Frequency and Intensity of Premonitory Urges-to-Tic in Tourette Syndrome Is Associated With Supplementary Motor Area GABA+ Levels

Abstract

Background:

Individuals with Tourette syndrome (TS) often report that they express tics as a means of alleviating the experience of unpleasant sensations. These sensations are perceived as an urge to act and are referred to as premonitory urges. Premonitory urges have been the focus of recent efforts to develop interventions to reduce tic expression in those with TS.

Objective:

The aim of this study was to examine the contribution of brain γ-aminobutyric acid (GABA) and glutamate levels of the right primary sensorimotor cortex (SM1), supplementary motor area (SMA), and insular cortex (insula) to tic and urge severity in children with TS.

Methods:

Edited magnetic resonance spectroscopy was used to assess GABA+ (GABA + macromolecules) and Glx (glutamate + glutamine) of the right SM1, SMA, and insula in 68 children with TS (M_Age = 10.59, SD_Age = 1.33) and 41 typically developing control subjects (M_Age = 10.26, SD_Age = 2.21). We first compared GABA+ and Glx levels of these brain regions between groups. We then explored the association between regional GABA+ and Glx levels with urge and tic severity.

Results:

GABA+ and Glx of the right SM1, SMA, and insula were comparable between the children with TS and typically developing control subjects. In children with TS, lower levels of SMA GABA+ were associated with more severe and more frequent premonitory urges. Neither GABA+ nor Glx levels were associated with tic severity.

Conclusions:

These results broadly support the role of GABAergic neurotransmission within the SMA in the experience of premonitory urges in children with TS.

Individuals with Tourette syndrome (TS) report that their tics are often preceded by a feeling of mounting inner tension that can only be relieved through the expression of a tic.¹ This sensory phenomenon is referred to as a “premonitory urge.” Premonitory urges are experienced as either generalized or localized tension and are sometimes described as a tactile sensation or a “tickle”^2,3. Although premonitory urges are related to tics, the nature of their relationship has not been fully established. Some have argued that because not all individuals with TS experience premonitory urges, tics cannot be caused by premonitory urges, and there is some evidence suggesting that the experience of premonitory urges is heightened when tics are suppressed.⁴ Indeed, only ~24% of children aged 8–10 years⁵ and 57% of adolescents aged 15–19 years⁵ report experiencing premonitory urges. However, by adulthood, up to 98% individuals with TS report experiencing premonitory urges,^6–8 suggesting that low estimates of premonitory urges experienced in TS may be partially because of difficulties describing premonitory urges in early childhood and adolesence.⁴ Importantly, if premonitory urges do indeed cause tics, understanding the underlying neurobiology of premonitory urges may help inform the development of interventions for alleviating tics in TS.

Much of the existing work toward understanding the neurobiology of premonitory urges has come from functional neuroimaging studies. Functional magnetic resonance imaging studies have shown that before the expression of tics, at a time that would be concurrent with the experience of a premonitory urge, there is an increase in the activation of the anterior cingulate cortex, primary sensorimotor cortex (SM1), supplementary motor area (SMA), insular cortex (insula), and parietal operculum in individuals with TS.^9–13 Studies using magnetoencephalography^14,15 and position emission topography^16,17 report similar findings. Although these studies suggest premonitory urges to be associated with these brain regions, premonitory urges could not be directly measured in these studies, meaning that activation of these brain regions could be caused by other behaviors, such as tic suppression. However, these findings are still sufficiently robust to warrant investigation of these neocortical and paralimbic brain regions using other modalities.

