Are ipsilateral motor evoked potentials subject to intracortical inhibition?

Abstract

Paired-pulse transcranial magnetic stimulation (TMS) can be used to examine intracortical inhibition in primary motor cortex (M1), termed short-interval intracortical inhibition (SICI). To our knowledge, SICI has only been demonstrated in contralateral motor evoked potentials (MEPs). Ipsilateral MEPs (iMEPs) are assumed to reflect excitability of an uncrossed oligosynaptic pathway, and can sometimes be evoked in proximal upper-limb muscles using high-intensity TMS. We examined whether iMEPs in the biceps brachii (BB) would be suppressed by subthreshold conditioning, therefore demonstrating SICI of iMEPs. TMS was delivered to the dominant M1 to evoke conditioned (C) and nonconditioned (NC) iMEPs in the nondominant BB of healthy participants during weak bilateral elbow flexion. The conditioning stimulus intensities tested were 85%, 100%, and 115% of active motor threshold (AMT), at 2 ms and 4 ms interstimulus intervals (ISI). The iMEP ratio (C/NC) was calculated for each condition to assess the amount of inhibition. Inhibition of iMEPs was present at 2 ms ISI with 100% and 115% AMT (both P < 0.03), mediated by a reduction in persistence and size (all P < 0.05). To our knowledge, this is the first demonstration of SICI of iMEPs. This technique may be useful as a tool to better understand the role of ipsilateral M1 during functional motor tasks.

the cortical control of ipsilateral proximal upper limb muscles is rarely studied in healthy people. This is surprising, given that ipsilateral control of the upper limb is evident in healthy people (e.g., Tazoe and Perez 2014; Verstynen et al. 2005), and the recovery of motor function after adult stroke (e.g., Ward et al. 2006; Schwerin et al. 2008) and cerebral palsy (e.g., Eyre et al. 2001, 2007). Further, the role of intracortical inhibition in control of the ipsilateral proximal upper limb is not well understood. Exploring new paradigms to probe intracortical inhibitory circuits in the ipsilateral primary motor cortex (M1) may help elucidate the role of ipsilateral M1 during upper limb movement.

Short-interval paired-pulse transcranial magnetic stimulation (TMS) is an established noninvasive measure of intracortical inhibition in M1. Typically, short-interval intracortical inhibition (SICI) is performed by delivering a subthreshold conditioning stimulus (CS), followed 1–5 ms later by a suprathreshold test stimulus to elicit motor evoked potentials (MEPs) in muscles of the contralateral target muscle. The suppressive effect of conditioning is thought to be mediated by GABA synaptic activity via GABA_A-mediated receptors, as SICI is enhanced by allosteric GABA_A receptor modulators (Ziemann et al. 1996a, b). To our knowledge, no study has examined the effect of short-interval paired-pulse TMS on ipsilateral MEPs. Evidence of MEP suppression from such a technique could yield new insights into the functional role of the ipsilateral M1 for upper limb control.

Ipsilateral MEPs (iMEPs) can be evoked in a moderate proportion of those studied by using high-intensity TMS applied over M1 and preactivation of the ipsilateral target muscle (Tazoe and Perez 2014; Ziemann et al. 1999). iMEPs are thought to reflect excitability of an uncrossed oligosynaptic pathway, such as the cortico-reticulospinal or cortico-propriospinal pathway (Ziemann et al. 1999). They are characterized by a long latency and a high threshold, are more likely found in proximal muscles (Bawa et al. 2004), and can be modulated by bilateral homonymous (agonist-agonist) and heteronymous (agonist-antagonist) elbow flexion contractions (Tazoe and Perez 2014), neck rotation (Tazoe and Perez 2014; Ziemann et al. 1999), and following repetitive TMS that is applied as a continuous τ-burst stimulation (Bradnam et al. 2010). iMEPs have been observed in a patient with complete agenesis of the corpus callosum (Ziemann et al. 1999), and they are known to be upregulated at the chronic stage after stroke, proportional to the severity of upper limb impairment (Schwerin et al. 2008).

