Evidence for normative intracortical inhibitory recruitment properties in cervical dystonia

Abstract

Objective:

Dystonia is associated with reduced intracortical inhibition as measured by the cortical silent period (cSP); however, this may be due to abnormal cSP threshold or input-output properties. This study evaluated cSP recruitment properties in people with cervical dystonia (CD).

Methods:

Bilateral electromyographic recordings were collected in the upper trapezius muscle in response to transcranial magnetic stimulation of the left and right primary motor cortex in a group with CD (n=19) and controls (n=21). cSP threshold, cSP input-output properties at stimulation intensities from 1 to 1.4x the cSP threshold, ipsilateral silent period duration (iSP) and timing and magnitude of the contralateral and ipsilateral motor evoked potential (MEP) were assessed.

Results:

The cSP threshold, input-output properties, and contralateral MEP magnitude were not significantly different between groups (all p>0.07). Hemispheric symmetry was present in the control group while the CD group had reduced iSP (p<0.01) and a trend for reduced ipsilateral MEP response (p=0.053) in the left hemisphere.

Conclusions:

Recruitment properties of intracortical inhibition are similar between control and CD groups. Transcallosal inhibition is asymmetric between hemispheres in people with CD.

Significance:

Evidence of normative intracortical inhibition recruitment properties challenge the commonly held view that cortical inhibition is reduced in dystonia.

Highlights

• Normative intracortical inhibition recruitment properties found in dystonic neck musculature.

• Transcallosal inhibition is symmetric in controls and asymmetric in people with cervical dystonia.

• Ipsilateral excitatory responses may be downregulated in people with cervical dystonia.

1. Introduction

Cervical dystonia (CD) is a neurologic movement disorder that is characterized by involuntary musculature contractions of neck musculature. Although CD has a poorly understood pathophysiology, there is a general consensus that idiopathic focal dystonia is associated with decreased inhibitory function (for review, Hallett 2011). One common method to assess cortical inhibition is via the use of transcranial magnetic stimulation (TMS) to evoke a cortical silent period (cSP) (Kimiskidis et al., 2006; Siebner et al., 1998). The cSP is characterized by a prolonged suppression of activity in the contralateral target muscle and is thought to be mediated, in part, by intracortical gamma aminobutyric acid (GABA) producing interneurons (Kimiskidis et al., 2006; Siebner et al., 1998; Ziemann et al., 1996). The duration of the cSP has been found to be reduced in various forms of idiopathic focal dystonia (Chen and Hallett, 1998; Filipovic et al., 1997; Kimberley et al., 2009; Rona et al., 1998; Samargia et al., 2016, 2014). Specific to CD, cSP duration has been found to be reduced in affected neck musculature (Amadio et al., 2000; Cakmur et al., 2004) and non-affected hand musculature (Odorfer et al., 2019; Quartarone et al., 2008). Yet, conflicting results have indicated normal cSP durations in CD affected (Odergren et al., 1997) and non-affected hand muscles (Cakmur et al., 2004). Disagreement between studies may be due to variable and suboptimal cSP testing methods that fail to precisely assess inhibitory pathway excitability that medicate the cSP response.

Traditionally, the stimulation intensity used to assess the cSP is ‘suprathreshold’, meaning it is a stimulus intensity determined from the resting motor threshold (RMT) or the minimum intensity required to evoke a motor evoked potential (MEP) (Rossini et al., 2015). Yet, this approach assumes that the cortico-motor threshold (i.e. RMT) and inhibitory pathway threshold (i.e. cSP) are the same. This assumption may not be appropriate because the neural elements mediating MEPs and cSPs are different. For example, the optimal cortical location to evoke an MEP compared to a cSP differ and the stimulus-response curves differ between the two outcomes (Kallioniemi et al., 2014; Kimiskidis et al., 2005; van Kuijk et al., 2009; Werhahn et al., 2007). Further, cSP duration is modulated by stimulus intensity (Kukowski and Haug, 1992) and thus precise calibration of the testing intensity to the inhibitory pathway under evaluation is essential. In a pathological condition that may be characterized by abnormalities in specific cortical pathways (i.e., inhibitory or excitatory circuits), it is feasible that the relationship between these thresholds are especially important to establish. One method to calibrate the stimulus intensity for precise cSP testing is to measure the cSP recruitment threshold, which may allow for an indirect estimation of inhibitory interneuron excitability (Wassermann et al., 1993). Using stimulus intensities based on the cSP threshold reduces the influence of excitatory components on the estimation of cSP duration, providing a more precise index of inhibitory pathway excitability (Kallioniemi et al., 2014; Wassermann et al., 1993). Estimation of subject-specific cSP recruitment thresholds or input-output properties has not been completed in people with dystonia.