Although functional neuroimaging studies can localize the brain regions likely to be involved in the experience of premonitory urges, a deeper mechanistic understanding of how these brain regions contribute to the experience of premonitory urges and tics requires the consideration of metabolite profiles and neurotransmission. Historically, investigations into the pathophysiology of TS have focused on the dopaminergic system.¹⁸ Despite there being many individual accounts of altered dopaminergic functioning in TS, the evidence in support of dopamine hypothesis of TS is mixed.¹⁹ An alternative hypothesis proposes that the symptoms of TS may be explained by an imbalance of inhibitory and excitatory neurotransmission. Specifically, altered phasic GABAergic neurotransmission and/or reduced tonic levels of γ-aminobutyric acid (GABA) in higher-order motor areas has been suggested to result in a failure of control over the gain of motor excitability and tic expression in TS.²⁰ In support of this alternative hypothesis, reduced short-interval intracortical inhibition (a marker of inhibitory motor control and GABA_Aergic neurotransmission) in left motor cortex (M1) has been identified in both children and adults with TS through paired-pulse transcranial magnetic stimulation (TMS) studies, although the stronger associations in children are with co-occurring ADHD.^21–24 A recent pilot study using double-coil TMS reported adolescents with TS have diminished inhibitory signaling from right SMA to left M1.²⁵ Reduced SM1²⁶ and ACC²⁷ GABA levels have also been identified in individuals with TS through the application of magnetic resonance spectroscopy (MRS). These findings align well with the body of animal studies showing that striatal injection of GABA_a antagonists, such as bicuculine and picrotoxin, can produce tic-like behaviors in both rodents^28,29 and primates.³⁰ Interestingly, reducing glutamatergic striatal innervation appears to reduce tic-like behaviors in rodents,²⁹ supporting the idea that the symptoms of TS may manifest out of an imbalance of excitatory and inhibitory signaling.

In this study, we used edited proton MRS to measure and compare both GABA and Glx (the combination of glutamate + glutamine because glutamate cannot be fully resolved at 3 Tesla [T]) levels as markers of inhibition and excitation in right SMA, right SM1, and right insula in children with and without TS. These three brain regions were selected based on their associations in prior TS imaging research with urges and tics. Concurrent with the period in which this study took place, other investigators have published studies on GABA and glutamate, with heterogeneous findings (see Mahone and colleagues³¹). Our specific objective was to characterize group differences in neurotransmitter levels in affected children and to quantify associations with tic and premonitory urge severity.

Subjects and Methods

Participants

Children with TS were recruited from specialty clinics at the Kennedy Krieger Institute and Johns Hopkins Hospital (the Pediatric Neurology Movement Disorders, OCD, and TS Specialty Clinics) and at Cincinnati Children’s Hospital Medical Center (Pediatric Neurology TS Clinic). Typically developing children (TDCs) were recruited through community advertisements as previously described.³² The final sample consisted of 116 children (44 Kennedy Krieger Institute, 72 Cincinnati Children’s Hospital Medical Center), with 74 children in the TS group and 42 children in the TDC control subjects group.

Diagnoses of TS, attention deficit hyperactivity disorder (ADHD), and obsessive-compulsive disorder (OCD) were based on Diagnostic and Statistical Manual of Mental Disorders, Fifth Edition³³ criteria. After a standard telephone screen, the diagnostic process began with a structured diagnostic interview using the Kiddie Schedule for Affective Disorders and Schizophrenia³⁴ to screen for excluded diagnoses, such as major depression, and a review of all available electronic health record clinical information, including, if available, any psychological testing. Symptom severity was assessed with standard clinical rating scales for tics (Yale Global Tic Severity Scale [YGTSS]³⁵), tic urges (Individualized Premonitory Urges in Tourette Scale [I-PUTS]³⁶), ADHD (Conners 3rd Edition ADHD Scale³⁷ and DuPaul’s ADHD rating scale³⁸), and OCD (Child Yale-Brown Obsessive Compulsive Scale³⁹). Assessment of TDCs involved the same process, except that scales for tics and OCD were not performed. Board-certified pediatric neurologists (S.H.M. and D.L.G.) with more than 20 years of clinical and research experience with TS and co-occurring conditions confirmed all diagnoses and clinical ratings.

