased inhibition from the contralesional hemisphere on the affected hemisphere interferes with motor recovery of the paretic hand (Murase et al. 2004; Nowak et al. 2009), although a recent meta-analysis does not fully support this conclusion (McDonnell and Stinear 2017). However, recovery of arm movements after stroke does not solely depend on hand and arm muscles. Axial muscles stabilize the trunk to provide a stable platform for purposeful arm movements. (Bergmark 1989; Cirstea and Levin 2000; Colebatch and Gandevia 1989; Dewald et al. 1995; Michaelsen et al. 2001). Numerous reports have demonstrated that trunk activation could compensate for decreased arm movement after a stroke (Cirstea and Levin 2000; Levin et al. 2002; Robertson and Roby-Brami 2011; Roby-Brami et al. 2003; Thielman 2013), but the mechanisms leading to impairment and eventually recovery of (or compensation for) axial muscles are still unknown, which echoes our limited knowledge of cortical control of axial muscles in healthy subjects. The current study provides normative data to be compared with changes occurring after stroke and that could underlie greater recovery or increased compensation of trunk muscles during reaching arm movement.

In conclusion, interhemispheric inhibition is observed in trunk muscles while these muscles are used to stabilize the trunk during a unilateral arm movement. Facilitation could also be observed in ES L1 during the stabilization task with the use of the iSP paradigm, and a control with the CMEP confirmed this, but not during dual-coil stimulation assessment. This suggests that although interhemispheric mechanisms appeared to be responsible for this observation, subcortical mechanisms or direct activation of spinal networks by ipsilateral pathways are likely involved. The functional role of these interactions remains to be identified, and the relationship between motor cortical excitability and muscle activity still needs to be investigated, notably by using approaches that will modify the cortical excitability, such as repetitive TMS.

GRANTS

This research was funded by the Natural Sciences and Engineering Research Council of Canada, Fonds de recherche du Québec en Santé, Canada Foundation for Innovation, and REA Foundation grants (to D. Barthélemy), as well as a Sensorimotor Rehabilitation Research Team scholarship (to L. Jean-Charles).

DISCLOSURES

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

AUTHOR CONTRIBUTIONS

D.B. and N.D. conceived and designed research; L.J.-C., J.-F.N., and D.B. performed experiments; L.J.-C., J.E.D., G.E., and D.B. analyzed data; L.J.-C., J.E.D., G.E., N.D., and D.B. interpreted results of experiments; L.J.-C. and D.B. prepared figures; L.J.-C. drafted manuscript; L.J.-C., J.-F.N., J.E.D., G.E., N.D., and D.B. edited and revised manuscript; D.B. approved final version of manuscript.