The primary aim of this study was to estimate subject-specific recruitment thresholds and input-output properties of inhibitory pathways that mediate the cSP in people with cervical dystonia (CD) compared to healthy, age-matched controls. Outcome measures were compared between groups and hemisphere to explore levels of hemispheric symmetry. We hypothesized that participants with CD would have lower excitability of inhibitory interneurons that mediate cSP duration compared to controls, generating shorter cSP durations on average.

In addition to examining the cSP threshold and input-output relationships, this study also examined the ipsilateral silent period (iSP) and contralateral and ipsilateral MEPs to provide an assessment of transcallosal and cortico-reticular pathways. The iSP measure provides an index of transcallosal inhibition (Karni et al., 1998; Meyer et al., 1995; Ziemann et al., 1999). In contrast, the ipsilateral MEP (iMEP) is independent of transcallosal pathways (Ziemann et al., 1999). The iMEP pathway to proximal and axial muscles is thought to be mediated by an oligosynaptic cortico-brainstem-spinal pathway (Ziemann et al., 1999) that may include a relay via the reticular formation (Bawa et al., 2004; Wassermann et al., 1994; Ziemann et al., 1999). Thus, the iSP and iMEP measures may provide complementary information about role of cortico-reticular and transcallosal pathways in the expression of dystonia

2. Methods

2.1. Participants

Participants included 21 people with CD (52yrs ± 11.47) and 25 healthy controls (CTL). (50 ± 9.48 years). All participants were screened for contraindications to TMS (Rossi et al., 2009). Exclusion criteria included medications acting on gamma-aminobutyric acid and dopaminergic neurotransmission, implanted devices, history of seizure in the last 2 years, pregnancy, or any neurologic conditions other than dystonia. Participants with CD were recruited from the Physical Medicine and Rehabilitation Neurotoxin Clinic at the University of Minnesota and local dystonia support groups. People with an acquired form of CD or any facial involvement were excluded. Each participant was screened for a diagnosis of idiopathic cervical dystonia with trapezii involvement, as the research goal was to assess an affected muscle group. Trapezii involvement was confirmed in this study using clinic notes regarding initial examination by the treating physician, botulinum toxin procedure notes, and a physical examination using the Toronto Western Spasmodic Torticollis Rating Scale (TWSTRS) by a licensed physical therapist. For participants with bilateral symptoms, activating maneuvers and muscle palpation were additionally used to identify the side of predominant symptoms. Participants with CD taking medications acting on gamma-aminobutyric acid and dopaminergic neurotransmission were instructed to work with their treating physician to wean off medication prior to participating in the study for at least two times the duration of the drug half-life. Participants receiving botulinum toxin injections participated in the study once toxin re-injection was clinically indicated. The timing of re-injection was based on patient report of symptoms (pain, range-of-motion, muscle tone, level of function) that had been deemed by the patient to have returned to pre-toxin baseline. Molecular genetic testing was not completed the CD participants. CTL participants were age- and sex-matched to participants with CD. Hand dominance was self-reported. The protocol was approved by the University of Minnesota’s Institutional Review Board and Clinical Translational Science Institute and conformed to the Declaration of Helsinki. Participants gave written informed consent to participate in the study.