Children were excluded if they had a presence or history of (1) a neurological disorder other than TS, including seizures, tumor, head injury, or stroke; (2) a severe chronic medical disorder; (3) presence of a major visual impairment; (4) substance abuse or dependency; or (5) any psychiatric (including autism spectrum disorder) or developmental disorders other than ADHD or OCD. Finally, both TS and TDC subjects had a full-scale intelligence quotient of >80 as measured by the Wechsler Intelligence Scale for Children-V.⁴⁰

Descriptive statistics for relevant demographic variables, including age, sex, handedness, prescribed medication, scores on the YGTSS, I-PUTS, Child Yale-Brown Obsessive and Compulsive Scale (CYBOCS), as well as GABA+ and Glx levels for each of the regions of interest (ROIs; see later), are available in Table 1.

TABLE 1.

Descriptive statistics

	TS			TDC
	n	M	SD	n	M	SD	P
Demographics
Age (y)	74	10.64	1.26	42	10.4	1.95	0.482
Sex (Female:Male)	14:60	–	–	17:25	–	–	0.0213
Handedness (Left:Right)	4:70	–	–	1:41	–	–	0.7678
Co-occurring diagnoses
ADHD	35:39	–	–	1:40	–	–	0
OCD	12:62	–	–	0:42	–	–	0.0147
Medication prescribed
ADHD stimulant (No:Yes)a	59:15	–	–	41:1	–	–	0.0162
ADHD nonstimulant (No:Yes)	73:1	–	–	42:0	–	–	>0.99
Dopamine receptor blockers (No:Yes)	69:5	–	–	42:0	–	–	0.213
Antiseizure drugs (No:Yes)	71:3	–	–	42:0	–	–	0.476
α2-Adrenergic agonists (No:Yes)	60:14	–	–	42:0	–	–	0.011
SSRI (No:Yes)	64:10	–	–	42:0	–	–	0.032
Clinical rating scale scores: TS
YGTSS
Total motor	74	13.76	11.29	42	–	–	–
Total phonic	74	9.44	5.97	42	–	–	–
Total impairment	74	14.30	13.60	42	–	–	–
Global severity	74	36.19	20.62	42	–	–	–
I-PUTS
Number of tics	74	7.44	5.95	42	–	–	–
Number of urges	74	4.57	5.53	42	–	–	–
Urge frequency	74	12.19	15.45	42	–	–	–
Urge intensity	74	11.06	14.93	42	–	–	–
Clinical rating scale scores: ADHD
Conners inattention	74	66.36	15.28	42	50.80	10.22	<0.01
Conners hyperactivity	74	69.31	16.06	42	49.96	9.42	<0.01
DuPaul’s inattention	74	12.57	7.74	42	4.14	4.78	<0.01
DuPaul’s hyperactivity	74	10.47	7.32	42	3.31	4.74	<0.01
Clinical rating scale scores: OCD
CYBOCS
Obsessions	74	2.87	4.22	42	0	–	–
Compulsions	74	2.97	4.52	42	0	–	–
Total	74	5.85	8.24	42	0	–	–
MRS of GABA and Glx
SM GABA (IU)	74	2.79	0.43	42	2.76	0.30	–b
SMA GABA (IU)	74	2.48	0.40	42	2.56	0.47	–b
Ins GABA (IU)	74	2.64	0.50	42	2.66	0.58	–b
SM Glx (IU)	74	6.10	1.45	42	5.91	1.58	–b
SMA Glx (IU)	74	6.58	1.92	42	6.50	1.67	–b
Ins Glx (IU)	74	6.87	1.78	42	6.68	2.01	–b
SM GABA/Cr	74	0.10	0.01	42	0.10	0.01	–b
SMA GABA/Cr	74	0.10	0.02	42	0.10	0.02	–b
Ins GABA/Cr	74	0.10	0.02	42	0.10	0.02	–b
SM Glx/Cr	74	0.06	0.01	42	0.06	0.02	–b
SMA Glx/Cr	74	0.07	0.02	42	0.08	0.02	–b
Ins Glx/Cr	74	0.07	0.02	42	0.08	0.02	–b

Continuous variables were compared using Welch’s t test. Categorical variables were compared using a Pearson’s χ² test with Yates’ continuity correction.