Another protocol presumed to assess putative intracortical inhibition in M1 is the use of subthreshold TMS during tonic isometric muscle contraction, and averaging over many trials (Davey et al. 1994). Using this technique subthreshold TMS over ipsilateral M1 exhibited more electromyography (EMG) suppression in the ipsilateral biceps brachii (BB) during bilateral elbow flexion than unilateral elbow flexion (Tazoe and Perez 2014). For this reason, we used a bilateral elbow flexion task in this experiment. Our hypothesis was that iMEPs in the BB would be suppressed by subthreshold conditioning at ISIs known to produce SICI of contralateral MEPs (cMEPs). On the basis of previous studies with cMEPs, we hypothesized that a 2-ms interstimulus interval (ISI) would induce more suppression of iMEPs than a longer ISI, and more suppression would be observed with stronger CS intensities (Chen et al. 1998; Peurala et al. 2008).

METHODS

Participants.

In total, 25 adults were screened for the presence of iMEPs in the nondominant BB. Ten neurologically healthy adults (mean age 25.1 yr, range 20–31 yr, 4 males, 6 females, 1 left-handed) met the study criteria and were included in the analysis. Participants were included if they produced iMEPs in >50% of trials with single-pulse TMS (14 excluded). All data were collected in a single session. Participants gave written informed consent, and the local ethics committee approved the study in accordance with the Declaration of Helsinki. Participants were assessed for contraindications to TMS by a neurologist, and handedness was assessed with the Edinburgh Handedness Inventory (Oldfield 1971).

Electromyography.

Surface EMG was recorded from the short head of left and right BB using disposable electrodes (Ambu Blue Sensor Paediatric NS, Ballerup, Denmark) placed over the muscle bellies 2.5 cm apart. Standard skin preparation procedures were used. EMG signals were amplified (CED 1902; Cambridge Electronic Design, Cambridge, UK), band-pass filtered (10-1,000 Hz), and sampled at 2 kHz (CED 1401).

Task configuration.

Participants were seated with shoulders neutral, elbows resting on a firm surface with an elbow angle of 90° and both forearms supine. A cuff was secured around each wrist, attached to metal rods embedded with force transducers. All participants performed two or three maximal isometric elbow flexion contractions with both arms for 3–5 s. The force was recorded using PowerLab and LabChart software (ADInstruments, Dunedin, New Zealand). For the remainder of the experiment, participants held bilateral elbow flexion at 10% of their maximum voluntary contraction (MVC), and targets were displayed on a screen to encourage accurate task performance.

Transcranial magnetic stimulation.

TMS was delivered to dominant M1 using a figure-of-eight D70² coil and two Magstim model 200 stimulators connected to a BiStim unit (Magstim, Whitland, UK). The coil was held tangential to the scalp, with cortical current directed posterior to anterior (Bradnam et al. 2011; Tazoe and Perez 2014), and the optimal site for eliciting cMEPs in the dominant BB was marked on the scalp. Active motor threshold (AMT) in the contralateral BB was defined as the minimum stimulus intensity to evoke a 200-μV cMEP in four out of eight trials during bilateral elbow flexion at 10% MVC [mean AMT = 36.9% maximum stimulator output (MSO), range 28–47% MSO].

To determine whether an individual produced iMEPs, 12 stimuli at 100% MSO were delivered to the dominant M1 during bilateral elbow flexion at 10% MVC. Rest breaks were given every four stimuli. iMEPs were identified from the average rectified EMG trace of the nondominant BB (Fig. 1). An iMEP was deemed acceptable when the EMG from the waveform average exceeded the mean background EMG (BG) + 1 SD for at least 5 ms (Ziemann et al. 1999).

Fig. 1.

Average-rectified electromyograph (EMG) recordings from the ipsilateral BB of a representative participant. The nonconditioned (NC) trace is shown in black, the conditioned trace (C) at 2 ms with CS₁₀₀ is shown in gray. The ipsilateral motor evoked potentials (iMEP) ratio was 0.42. iMEPs were deemed present in an individual if the EMG exceeded the mean background (BG) EMG + 1 SD for >5 ms (Ziemann et al. 1999). Ipsilateral MEPs were measured as the area between the onset and offset, less an equivalent window of BG EMG, iMEP = iMEP area − BG area.