2.2. Transcranial magnetic stimulation

Participants were seated in a semi-reclined chair during cortical excitability testing and calibrated to a frame-less stereotactic neuronavigation system (BrainSight, Rogue Research Inc., Montreal, QC, Canada) to verify consistency of the TMS coil location during testing. Electromyography (EMG) was collected from bilateral upper trapezii muscles with the active electrode placed on the approximate motor point of the upper trapezius fibers and the reference placed 3cm lateral along the orientation of the muscle fibers. The ground electrode (a flexible, conductive fabric strip) was taped vertically along the spinous process, centered at C7. Participants performed an isometric shoulder shrug to activate bilateral upper trapezius muscle during TMS testing. Participants who were unable to generate an EMG signal >20μV with the cued contraction were first given verbal and tactile cues to perform the proper muscle contraction and if needed, they were given a one-pound weight to hold at their side to achieve the desired activation. The cSP response was assessed using a 70-mm figure-of-eight TMS coil connected to a Magstim 200² stimulator (The Magstim Company Ltd., Carmarthenshire, UK). During testing, participants were given breaks after every 15 trials, or if they self-reported fatigue. Real-time EMG was displayed on a screen visible to all participants and all were instructed to maintain a stable contraction level during testing trials. For the primary aim, the cSP threshold was bilaterally determined by observing a 30ms or greater duration cSP in 5 of 10 trials (Cantello et al., 1992; Lo and Fook-Chong, 2005). Next, cSP input-output properties were assessed in each hemisphere using testing intensities of 100%, 110%, 120%, 130%, 140% of the cSP threshold, with 10 trials collected at each intensity. For the secondary aims, the iSP, cMEP and iMEP were extracted from bilateral EMG traces generated during unilateral cSP testing.

2.3. Data Processing

The contralateral responses (cSP and cMEP) were extracted from the EMG channel contralateral to TMS application, while the ipsilateral responses (iSP and iMEP) were extracted from the EMG channel recording ipsilateral to TMS application. cSP and iSP data were rectified and a 10ms moving standard deviation (SD) calculation was constructed to visualize the waveform. The average pre-stimulus SD (from −100ms to −5ms) was used to construct a threshold to determine the offset of each silent period according to previously established methods (Chen et al., 2015). cSP offset was defined as the point that the moving 10ms SD value returned to the mean pre-stimulus level. Duration of cSP and iSP was calculated from the time point of TMS artifact to cSP or iSP offset. MEP traces were rectified and averaged for all trials at each intensity. An iSP was not consistently present across participants at stimulation intensities <140% cSP threshold, thus the average iSP waveform for each participant was derived from a minimum of 5 trials where an iSP was clearly observable at 140% of cSP threshold. The MEP onset and offset latency was manually determined from a rectified mean EMG trace constructed from ten trials using 140% of the silent period threshold. MEP onset and offset were set at each prominent EMG trace deflection rising or falling outside of a three SD threshold, constructed from baseline EMG activity. MEP size was then calculated from each individual trace using this onset and offset point with the following equation: MEP size = MEP area − baseline EMG area where MEP area is the area under the MEP curve and EMG area is the area under the curve for a time-equivalent period of pre-stimulus activity (Bradnam et al., 2010). To assess baseline spinal motor neuron activity, a 30ms window of EMG prior to the TMS pulse (−5 to −35ms) was compared between groups using area under the curve.

2.4. Data Analysis

Normality was evaluated using Shapiro-Wilks test with log transformations completed when appropriate. For all measures within each subject, outliers that were greater than three SD from mean were removed. Group comparisons of binary demographic data were compared with the Fisher’s exact test. For cSP and cMEP data, a mixed model ANOVA were fit using JMP Pro software (JMP® Pro. SAS Institute Inc., Cary, NC, 1989–2007). Group (CD/CTL) and stimulus intensity (100%, 110%, 120%, 130%, 140%) were fixed factors with Subject as a random factor, and a repeated correlation structure (Toplitz) was used. To estimate iSP and iMEP response rate (ratio of participants with an iMEP or iSP response) between groups, the Cochran-Mantel-Haenszel Test (stratified by stimulation level) was used. iMEP and iSP responses were further compared between groups using the maximal stimulus intensity (140%) due to non-random data loss secondary to reduced response rates at lower intensities. Within-group evaluations of hemispheric symmetry were explored with matched pairs analyses. Comparisons between groups on iSP duration and iMEP latency were completed with t-tests and Wilcoxon non-parametric equivalents when appropriate. Graphical data shown with box and whisker plot with median, box at 25^th-75^th percentiles and whiskers indicating range. The relationship of the iSP and TWSTRS sub-scores were explored with the Spearman’s correlation coefficient. Significance was set to p<0.05.