TS, Tourette syndrome; TDC, typically developing children; SD, standard deviation; GABA+, γ-aminobutyric acid + macromolecules; ADHD, attention deficit hyperactivity disorder; OCD, obsessive-compulsive disorder; SSRI, selective serotonin reuptake inhibitor; YGTSS, Yale Global Tic Severity Scale; I-PUTS, Individualized Premonitory Urges in Tourette Scale; CYBOCS, Child Yale-Brown Obsessive Compulsive Scale; MRS, magnetic resonance spectroscopy; Glx, glutamate + glutamine; SM1, primary sensorimotor cortex; IU, institutional units; SMA, supplementary motor area; Ins, insular cortex; Cr, creatine.

^aStimulant medication temporarily discontinued day before study visit.

^bCompared using general linear model with covariates.

MRS

Before MRI scanning, children underwent a mock scan with a research assistant to improve their comfort and decrease their anxiety. To motivate children to lie still during the scan, the children were allowed to watch movies during the acquisition and received “points” for lying still. These points could be exchanged for prizes after the scan. MRI and MRS scans were conducted using Phillips 3T Achieva scanners (Best, the Netherlands; 32-channel head coil for receive and body coil for transmit) at both sites using identical examination cards for the pulse sequences. One person (N.A.J.P.) traveled to both sites to ensure cross-site consistency and train in voxel placement. For voxel placement and segmentation, a 1-mm³ isotropic Tl-weighted MP-RAGE (magnetization-prepared rapid gradient-echo) image was acquired first (repetition time = 7.99 ms, echo time = 3.76 ms, flip angle = 8°). Voxels of 3 × 3 × 3 cm³ were then positioned in three ROIs: the right SMA, right SM1, and right insula (see Fig. 1). We chose to examine right hemisphere voxels because of previous work having suggested that the right rather than left insula was specifically involved in the experience of the premonitory urges-to-tic and to correspond with a separate study of tactile function, which has traditionally been examined in the left hand. Full details of the MRS data acquisition can be found in the Supporting Information (Section 0). The results presented in this article are referenced to water and are presented in institutional units. (When MRS is quantified relative to unsuppressed water signal, it is typically accounting for metabolite- and compartment-specific relaxation values, as well as editing efficiency, approximating concentration values in the millimolar range. However, because this is not a direct comparison with millimolar values, it is typically presented as institutional units.) There were no significant group differences on any of the quality assurance metrics (see Supporting Information Table S1).

FIG. 1.

Voxel placement and resultant spectra. Voxel placement in SM1 (A) and accompanying spectra (B). Voxel placement in SMA (C) and accompanying spectra (D). Voxel placement in insula (E) and accompanying spectra (F). GABA+, GABA + macromolecules; Glx, glutamate + glutamine; Insula, insular cortex; ppm, parts per million; SM1, primary sensorimotor cortex; SMA, supplementary motor area.

Statistical Analyses

All statistical analyses were conducted using the R programming language (v4.0.3) in Rstudio (v1.3.1093). The code used to generate the results and figures of this manuscript are available online through the Open Science Framework (https://osf.io/e7yw9/). The α level for all analyses was set to 0.05. For group comparisons using analysis of variance, partial eta-squared (η²_p) was estimated using the “effect size” package,⁴¹ and Bayes factors were estimated using the “BayesFactor” package.⁴² The evidence for H1 against H0 is denoted as BF₁₀. BF₁₀ was estimated for each variable of interest by comparing the BF₁₀ of the model with the variable of interest included against the model without the variable of interest included. Similarly, to determine the BF₁₀ of interaction effects, the BF₁₀ of the model with the interaction term included was compared against the model without the interaction term included.