For paired-pulse TMS, the test stimulus intensity was the lowest intensity that produced above threshold (>100 μV/ms) and persistent (>50%) iMEPs. The test stimulus was determined by delivering blocks of 12 stimuli at decreasing intervals of 5–10% MSO from maximum, until an iMEP was no longer deemed acceptable from the waveform average of the ipsilateral EMG. The CS intensities were 85%, 100%, and 115% of the contralateral BB AMT (termed CS₈₅, CS₁₀₀, CS₁₁₅) and delivered before the test stimulus, with an interstimulus interval (ISI) of 2 ms or 4 ms. Twenty-four nonconditioned (NC) and twelve conditioned (C) traces were collected in a randomized order for each condition (2 ISI × 3 CS). In total, 96 traces were collected. Stimulation frequency was 0.2 Hz, and a rest break was given after every six stimuli.

Data processing.

Ipsilateral MEPs were measured from rectified EMG for the nondominant BB. The iMEP onset and offset were determined from the averaged waveform trace at 100% MSO and were used as the individualized iMEP window for each participant. The onset was the earliest deflection of the EMG that was maintained above the BG mean + 1 SD for at least 5 ms. The offset was the first instance the EMG returned to the BG mean + 1 SD. From each trace, iMEPs were measured as the area within the iMEP window, less an equivalent window of BG EMG area; iMEP (μV/ms) = iMEP area − BG area.

Persistence of iMEPs was calculated as the number of trials in which the iMEP was >100 μV/ms out of the total for each condition. A threshold of 100 μV/ms was chosen on the basis that it provides an objective criterion to exclude trials in which an iMEP is not present. Previous studies have relied on visual inspection of the trace (Schwerin et al. 2008, 2011). The iMEP latency was obtained from trials in which the iMEP was >100 μV/ms. The iMEP latency was measured from the raw EMG as the first prominent deflection of the EMG within a predetermined iMEP window (i.e., 3–15 ms later than the cMEP latency).

The cMEP latency was measured from the raw EMG as the first deflection at least 8 ms after TMS. Contralateral MEPs were measured by calculating the integral of rectified EMG for the dominant BB. The cMEP area was calculated in a 20-ms window from the cMEP latency (mean = 10.1 ms, range 9.0–11.5 ms) and was expressed as the difference between the cMEP area and an equivalent window (i.e., 20 ms) of background EMG; cMEP (μV/ms) = cMEP area − BG area.

The root mean square of the EMG (rmsEMG) was calculated from the prestimulus EMG between −100 and −10 ms in the ipsilateral and contralateral BB to ensure background EMG activity was not different between conditions.

Statistical analysis.

The iMEP and cMEP ratios were calculated (C/NC) for each condition. The iMEP ratio was an average of all trials per condition, and the iMEP+ ratio was an average of trials in which the iMEP was >100 μV/ms. One-sample t-tests of the iMEP and cMEP ratios were used to detect differences from 1. Delta (Δ) iMEP persistence was calculated as the difference between C and NC iMEP persistence (Δ persistence = C − NC). The iMEP ratio, iMEP+ ratio, cMEP ratio, Δ persistence, and iMEP latency were analyzed with a 2 ISI (2 ms, 4 ms) × 3 CS intensity (CS₈₅, CS₁₀₀, CS₁₁₅) repeated-measures ANOVA. To determine whether MEP ratios were affected by NC MEP size, linear regression analyses were conducted between the NC MEP area for each condition and the respective iMEP, iMEP+, and cMEP ratio. Paired t-tests between the NC iMEP latency and C iMEP latency for each condition were also performed. The rmsEMG for the ipsilateral and contralateral BB were analyzed in separate one-way ANOVAs.

Effect sizes from ANOVA were determined as partial ε-squared (η_p²), and Cohen's d are reported from one-sample t-tests. Effects were deemed significant if P < 0.05 and post hoc tests were conducted using paired and one-sample t-tests. For multiple pairwise comparisons a modified Bonferroni procedure was used (Rom 1990). Means ± SE are reported in the text.