3. Results

All participants completed the study with three subjects excluded from the final analysis. Three subjects were excluded from the analysis due to: insufficient bilateral muscle activation (n=1, CD group), left handedness (n=1, CTL group) and one participant with CD that had marked chorea-like symptoms during the exam and presented with cSP and iSP outlier data (prolonged durations of cSP and iSP), which were defined as data >3 SD. Data reported in the results include 19 participants with CD (age 52.3±11 yrs) and 24 CTL (age 49.4±9 yrs). Five participants had missing data from one hemisphere due to the following reasons: TMS was reported to be too painful at the scalp site so data were not collected bilaterally (CTL n=1), unable to evoke a cSP in the left hemisphere (CTL n=1), TMS machine overheating (CTL n=1), and equipment error/data loss (CD n=2). Neither age (t=0.91, p=0.37) or sex (p=1.00) was significantly different between groups. The demographics of the participants with CD are presented in Table 1.

Table 1.

Demographics of participants with cervical dystonia

Participant	Age	Sex	Sx Side	BoNT (days post)	Total BoNT Dose	Left Trapezii BoNT	Right Trapezii BoNT	Sx (years)	Tremor	TWSTRS Pain	TWSTRS Disability	TWSTRS Motor Score
CD02	41	F	L	69	275	5	5	1.5	Y	13.25	18	5
CD06	54	F	R	75	175	0	25	7	N	0	8	7
CD08	47	M	R	71	500	100	100	10	N	12.25	20	7
CD10	40	F	R	55	200	0	55	4	Y	8.25	10	18
CD11	53	F	R	>365	160	5	5	7	Y	19.5	28	25
CD13	50	F	L	87	125	25	20	25	Y	9.2	6	6
CD14	62	F	R	105	100	5	5	25	Y	1.75	5	12
CD15	55	F	L	84	100	5	5	5	Y	3	6	7
CD17	59	F	R	78	150	10	20	17	Y	9.5	9	13
CD18	63	M	R	70	475	10	20	10	N	6.5	9	16
CD19	66	F	L	86	225	20	0	13	N	13	8	10
CD20	56	M	R	84	300	20	20	28	N	0	1	3
CD21	61	F	L	91	225	30	0	28	Y	4	3	27
CD26	30	F	L	83	200	0	0	16	Y	10.5	19	19
CD31	46	F	L	71	200	10	10	9	N	11.25	12	NA
CD33	64	F	R	83	350	10	30	6	Y	13.5	16	27
CD35	62	F	R	84	425	15	55	4	N	1.5	8	8
CD40	55	F	R	79	250	15	45	6	Y	17.5	20	20
CD43	29	F	R	90	285	0	60	5	N	12.5	16	16

M = male; F = female; Sx = symptoms; BoNT = total botulinum neurotoxin dose in all muscles injected; Hx = history, L = left; R = right; Y = yes; N = no; NA = not available; TWSTRS = Toronto Western Spasmodic Torticollis Rating Scale. Symptom side refers to the predominantly or more affected side of muscle involvement.

EMG levels prior to TMS application were not significantly different between groups during the assessment of either hemisphere (p=0.10–0.49) and not significantly different within groups when comparing between sides (p=0.37–0.78). A representative example of contralateral and ipsilateral EMG responses to TMS is provided in Figure 1 from a control participant.

Figure 1.

Sample bilateral electromyographic responses from a unilateral transcranial magnetic stimulation (TMS) pulse in a single control participant. Stimulation at 140% of contralateral cortical silent period (cSP) threshold. Responses in the contralateral muscle (top trace) produced an easily identifiable contralateral motor evoked potential (cMEP) and cSP. Responses in the muscle ipsilateral to TMS produced an ipsilateral motor evoked potential (iMEP) and silent period (iSP). Ipsilateral responses were observed most frequently at 130–140% of cSP threshold.

3.1. Intracortical Inhibition (cSP)

The cSP threshold, as a measure of intracortical inhibitory pathway excitability, was not significantly different between groups (right hemisphere: t_38.4 = 0.84, p = 0.41, left hemisphere: t₃₃ = 0.92, p = 0.37). The mean ± standard deviation stimulator output corresponding to cSP threshold was 61.8±9.9% maximal stimulus output (MSO) for left, and 58.2±10.4% MSO for right hemisphere stimulation in the CTL group, and 64.8±10.7% MSO for left and 60.47±7.9% MSO for right hemisphere stimulation in the CD group.

The mixed model was performed separately for each hemisphere to explore group differences. In each mixed model, Intensity was a significant factor for cSP duration (p<0.001), while there was no significant Group or Group*Intensity effect (Table 2 and Figure 2). The CD group demonstrated consistently shorter durations of cSP across all intensities on average.