First, after controlling for age and sex, a 3 × 2 analysis of variance was conducted to determine whether there were differences with regard to GABA+ and Glx levels between our ROIs (SMA, SM1, and insula) and groups (TS and TDC). Where there was evidence of an interaction effect, which was determined through joint consideration of P values, η²_p, and Bayes factors, simple main effect analyses were conducted. If there was evidence of a main effect, post hoc comparisons were conducted, and Tukey’s honestly significant difference was used to control for inflation of Type I error. Second, to determine whether individual differences in GABA+ and Glx levels were related to urge and tic severity, Pearson’s correlation analyses were conducted between metabolite concentration levels of each ROI and scores on the PUTS and YGTSS. Due to the large number of correlations being conducted in this study, the more conservative Bonferroni’s method was used to correct for inflation of Type I error. Additional analyses where co-occurring ADHD and/or OCD are considered are presented in the Supporting Information (Section 0.5).

Results

Comparison of GABA+ and Glx Levels Between Regions and Groups

GABA

After controlling for age [F(1, 299) = 0.22, P = 0.639; η²p = 0.00; BF₁₀ = 0.14] and sex [F(1, 308) = 0.17, P = 0.681; η²p = 0.00; BF₁₀ = 0.16], there was a significant main effect of region [F(2, 299) = 9.16, P < 0.001; η²p = 0.06; BF₁₀ = 164.30), but not group, on GABA+ levels [F(1, 299) = 0.13, P = 0.724; η²p = 0.00; BF₁₀ = 0.13]. There was no significant region by group interaction effect [F(2, 299) = 0.35, P = 0.705; η²p = 0.00; BF₁₀ = 0.09]. Together, the results of these analyses suggest that although GABA+ levels were different across the three regions, GABA+ levels across these regions were comparable between children in the TS and TDC groups. Subsequent post hoc comparisons showed that GABA+ levels were higher in the SM than SMA (P_Tukey < 0.001), and that GABA+ levels were otherwise comparable between SM and insula (P_Tukey = 0.089) and SMA and insula (P_Tukey = 0.088) (see Fig. 2A). Although there was no significant main effect of group, we visualized the post hoc group comparisons across the regions in Fig. 2B–D.

FIG. 2.

Comparing GABA+ and Glx across regions and between groups. Analyses were conducted on (A–D) GABA+ and (E–H) Glx. (A) Group x region on GABA+ analysis. There were differences in GABA+ between regions, but no differences of GABA+ between groups. The latter was confirmed by subsequent post hoc comparisons showing comparable GABA+ between TDC and TS in the (B) SM1, (C) SMA, and (D) insula. (D) Group x region on Glx analysis. For Glx, again there were significant differences between the regions (E), but no significant differences between the groups in terms of (F) SM1, (G) SMA, and (H) insula Glx. Error bars in (A) and (E) represent standard error of the mean. Values on the y axis have been corrected for age and sex. GABA+, GABA + macromolecules; Glx, glutamate + glutamine; Insula, insular cortex; IU, institutional units; SM1, primary sensorimotor cortex; SMA, supplementary motor area; TDC, typically developing control subjects; TS, Tourette syndrome.

Glx

After controlling for age [F(1, 299) = 4.95, P = 0.027; η²p = 0.02; BF₁₀ = 1.24] and sex [F(1, 299) = 3.83, P = 0.051; η²p = 0.01; BF10 = 0.89], there was a significant main effect of region [F(2, 299) = 5.24, P = 0.006; η²p = 0.03; BF₁₀ = 4.31], but no significant main effect of group on Glx levels [F(1, 299) = 0.03, P = 0.973; η²p = 0.00; BF₁₀ = 0.15]. There was no significant region by group interaction either [F(2, 299) = 0.03, P = 0.973; η²p = 0.00; BF10 = 0.06]. Post hoc comparison showed that Glx levels were higher in SMA than in SM1, although this effect fell shy of significance (pTukey = 0.074). There were no significant differences in Glx levels between insula and SMA (pTukey = 0.575), but Glx levels were higher in insula than in SM (P_Tukey = 0.005) (see Fig. 2E). Again, although there was no significant main effect of group, we visualized the post hoc group comparisons across the regions in Fig. 2G,H.