RESULTS

One participant was excluded as an outlier because their paired-pulse TMS measures were >3 SD from the population mean. The main result is shown in Fig. 2A. One-sample t-tests of the iMEP ratios confirmed that iMEPs were suppressed with an ISI of 2 ms with CS₁₀₀ (0.65 ± 0.14; t₉ = −2.53, P = 0.03, d = −0.80) and CS₁₁₅ (0.64 ± 0.11; t₉ = −3.44, P = 0.007, d = −1.09), but not CS₈₅ (0.85 ± 0.13; t₉ = −1.12, P = 0.29, d = −0.35). There was no iMEP suppression at 4 ms with any conditioning intensity (CS₈₅: 1.30 ± 0.26; t₉ = 1.17, P = 0.27, d = 0.37; CS₁₀₀: 1.05 ± 0.15; t₉ = 0.35, P = 0.74, d = 0.11), although there was a trend with CS₁₁₅ (0.72 ± 0.14; t₉ = −2.06, P = 0.07, d = −0.65). ANOVA indicated a main effect of ISI (F_1,9 = 9.52, P = 0.01, η_p² = 0.51) and CS intensity (F_2,18 = 6.72, P = 0.01, η_p² = 0.43), and no interaction (F_2,18 = 1.30, P > 0.30, η_p² = 0.13). Suppression was greater with 2 ms ISI compared with 4 ms ISI (2 ms = 0.71 ± 0.10; 4 ms 1.02 ± 0.15). Post hoc analyses revealed more iMEP suppression at CS₁₁₅ than CS₈₅ (P = 0.02), but there were no differences between CS₈₅ and CS₁₀₀ (P = 0.10), or CS₁₀₀ and CS₁₁₅ (P = 0.14) [CS₈₅ = 1.08 ± 0.17; CS₁₀₀ = 0.85 ± 0.12; CS₁₁₅ = 0.68 ± 0.10].

Fig. 2.

Group averages (n = 10) of iMEP ratios and contralateral (cMEP) ratios for each conditioning stimulus intensity at 2 ms (black) or 4 ms (gray) interstimulus intervals (ISI). A: iMEP ratios included all trials. The repeated-measures ANOVA revealed a main effect of ISI (P = 0.01) and CS (P = 0.01). Suppression was present at 2 ms with CS₁₀₀ and CS₁₁₅ (*P < 0.05) and a trend for 4 ms at CS₁₁₅ (#P < 0.1). B: iMEP+ ratios included trials with an iMEP > 100 μV/ms and, therefore, excluded trials in which an iMEP was not present. Suppression of iMEPs was found at 2 ms with CS₁₀₀ (*P < 0.05) and a trend for both ISIs at CS₁₀₀ (#P < 0.1). C: cMEP ratios. Inhibition was present at 2 ms and 4 ms with CS₁₀₀ and CS₁₁₅ (*P < 0.05) and a trend for 2 ms with CS₈₅ (#P < 0.1). D: dot plot of individual iMEP (circle) and cMEP (triangle) ratios for each CS intensity at 2 ms (black) and 4 ms (gray) ISIs. Each data point within a condition represents one participant.

The persistence of iMEPs was 74.1 ± 5.6% (Fig. 3A). For ΔiMEP persistence, there was a main effect of ISI (F_1,9 = 5.85, P = 0.04, η_p² = 0.39) and a trend for CS (F_2,18 = 2.79, P = 0.09, η_p² = 0.24) and no ISI × CS interaction (F_2,18 = 0.32, P = 0.73, η_p² = 0.03) (Fig. 3B). ΔiMEP persistence was less at 2 ms than 4 ms (2 ms = −7.5 ± 5.7%; 4 ms = −1.6 ± 5.7%; Fig. 3B), and one-sample t-tests confirmed lower ΔiMEP persistence at 2 ms (t₂₉ = −2.13, P = 0.04, d = −0.39) but not 4 ms (t₂₉ = −0.45, P = 0.66, d = −0.08).

Fig. 3.

A: group averages (n = 10) of iMEP persistence. Persistence was calculated as the number of trials the iMEP was >100 μV/ms out of the total number of trials. NC iMEP persistence is shown as the open bar. B: ΔiMEP persistence was the difference between each condition (C) from NC. The repeated-measures ANOVA revealed a main effect of ISI (*P = 0.04), with persistence at 2 ms lower than 4 ms.