Table 2.

Mixed model results

Outcome		Effect	df	F	p
cSP duration	Left Hemisphere	Group	44.8	3.31	0.076
		Intensity	79.7	25.7	<0.001*
		Group*Intensity	79.7	0.69	0.6
	Right Hemisphere	Group	53.7	1.9	0.17
		Intensity	81.9	26.27	<0.001*
		Group*Intensity	81.9	1.38	0.25
iSP duration		Group	38.9	1.96	0.17
		Hemisphere	34.3	3.45	0.072
		Group*Hemisphere	34.3	4.22	0.048*
cMEP size	Left Hemisphere	Group	39.3	0.32	0.58
		Intensity	78.9	51.62	<0.001*
		Group*Intensity	78.9	1.05	0.38
	Right Hemisphere	Group	36.2	0.10	0.75
		Intensity	78.3	0.76	<0.001*
		Group*Intensity	48.3	0.08	0.56
iMEP size		Group	35.9	0.007	0.93
		Hemisphere	28.7	5.41	0.027*
		Group*Hemisphere	28.7	5.0	0.033*

cSP = contralateral cortical silent period; iSP = ipsilateral cortical silent period; cMEP = contralateral motor evoked potential; iMEP = ipsilateral motor evoked potential, df = degrees of freedom. Note: the factor of Hemisphere is not present for the contralateral data (cSP and cMEP) because separate mixed models were used to evaluate within hemisphere effects of stimulation intensity. Ipsilateral data (iSP and iMEP) do not have the factor of Intensity due to missing data at low intensities.

Figure 2.

Contralateral silent period (cSP) from left (A) and right (B) hemisphere stimulation. No statistical differences between groups, but consistency for cervical dystonia (CD) group to have reduced cSP duration on average. cSP duration measured in milliseconds (ms). CTL = healthy control; CD = cervical dystonia.

3.2. Transcallosal Inhibition (iSP)

An iSP was not detected in one hemisphere in six subjects (three CD and three CTL). The incidence of iSP observations was similar between groups from left (χ²=0.021, p=0.89) and right (χ²=2.63, p=0.10) hemisphere stimulation, suggesting similar response rate between groups. The mixed model for the iSP response revealed that there was a significant Group*Hemisphere interaction (p=0.048), but no Group or Hemisphere effect (Table 2). The iSP was not significantly different between hemispheres in the CTL group (t₁₈ = 0.11, p = 0.91). In the CD group, the iSP from left hemisphere stimulation was significantly shorter compared to the right hemisphere (t₁₆ = 3.25, p = 0.005), suggesting asymmetrically reduced transcallosal inhibition in CD (Figure 3).

Figure 3.

Ipsilateral silent period (iSP) duration between groups and hemispheres. Control (CTL) group demonstrated hemispheric symmetry, while cervical dystonia (CD) group show similar transcallosal inhibition compared to CTL in the right hemisphere but reduced transcallosal inhibition in the left hemisphere. Single subject data as single data points, Left hemisphere N=37, right hemisphere, N=40. *p<0.05. iSP duration measured in milliseconds (ms). CTL = healthy control; CD = cervical dystonia.

To explore if side of symptoms could explain the left hemisphere asymmetry noted in iSP, CD participant iSP duration by side of predominant dystonic symptoms was evaluated. iSP duration was similar between the more affected and less affected muscles (t₁₆ = 0.8, p = 0.42), suggesting the reduced asymmetry of transcallosal inhibition was not explained by the side of symptoms. To explore if there was a relationship between symptom severity and left iSP duration, correlational strength was assessed. There was no significant relationship between TWSTRS sub-scores and left iSP duration (Spearman’s ρ=0.10–0.16, p=0.53–0.67).