Associations Between Brain Metabolite Levels and TS Symptom Severity

A heatmap of the associations between brain metabolite levels (ie, GABA+ and Glx) and TS symptom severity (ie, urges and tics as measured using the I-PUTS and YGTSS, respectively) is presented in Fig. 3A.

FIG. 3.

Associations of urge and tic with metabolite levels. (A) Heatmap depicting the associations between GABA+ and Glx across the regions of interest on the x axis, with urge and tic severity on the y axis. Scatterplots depicting the associations between SMA GABA+ and the number (B), frequency (C), and intensity (D) of urges as assessed by the I-PUTS. Note that although we corrected for multiple comparisons using the Bonferroni method in the main text, the P values presented in (B)–(D) are uncorrected. Values on the y axis have been corrected for age and sex. GABA+, GABA + macromolecules; Glx, glutamate + glutamine; Insula, insular cortex; I-PUTS, Individualized Premonitory Urges in Tourette Scale; IU, institutional units; SM1, primary sensorimotor cortex; SMA, supplementary motor area; YGTSS, Yale Global Tic Severity Scale.

Urges

Among children with TS, lower levels of SMA GABA+ levels were correlated with number of urges (r = −0.40, P_Bonferroni = 0.020; Fig. 3B), urge frequency (r = −0.39, pBonferroni = 0.022; Fig. 3C), and urge intensity (r = −0.42, P_Bonferroni = 0.011; Fig. 3D) scores of the I-PUTS. There were no correlations between GABA+ levels in insula or SM cortex with any index of urges (all P_Bonferroni > 0.999). Glx levels were also not associated with any index of urges (all P_Bonferroni > 0.150).

Tics

Neither GABA+ nor Glx levels were associated with YGTSS scores or number of tics on the I-PUTS (all P_Bonferroni > 0.140).

Parallel Analyses With Creatine- Rather Than Water-Referenced Metabolites

As with the water-referenced metabolites, group differences of GABA+ and Glx referenced to creatine (i.e., GABA+/Cr and Glx/Cr) levels were comparable between groups. With regard to the associations between brain metabolites and TS symptom severity, the effects were even stronger for the associations between SMA GABA+/Cr and urge severity. For example, the Pearson’s r value for the associations between SMA GABA+/Cr and the number of urges (r = −0.47, P_Bonferroni = 0.003), urge frequency (r = −0.48, P_Bonferroni = 0.002), and urge intensity (r = −0.48, P_Bonferroni = 0.003) were ~10%–20% stronger than those shown in Fig. 3A (see Supporting Information Fig. S2). Interestingly, insula GABA+/Cr levels also showed significant associations with the number of urges (r = −0.40, P_Bonferroni = 0.032), urge frequency (r = −0.46, P_Bonferroni = 0.008), and urge intensity (r = −0.45, P_Bonferroni = 0.011). Given that the associations between SMA GABA+ and urge severity remained significant across reference variables, while insula GABA+ did not, these results suggest that the SMA GABA+ associations are not being driven by the reference variable (and are hence reliable), whereas the insula GABA+ associations might be spurious. We discuss these reference variable discrepancies and the robustness of our correlations in more detail in the Supporting Information.

Effects of Co-occurring ADHD and OCD

The statistical analyses used to investigate the effects of co-occurring ADHD and OCD, as well as their results, are described comprehensively in the Supporting Information (Section 0.5). Broadly, TS individuals with co-occurring ADHD appeared to have significantly higher Glx levels than those without co-occurring ADHD (see Supporting Information Fig. S10). Co-occurring ADHD also appeared to moderate associations between brain metabolite levels and TS symptom severity (Supporting Information Figs. S11 and S12). For instance, higher insula GABA+ levels were associated with a higher number of urges in TS individuals without co-occurring ADHD, but a lower number of urges in those with co-occurring ADHD, suggesting ADHD as an important diagnostic moderator of TS. Regarding OCD, there was no difference in GABA or Glx levels between patients with TS with and without co-occurring OCD (Supporting Information Figs. S13 and S14). Although co-occurring OCD also appeared to moderate some of the associations between brain metabolite levels and TS symptom severity (Supporting Information Figs. S15 and S16), none of the interaction effects would have survived correction for multiple comparisons.