Analysis of iMEP+ ratio included only trials where an iMEP was >100 μV/ms. The pattern of results was similar to that above. iMEP+ ratios were suppressed at 2 ms with CS₁₀₀ (0.81 ± 0.08; t₉ = −2.27, P = 0.049, d = −0.72) and trends with CS₁₁₅ at 2 ms (0.80 ± 0.09; t₉ = −2.09, P = 0.066, d = −0.66), and 4 ms (0.83 ± 0.09; t₉ = −1.84, P = 0.099, d = −0.58) (all other P > 0.16; Fig. 2B). There was a main effect of CS intensity (F_2,18 = 4.88, P = 0.02, η_p² = 0.35), and no other main effects or interactions (all P > 0.12). The iMEP+ ratio at CS85 showed less suppression than at CS₁₀₀ (P = 0.048) or CS₁₁₅ (P = 0.02), with no difference between them [CS₈₅ = 1.10 ± 0.08; CS₁₀₀ = 0.94 ± 0.07; CS₁₁₅ = 0.82 ± 0.09].

iMEP latency was consistent across conditions and compared with NC (all P > 0.30). The average iMEP latency was 18.81 ± 0.29 ms, which was 8.72 ± 1.89 ms later than the cMEP latency.

As expected, cMEPs were suppressed at ISI of 2 ms with CS₁₀₀ (0.77 ± 0.06; t₉ = −3.69, P = 0.005, d = −1.17) and CS₁₁₅ (0.76 ± 0.09; t₉ = −2.75, P = 0.022, d = −0.87) and ISI of 4 ms with CS₁₀₀ and CS₁₁₅ (CS₁₀₀: 0.78 ± 0.05; t₉ = −4.85, P = 0.001 d = −1.53, CS₁₁₅: 0.71 ± 0.06; t₉ = −5.03, P = 0.001, d = −1.59, Fig. 2C). There was also a trend for inhibition at 2 ms with CS₈₅ (0.92 ± 0.05; t₉ = −1.85, P = 0.097, d = −0.58). There was a main effect of CS intensity (F_2,18 = 9.32, P = 0.002, η_p² = 0.51) and no other main effects or interactions (all P > 0.70). The cMEP ratio at CS₈₅ was higher than CS₁₀₀ and CS₁₁₅ (both P < 0.01), and there was no difference between CS₁₀₀ and CS₁₁₅ (P = 0.38) [CS₈₅ = 0.92 ± 0.02; CS₁₀₀ = 0.78 ± 0.04; CS₁₁₅ = 0.73 ± 0.06]. For comparison, Fig. 2D depicts the iMEP and cMEP ratios for each participant.

The average NC MEP was 2,153 ± 500 μV/ms for the iMEP ratio, 2,792 ± 460 μV/ms for the iMEP+ ratio, and 29,628 ± 7,046 μV/ms for the cMEP ratio. There was no association between NC area and the respective iMEP ratio (R < 0.23, all P > 0.52), iMEP+ ratio (R < 0.48, all P > 0.17), or cMEP ratio (R < 0.40, all P > 0.26).

The pretrigger rmsEMG was not different between conditions for the ipsilateral (grand average = 152.86 μV ± 6.99) and contralateral BB (grand average = 153.71 ± 4.54) (both P > 0.998).

DISCUSSION

To our knowledge, this is the first demonstration of iMEP suppression using short-interval paired-pulse TMS. Suppression of iMEPs occurred via reduced iMEP persistence and size. Suppression of cMEPs also occurred at the same ISIs, which are known to elicit SICI of cMEPs, a GABA_A receptor-mediated inhibitory process. There is a growing body of evidence from neurophysiological and neuroimaging studies that suggest a role of ipsilateral M1 during skilled upper limb movement in healthy individuals (Diedrichsen et al. 2013; McCambridge et al. 2011; Uehara and Funase 2014; Verstynen et al. 2005) and those affected by stroke (Bradnam et al. 2013; Riecker et al. 2010; Ward et al. 2006). Short-interval paired-pulse TMS of iMEPs could be useful for understanding how intracortical inhibitory circuits that act on ipsilateral motor pathways are modulated during functional motor tasks.