3.3. Corticomotor Excitability (MEPs)

cMEPs were observed in all participants. There was a significant effect of TMS Intensity (p<0.001), but no significant effect of Group and no Group*Intensity interaction on MEP size (Table 2). cMEP onset latency (CTL left: 9.3±1.5, right: 9.3±1.5; CD left: 9.6±1.5, right: 8.8±1) was similar between groups and between hemispheres (all comparisons p > 0.10).

iMEPs were not observed in all participants. The CTL group had iMEPs in 18/23 (78%) participants from left hemisphere stimulation and 17/22 (77%) participants from right hemisphere stimulation. The CD group had iMEPs from 8/17 (47%) participants from left hemisphere stimulation and 13/19 (68%) from right hemisphere stimulation. There was a trend for reduced left iMEP response rate from left hemisphere stimulation in the CD group compared to CTL (Fisher’s Exact Test, p=0.053), but no significant difference for iMEPs generated from right hemisphere stimulation (Fisher’s Exact Test, p=0.73).

A mixed model analysis of iMEP size indicated a significant effect of Hemisphere (p=0.027) and a significant Group*Hemisphere interaction (p=0.033), but no significant effect of Group (Table 2). The control group demonstrated hemispheric symmetry in iMEP size (t₁₀ = 0.009, p = 0.99). Figure 4 illustrates that iMEPs from left hemisphere stimulation were reduced in magnitude compared to the right hemisphere in the CD group, but this difference was not significant (t₅= 1.83, p = 0.12), likely due to the high variance in the right iMEPs and low number of participants with bilateral iMEP responses.

Figure 4.

Ipsilateral motor evoked potential (iMEP) from left and right hemisphere stimulation. Control group (CTL, black) show symmetry between hemispheres, while cervical dystonia (CD, grey) show reduced response in the left hemisphere compared to CTL. *p=0.026. iMEP size measured in millivolts (mV) per milliseconds (ms). CTL = healthy control; CD = cervical dystonia.

In those participants with measurable iMEP, the onset latency was similar between groups from left (CTL: 18.8±6ms; CD: 18.8±4.9ms; Wilcoxon, Z=0.50, p=0.62) and right (CTL: 16.7±5.6ms; CD: 16.9±3.2ms; Wilcoxon, Z=0.57, p=0.57) hemisphere stimulation, indicating similar conduction times along the ipsilateral descending pathways between groups.

4. Discussion

The primary finding of this work was that the recruitment threshold and input-output properties of the cSP were similar between people with CD and age-matched controls. This finding suggests that the excitability of cortico-motor GABAergic inhibitory mechanisms related to the affected upper trapezius muscle are not dysregulated in people with CD. However, it is important to consider that the primary phenomenon of shortened cSP (using traditional cSP testing methods), which has been reported in CD, was not evaluated in this study. The novel methods employed here (i. e. cSP recruitment threshold testing) may be a solution to reduce uncertainty between studies by more precisely assessing inhibitory pathway excitability. The finding of unaffected cSP recruitment threshold and input-output properties in an affected dystonic muscle is provocative because it challenges the commonly held view that reduced or absent GABAergic inhibition is characteristic of idiopathic focal dystonia. The methods employed in this study may provide more precise approach to quantify GABAergic tone and lead to greater agreement between studies.

The results also demonstrated hemispheric symmetry in all outcomes for participants in the control group, while the responses in the people with CD were characterized by asymmetric ipsilateral inhibitory responses. This finding may indicate dysregulation of transcallosal pathways in people with CD.

4.1. Implications for cSP interpretation

Despite a comparable recruitment threshold for evoking the cSP response and input-output properties, the duration of the cSP was consistently shorter in the CD group but did not reach statistical significance (Figure 2). This finding was not anticipated because several previous studies have shown significantly reduced cSP duration in people with CD, specifically in affected neck musculature (Amadio et al., 2000; Cakmur et al., 2004) and unaffected hand muscles (Odorfer et al., 2019; Quartarone et al., 2008). Yet, there is disagreement between studies as two studies have concluded normative cSP durations in affected neck musculature (Odergren et al., 1997) and non-affected hand musculature (Cakmur et al., 2004) in people with CD. Comparing cSP results between the previous studies in CD and our work is challenging because each study utilized different testing methods based on a traditional suprathreshold stimulus approach. Other methods between past studies also differ based on the coil configuration used (Figure-8, circular, double-cone), the cortical location stimulated (Cz, motor cortex, midline between Cz and external auditory meatus), or the stimuli intensity during testing (100% MSO or %RMT). Differing methods to assess the cSP may be a key source of disagreement between studies evaluating the cSP in focal dystonia. The methods utilized here (i. e. cSP recruitment threshold testing) may be a solution to reduce uncertainty between studies by more precisely assessing inhibitory pathway excitability.