Discussion

We measured, in 8- to 12-year-old children with TS and TDCs, GABA+ and Glx in the right SM1, SMA, and insula. We found that although both groups have comparable GABA and Glx profiles, lower SMA GABA + levels were associated with more frequent and more intense premonitory urges-to-tic in TS. In contrast, GABA+ of SM1 and insula did not show any significant association to urge severity. GABA+ also was not associated with tic severity. We also did not identify any associations between Glx and urge or tic symptom severity.

SMA GABA+ Levels Are Associated With Premonitory Urge-to-Tic Severity in Children With TS

Our most consistent finding was the associations between SMA GABA+ levels with urge severity, which were present regardless of whether GABA+ was referenced to water or Cr. Based on the putative role of the insula (eg, its role in interoceptive awareness⁴³ and awareness of the urge for action in other behavioral domains⁴), it is likely that premonitory urges do have their origins in the insula (specifically the anteriordorsal region of the right insula, which has been demonstrated to be positively associated with premonitory urge scores but uncorrelated with motor tic severity⁴). However, because the posterior of the insula is anatomo-functionally connected to the middle cingulate around the area of the SMA,^44–47 we speculate that although premonitory urges may be initially generated within the insula, awareness of these urges may be partially mediated by tonic GABA levels of the SMA. This would explain our results showing that children with TS who had lower SMA GABA+ levels reported more urges, more frequent urges, and more severe urges. Consistent with our findings of a role for SMA, Tinaz and colleagues¹³ found that, in 13 adults, functional connectivity between the right dorsal anterior insula and bilateral SMA was positively associated with urge severity, assessed using the PUTS. Still, our interpretation of the results here assumes that tics are caused by urges (ie, premonitory urges generated in the insular drive tics). However, if premonitory urges are experienced only when tics are suppressed, a reasonable alternative interpretation could be that lower SMA GABA+ levels result in less efficient tic suppression, which in turn results in an increase in premonitory urges-to-tic.

Draper and colleagues⁴⁸ had previously reported higher GABA levels in SMA in subjects with TS (n = 8, M_Age = 18.30, SD_Age = 2.7) compared with TDC control subjects (n = 9, M_Age = 17.20, SD_Age = 3.30). GABA levels in the SMA also predicted the amount of change in blood oxygen-level dependent activity in the same voxel before voluntary action, with an increase in SMA GABA being interpreted as a possible mechanism by which tic suppression is gained with increasing age.²⁰ Although we did not identify increased SMA GABA+ levels in our sample, our sample contained an appreciably larger and younger sample than that of Draper and colleagues⁴⁸; thus, the results are not necessarily in conflict. However, because our results suggest that GABA+ levels of the SMA mediate urge rather than tic severity in younger children with TS, it is possible that the mechanism of gaining control over tics first occurs through the suppression of urges, rather than simply through control over tics.

Comparable GABA and Glx Levels Between Children With TS and TDCs

Perhaps to our surprise, we were unable to reproduce our prior finding of reduced SM1 GABA+ in children with TS.²⁶ We do not believe that the discrepancy between our prior and current findings are due to the differences in MRS acquisition, because both studies used methods to minimize motion artifact. Similarly, the demographic characteristics of both studies were similar (children aged 8–12 years, with comparable mean age and sex ratios). Instead, we believe the discrepancy to be better explained by the larger sample size (n = 68 in the TS group for the SM1 group comparison vs. n = 17 in our pilot study). Although small samples are not necessarily an issue in the presence of large effects, this larger study likely generated more valid and generalizable results. When taken with the work of others assessing GABA of SM1 in adolescent⁴⁸ and adult¹⁵ TS populations (which were conducted at 7 and 3 T, respectively), our results suggest that SM1 GABA+ levels are comparable between children with TS and TDC.