In the present study, 10 of 25 participants produced acceptable iMEPs based on previously described acceptance criteria (Ziemann et al. 1999). This is a low to moderate proportion of responders compared with other studies targeting iMEPs in the BB (Bradnam et al. 2010; McCambridge et al. 2011; McCambridge et al. 2014; Tazoe and Perez 2014; Ziemann et al. 1999). One reason for this low proportion may have been the relatively weak task contraction compared with previous studies (Tazoe and Perez 2014; Ziemann et al. 1999). However, given that intracortical inhibition is downregulated with voluntary contraction (Reynolds and Ashby 1999; Roshan et al. 2003), we limited the contraction strength to 10% MVC to determine whether iMEP suppression would be evident. The low proportion of responders is, therefore, likely reflective of the weaker contraction, and not indicative of less reliance on ipsilateral control of the upper limb.

Ipsilateral MEPs were present in 74% of trials overall, with a latency of 18.8 ± 0.2 ms, which is consistent with other studies in the BB (Bradnam et al. 2010; Lewis and Perreault 2007; McCambridge et al. 2011, 2014; Tazoe and Perez 2014). Paired-pulse TMS suppressed iMEPs with an ISI of 2 ms at stronger CS intensities (CS₁₀₀ and CS₁₁₅, Fig. 2A). The persistence of iMEPs was reduced by 7.5% at 2 ms ISI (Fig. 3B). Of the trials in which an iMEP was present (i.e., iMEP+ ratio included only iMEPs > 100 μV/ms), the suppressive effect of paired-pulse TMS with 2 ms at CS₁₀₀ was still evident (Fig. 2B). Therefore, suppression of iMEPs was mediated by both reduced iMEP persistence and size.

In the present study, cMEPs were suppressed by both ISIs with stronger CS intensities (CS₁₀₀ and CS₁₁₅, Fig. 3A). In general, short-interval paired-pulse TMS similarly modulated contralateral and ipsilateral MEPs across each condition. The main contrast between conditions was seen at 4 ms with CS₁₀₀, where suppression was found for cMEPs but not iMEPs. One reason for this could relate to differences in the threshold of ipsilateral vs. contralateral pathways (Bawa et al. 2004; Ziemann et al. 1999), as a slightly higher CS intensity (i.e., CS₁₁₅ at 4 ms) did produce iMEP suppression. Alternatively, it could be related to interindividual variability, as iMEP ratios appeared to be more variable than cMEP ratios (Fig. 2D). A limitation of the study was that the TMS hotspot was not optimized for iMEPs; therefore, this could have influenced iMEP variability. Because paired-pulse TMS suppressed both contralateral and ipsilateral motor pathways at known ISIs and CS intensities for SICI, we speculate that these effects reflect intracortical inhibition in M1, whereby subthreshold conditioning activated intracortical interneurons that inhibited pyramidal neurons of both motor pathways.

As far as we know, this is the first report of intracortical inhibition of iMEPs. There are several possible avenues for future neurophysiological investigations. Pharmacological studies could specifically determine whether the observed suppression of iMEPs is dependent on a particular GABA_A-receptor subunit (e.g., Di Lazzaro et al. 2006). Another question is how pyramidal neurons in M1 are influenced by both facilitatory and inhibitory networks. For example, suprathreshold paired-pulse TMS can facilitate iMEPs in both healthy and chronic stroke participants (Schwerin et al. 2011). Ipsilateral motor pathways from contralesional M1 may be upregulated after stroke (Caramia et al. 2000; Schwerin et al. 2008), especially when the ipsilesional corticospinal tract has been damaged (Bradnam et al. 2013; Ward et al. 2006). It remains to be determined whether intracortical inhibitory networks functionally modulate ipsilateral pathways in a task-dependent manner. Such investigations may provide insight into neural reorganization and motor recovery in conditions such as stroke or cerebral palsy.

GRANTS

A. B. McCambridge was supported by a Health Research Council PhD Scholarship. This research was supported by a Faculty Research Development Grant to W. D. Byblow.

DISCLOSURES

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

AUTHOR CONTRIBUTIONS

Author contributions: A.B.M. and W.D.B. conception and design of research; A.B.M. performed experiments; A.B.M. and W.D.B. analyzed data; A.B.M. and W.D.B. interpreted results of experiments; A.B.M. prepared figures; A.B.M. and W.D.B. drafted manuscript; A.B.M., J.W.S., and W.D.B. edited and revised manuscript; A.B.M., J.W.S., and W.D.B. approved final version of manuscript.