In the present study, the cSP recruitment threshold was defined as the MSO intensity that produced a suppression of EMG activity of at least 30ms duration in 5/10 trials during a tonic contraction. The suppression of EMG (i.e. cSP) was typically observed in the absence of a cMEP, suggesting that TMS may activate low-threshold intracortical inhibitory interneurons without activating excitatory interneurons that generate the cMEP. The relationship between TMS intensity and the threshold for excitatory (i.e. MEP) or inhibitory (i.e. cSP) responses is not well defined and may vary depending on the level of corticospinal input to the target muscle (Cantello et al., 1992; van Kuijk et al., 2009). However, observation of a cSP without a preceding MEP is consistent with previous research that has shown TMS responses to show initial suppression of tonic muscle activity at low intensities, reflecting the putative activation of low threshold cortical inhibitory interneurons, followed by a net facilitation of EMG activity reflecting the progressive activation of excitatory interneurons networks and output from the corticospinal neurons (Klimpe et al., 2009; MacKinnon et al., 2005; Werhahn et al., 2007).

With regard to cSP testing in hyperkinetic movement disorders, traditional approaches used to estimate the cSP may be inadequate because abnormal or excessive excitatory drive may be characteristic to the disorder and influence the cSP response. For example, both the input-output properties of the cMEP and brainstem reflexes have been found to be excessively facilitated in people with CD (Amadio et al., 2000; Antelmi et al., 2016). Elevated excitatory drive may lead to the use of lower testing stimulus intensities due to an enhanced size of the MEP or may influence the resumption of EMG activity. Further, the testing stimulus may need to be established during an active contraction to account for differences in spinal excitability and neuromuscular drive. Thus, the use of testing intensities calibrated to the cSP response itself may offer an advantageous method to study inhibitory responses while reducing the influence of excitatory components (Kallioniemi et al., 2014; Werhahn et al., 2007).

Extending this methodology of cSP threshold or cSP stimulus-response curve testing to other dystonias or muscular targets may help to elucidate which aspects of cortical inhibition are dysregulated in idiopathic focal dystonia. Although our findings indicate unaffected recruitment threshold and input-output properties in a cohort with CD, it remains unknown if the maximal cSP duration is affected. This could be an important task for future study because testing the cSP at various points along the stimulus-response curve may unveil if GABA_A or GABA_B medicated pathways are predominantly affected in models of dystonia. It has been suggested that cSPs evoked at higher intensities (at the top of the stimulus-response curve) are predominantly mediated by GABA_B, while cSPs evoked at low intensities reflect predominantly GABA_A (Kimiskidis et al., 2006; Rossini et al., 2015; Ziemann et al., 1996). It is possible that previous cSP studies in dystonia may have tested the cSP in a higher range of stimulus intensities (near the maximal cSP duration), reflecting a loss of GABA_B inhibition that is sensitive to high intensity stimulus intensities. In contrast, the results here were tested near the bottom of the curve and may indicate normal recruitment of GABA_A pathways that are sensitive to low stimulus intensities.

This study provides insight into the excitability of GABAergic inhibitory pathways in people with dystonia, but the results may be unique to axial musculature or CD. The upper trapezius is an axial muscle that receives extensive cortico-reticular and cortico-bulbar input (Strominger et al., 2012) and therefore the cortical contributions to the control of tonic muscle contractions may be very different from the more frequently tested distal arm muscles such as the first dorsal interosseous. Thus, our conclusions may be specific to those with idiopathic CD affecting axial muscles. An additional limitation in this study was that we could not assess the stimulus-response curve fully to evaluate the maximal cSP duration. Given the high cSP threshold of the UT muscle target (~58–65% MSO), a plateau in cSP duration could not be estimated before reaching the maximal output of the TMS device. Future studies that implement a cSP stimulus-response curve protocol (Kimiskidis et al., 2005) may need to consider the feasibility of reaching the plateau in cSP duration and choose a muscular target accordingly.

Because an affected muscle was targeted in this study, it should be noted that a history of botulinum toxin injections may influence our cSP findings. Botulinum toxin may have lingering central effects, possibly changing measures of corticobulbar excitability or reducing the size of the MEP. However, TMS measures such as motor thresholds, cSP duration, intracortical facilitation and input-output curves have not been shown to modulate following botulinum toxin injection in people with dystonia thus these concerns are lessened (Allam et al., 2005; Boroojerdi et al., 2003; Kojovic et al., 2011). Additionally, participants in this study were evaluated once botulinum clinical effectiveness had declined.