With regard to the lack of group differences and associations identified for Glx, few studies have assessed glutamate levels in TS. Of the studies that have compared Glx (or glutamate at 7 T, where glutamate and glutamine can be resolved) levels between individuals with TS and TDC, the findings are mixed. At 7 T, Mahone and colleagues³¹ found that glutamate levels were increased in the primary motor cortex of young children with TS (TS: n = 32, M_Age = 9.88, SD_Age = 1.93; TDC: n = 43, M_Age = 8.11, SD_Age = 1.93). For studies at 3 T, Naaijen and colleagues (https://www.sciencedirect.com/science/article/pii/S2213158216302194) found normal Glx levels in the dorsal striatum and anterior cingulate cortex of children with TS (TS: n = 15, M_Age = 10.40, SD_Age = 1.20; TDC: n = 53, M_Age = 10.00, SD_Age = 1.00) and TS + ADHD (n = 28, M_Age = 10.70, SD_Age = 1.60). Kanaan and colleagues⁴⁹ found reduced striatal and thalamic Glx in adults with TS (n = 37, M_Age = 38.30, SD_Age = 11.10; TDC: n = 36, M_Age = 38.40, SD_Age = 11.10). Given the variability in the sample size, MRS acquisition, and ROIs across these studies, it is perhaps too premature to try and draw any firm conclusions regarding the status of glutamate levels in TS based on these studies.

Limitations

GABA concentrations in the brain are low, necessitating measurement and averaging of GABA+ over relatively large brain areas. This could limit the ability to detect important differences within smaller regions. In addition, over and above the limitations of using 3 T rather than 7 T, and the degree to which MRS-derived measures of GABA and glutamate can be used as markers of inhibition and excitation (both of which we discuss in the Supporting Information), many of the correlations identified in this study are influenced by outliers. To be more confident that the associations identified in our study were true associations rather than spurious, we tested how robust the significant associations were to various stages of outlier removal. We also used additional measures of association, such as Spearman’s Rho and percentage-bend.^50,51 The results of those analyses are presented in the Supporting Information. Broadly, although removal of outliers did reduce the strength of the associations, the significant associations between SMA GABA+ and urge often remained significant, making us confident that the associations presented in this study are not simply driven by outliers. Importantly, in comparing our results with those of others, it is important to note differences not only in imaging techniques but also that most other studies focus on adults with TS, whose anatomy and physiology may reflect compensatory processes. Finally, with respect to urge assessment, we used the I-PUTS, whereas many other studies used the PUTS. Correlations between these two scale scores are modest, particularly in children with comorbid diagnoses.³⁶

Future Directions

If SMA GABA levels mediate the conscious experience of premonitory urges generated by the insula, it follows that targeting SMA GABA levels (or GABAergic functioning within the SMA) may be a plausible way of treating the symptoms of TS. Suppression of the SMA using high-frequency repetitive TMS has already been shown to temporarily reduce tic severity in individuals with TS, although the findings have been mixed.^52–54 Increasing tonic GABA levels or improving GABAergic functioning in the right dorsal anterior insula may also help dampen the limbic drive it imposes on the SMA. However, there have been only a few studies that have demonstrated direct stimulation of the insula with TMS^55,56 with some questioning its feasibility. Further discussion of future directions and additional considerations of our interpretation of the results can be found in the Supporting Information.

Conclusions

Using edited MRS, we found that although children with TS had comparable GABA+ and Glx levels as their typically developing peers, SMA GABA+ levels were specifically associated with the number, frequency and intensity of premonitory urges in children with TS. That is, the children with TS who had higher SMA GABA+ levels were also those who had less severe premonitory urges. When taken with existing work, our results suggest that targeting SMA GABA levels may reduce both urge and thereby tic severity in children with TS.

Data Availability Statement

TBC