Brain Stimulation 7 (2014) 212e218

Contents lists available at ScienceDirect

Brain Stimulation journal homepage: www.brainstimjrnl.com

Functional MRI-navigated Repetitive Transcranial Magnetic Stimulation Over Supplementary Motor Area in Chronic Tic Disorders Steve W. Wu a, *, Thomas Maloney b, Donald L. Gilbert a, Stephan G. Dixon a, Paul S. Horn a, David A. Huddleston a, Kenneth Eaton b, Jennifer Vannest a, b a b

Cincinnati Children’s Hospital Medical Center, Division of Neurology, 3333 Burnet Ave., MLC 2015, Cincinnati, OH 45229, USA Cincinnati Children’s Hospital Medical Center, Division of Radiology, 3333 Burnet Ave., MLC 5033, Cincinnati, OH 45229, USA

a r t i c l e i n f o

a b s t r a c t

Article history: Received 30 July 2013 Received in revised form 12 October 2013 Accepted 24 October 2013 Available online 22 November 2013

Background: Open label studies have shown repetitive transcranial magnetic stimulation to be effective in reducing tics. Objectives: To determine whether 8 sessions of continuous theta burst stimulation (cTBS) over supplementary motor area (SMA) given over 2 days may reduce tics and motor cortical network activity in Tourette syndrome/chronic tic disorders. Methods: This was a randomized (1:1), double-blind, sham-controlled trial of functional MRI (fMRI)navigated, 30 Hz cTBS at 90% of resting motor threshold (RMT) over SMA in 12 patients ages 10e22 years. Comorbid ADHD (n ¼ 8), OCD (n ¼ 8), and stable concurrent medications (n ¼ 9) were permitted. Neuronavigation utilized each individual’s event-related fMRI signal. Primary clinical and cortical outcomes were: 1) Yale Global Tic Severity Scale (YGTSS) at one week; 2) fMRI event-related signal in SMA and primary motor cortex (M1) during a finger-tapping motor task. Result: Baseline characteristics were not statistically different between groups (age, current tic/OCD/ ADHD severities, tic-years, number of prior medication trials, RMT). Mean YGTSS scores decreased in both active (27.5  7.4 to 23.2  9.8) and sham (26.8  4.8 to 21.7  7.7) groups. However, no significant difference in video-based tic severity rating was detected between the two groups. Two-day posttreatment fMRI activation during finger tapping decreased significantly in active vs. sham groups for SMA (P ¼ 0.02), left M1 (P ¼ 0.0004), and right M1 (P < 0.0001). No serious adverse events occurred. Conclusion: Active, fMRI-navigated cTBS administered in 8 sessions over 2 days to the SMA induced significant inhibition in the motor network (SMA, bilateral M1). However, both groups on average experienced tic reduction at 7 days. Larger sample size and protocol modifications may be needed to produce clinically significant tic reduction beyond placebo effect. Ó 2014 Elsevier Inc. All rights reserved.

Keywords: Repetitive transcranial magnetic stimulation Theta burst stimulation Plasticity Tourette syndrome Supplementary motor area

Introduction Repetitive transcranial magnetic stimulation (rTMS) is a method of noninvasive brain stimulation with therapeutic potential in neuropsychiatric conditions. Regulatory agencies in several

Abstract presentation: This abstract was presented at the 42nd Annual Meeting of the Child Neurology Society in Austin, TX (Oct 30eNov 2, 2013). Drs. Wu, Vannest, Eaton and Mr. Huddleston received grant funding from the Tourette Syndrome Association to conduct this study. Financial disclosures: None of the authors have any financial conflict of interest to disclose. * Corresponding author. Tel.: þ1 513 803 1510; fax: þ1 513 636 1063. E-mail address: [email protected] (S.W. Wu). 1935-861X/$ e see front matter Ó 2014 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.brs.2013.10.005

countries have approved rTMS for the treatment of severe depression. Due to its potential clinical impact, researchers have been studying rTMS in other conditions, such as Tourette Syndrome (TS), for symptom reduction. Since the first report of tic reduction by rTMS [1], several studies have shown rTMS to be effective in decreasing tics [2e6] while others demonstrated no benefit [7,8]. Both study design and stimulation parameters likely contribute to the mixed results. Open-label studies were more likely to show benefit [1e5] than sham-controlled and blinded studies [6e8]. Differences in type of stimulation coil, pulses/day, stimulation intensity, duration and cortical target may also influence outcomes. Earlier rTMS-TS studies focused on prefrontal and premotor cortex as TS imaging studies have shown increased activities in