4.2. Transcallosal inhibition is asymmetrical in CD

It has been demonstrated that transcallosal inhibition, as measured by the iSP outcome, is symmetric in healthy populations independent of handedness (Davidson and Tremblay 2013b, 2013a; Kuo et al. 2019). Factors that have been shown to significantly influence transcallosal inhibition are age, hand dexterity, and performance features related to timing (Davidson and Tremblay 2013a, 2013b; Kuo et al. 2019), which may indicate that aging, coordination and skill may play an important role in transcallosal regulation.

The finding of asymmetric transcallosal inhibition in people with CD may indicate interhemispheric dysregulation. However, there was no association between dystonia symptom presentation (by side or severity) and the iSP duration. In this study, all CD participants were right-handed and left hemisphere iSP abnormalities were noted, suggesting a possible role of hemispheric dominance on iSP dysregulation. We are unable to confirm if hemispheric dominance may explain our findings of asymmetric transcallosal inhibition in CD due to the lack of left-handed individuals.

Only one previous study has evaluated the iSP in people with dystonia (Niehaus et al., 2001). That study assessed individuals with focal hand dystonia and found a significantly increased duration of transcallosal inhibition (Niehaus et al., 2001), which is in contrast to findings reported here. These opposing findings might be related to differences in the TMS protocol used, the target muscle or the dystonia phenotype. Complementary evidence of disordered and reduced transcallosal connectivity comes from studies using the interhemispheric inhibition (IHI) paired pulse TMS protocol. These studies have shown that transcallosal inhibition is reduced in people with focal hand dystonia (Bäumer et al., 2016; Beck et al., 2009; Nelson et al., 2010). Similar to the finding of asymmetry in the transcallosal connections, Nelson et al. (2010) reported reduced left to right transcallosal inhibition in a group of people with focal hand dystonia compared to controls, while right to left transcallosal tone was unaffected (Nelson et al., 2010). Although IHI and iSP do not measure the same aspect of transcallosal inhibition (Chen et al., 2003), the studies of reduced IHI in dystonia are in agreement with the current findings. Continued evaluations of the iSP and IHI in people with various phenotypes of dystonia and across different muscles (e.g. appendicular vs. axial) are warranted to better characterize transcallosal abnormalities and determine clinical implications.

4.3. Reduced iMEP response in CD

In keeping with previous studies of iMEPs in healthy adults (Bawa et al., 2004), the latency of the iMEPs in the trapezius muscle were markedly prolonged compared with the cMEPs (latency differences of 8–10 ms). The pathways mediating ipsilateral MEPs are poorly understood. The balance of evidence suggests that iMEPs are mediated by polysynaptic inputs from corticoreticulospinal pathways (Bawa et al., 2004; Chen et al., 2003; Wassermann et al., 1994; Yeo et al., 2012; Ziemann et al., 1999). In the present study iMEP conduction times were not significantly different between groups, but the size and response rate of the left iMEP was reduced in people with CD compared to controls. These results may indicate asymmetric and downregulated cortico-reticular excitability in people with CD.

5. Conclusion

The recruitment threshold and input-output properties of cortico-motor GABAergic inhibitory pathways mediating the cSP in the upper trapezius muscle were not significantly altered in people with CD compared with healthy adults. This work is the first to examine individualized excitability of inhibitory pathways that mediate the cSP response in people with CD; however, it remains to be seen if normative cSP thresholds can be reproduced when the primary phenomenon of reduced cSP is demonstrated. It also remains unknown if normative cSP excitability is present across dystonic phenotypes and muscles.

This study presents novel evidence for normative intracortical inhibition excitability, asymmetrically reduced transcallosal inhibition, and asymmetric ipsilaterally descending cortical responses in people with CD. The asymmetry in outcomes were not explained by stimulation intensity, laterality of symptoms, or altered cMEP excitability. One possible explanation for the hemispheric asymmetry could be due to hemispheric dominance, yet we are unable to explore this possibility without left-handed individuals. Future studies of cSP in dystonia may benefit from the use of testing intensities calibrated to the cSP response in effort to reduce possible confounding effects from excitatory mechanisms and increase agreement between studies.