S.W. Wu et al. / Brain Stimulation 7 (2014) 212e218

these regions [9,10]. However, only one [6] of three studies [6e8] targeting these sites showed benefit. More recent open-label rTMS-TS studies were effective in tic reduction when targeting supplementary motor area (SMA) [2e5]. SMA plays a critical role in motor control [11e13] and is accessible to noninvasive brain stimulation. Several studies provide support for targeting SMA in TS. Functional MRI (fMRI) studies have found that SMA appears to be overactive in TS patients when performing motor tasks [14] and bilateral SMA are the most active regions immediately before tic execution [15]. Both fMRI and magnetoencephalography have shown increased functional connectivity between SMA and motor cortex in chronic tic patients [16,17]. A recent study of excitatory rTMS over SMA was shown to induce echophenomena in healthy adults [18]. Together these findings suggest SMA as a reasonable target for brain stimulation intervention in TS. The objectives of this randomized, sham-controlled pilot study was to use fMRI to 1) provide individualized neuro-navigated rTMS stimulation of the SMA in TS/chronic tic disorder patients, and 2) to determine treatment-associated cortical changes. We hypothesized that repeated active rTMS stimulation of the SMA over 2 consecutive days would reduce tics compared to sham stimulation. We also hypothesized that this stimulation would alter the blood oxygenation level dependent (BOLD) signal change in SMA and primary motor cortex (M1) during a finger tapping task. We chose to deliver a specific paradigm of rTMS called continuous theta burst stimulation (cTBS) [19]. The cTBS-induced neurophysiologic effect on M1 is similar to that of conventional inhibitory 1 Hz rTMS [20,21]. However, cTBS has benefits over 1 Hz rTMS in that the stimulation duration is much shorter and the stimulation intensity is lower, thus making this technique more feasible for pediatric patients. To further enhance the convenience and feasibility of this intervention, we compressed eight sessions of cTBS over 2 consecutive days as repeated cTBS sessions have been shown to safely produce significant neurophysiologic and clinical changes [22e25].

213

Experimental design SWW, the only un-blinded investigator throughout the entire study, generated the randomization of active vs. sham stimulation (1:1) in blocks of four using Microsoft Excel. No other investigator had access to the randomization. SWW and DLG offered participation to all eligible patients. Agreed participants were sequentially assigned to a randomized condition after the consent/assent process during the first visit (Fig. 1). During the first visit, each participant underwent clinical assessment and fMRI by investigators blinded to treatment assignment. The TMS operators also demonstrated the TMS machines and obtained resting motor threshold (RMT). On visit 2, each patient underwent videotaping [31] and had repeat RMT measurements. Afterward, cTBS was delivered over the SMA using BrainSight2Ò software (Rogue Research Inc., Montreal, Canada) for individualized neuronavigation based on fMRI data from visit 1. On the following day, each patient received repeated cTBS again over SMA. Immediately after the stimulation, the participant was videotaped [19] again and then underwent the second fMRI. Seven days later, the participants returned for videotaping [31] and clinical assessments. Behavioral tasks The finger-tapping motor task was used to isolate neural activity involved in motor planning and execution. This task had active and control conditions, both of which presented the participant with audio and visual stimuli. During both conditions, the participant heard a sequence of 1e4 tones (100 ms duration, 100 ms between tones, 100 Hz) and was presented with an image indicating to them as to whether they were on the active or control portion of the task. The active condition required the participants to tap their bilateral thumbs and index fingers together with the number of taps matching the number of tones heard. The control condition required the participants to simply listen to the tones. In total this task had 5 cycles of active and control conditions implemented as a block design with 4 active and 4 control conditions per cycle, one

Methods and materials Patient recruitment, diagnosis, clinical assessment Patients (10 years old) with chronic tics were recruited from the Tourette Syndrome Clinic at the Cincinnati Children’s Hospital Medical Center (CCHMC). Diagnoses were based on DSM-IV-TR criteria using direct physician interview. Inclusion criteria were 1) diagnosis of TS or chronic motor/vocal tic disorder, 2) Yale Global Tic Severity Scale (26) (YGTSS) score 20, 3) no medication changes within 10 days of receiving cTBS and throughout the remainder of the study and 4) no botulinum toxin injection within 12 weeks of starting and throughout the study. Patients with implanted metallic medical devices, pregnancy and epilepsy were excluded. Tic, attention deficit hyperactivity disorder (ADHD), obsessive compulsive disorder (OCD) symptom severities were assessed using validated scales e YGTSS, Gilles de la Tourette Syndrome Quality of Life Scale (GTS-QOL) [27], Premonitory Urge for Tics Scale (PUTS) [28], Children’s Yale-Brown Obsessive Compulsive Scale (CYBOCS) [29] and DuPaul ADHD Rating Scale [30]. Patients were also videotaped according to the Rush Video-Based Tic Rating Scale (RVTRS) [31]. These ratings were done by a blinded rater. Adult participants and parent(s) of pediatric patients gave written informed consent for the study, which was approved by the CCHMC Institutional Review Board. Children also gave written assent for the study. This study was registered with ClinicalTrials.gov (NCT01258790).

Figure 1. Study flow.

214

S.W. Wu et al. / Brain Stimulation 7 (2014) 212e218

for each of the 1e4 tone conditions, with the number of tones presented in each being randomized within the block. A continuous tic fMRI acquisition was used to capture neural activation involved in motor tics vs. rest. During this acquisition, SWW pressed a button whenever a motor tic was observed. The button-push marked the time of the tic in order to use the timing in an event related analysis of the fMRI scan. A sparse acquisition was used to detect neural activation associated with phonic tic production contrasted with rest. During this task, three volume acquisitions were acquired for 6 s, then the scanner paused for 6 s during which time audio from the participant was recorded. The timing was such that the peak of the BOLD response elicited during a scanner pause period was captured while the scanner was on again. Following the scan, SWW listened to the audio files to identify periods of phonic tics for event-related fMRI analysis. Functional MRI All MRI scans were taken on the same 3T Philips Achieva MRI Scanner. High-resolution T1-weighted anatomical images were acquired with TR/TE ¼ 8.0/3.7 ms, FOV 25.6  25.6  18.0 cm, matrix 256  256, voxel size 1 mm isotropic. A single shot gradientecho echo planar imaging (EPI) sequence was used to acquire functional images. Each functional image volume for the motor task consisted of thirty-two 5 mm T2*-weighted axial slices (TR/ TE ¼ 2000/38 ms, FOV 25.6  25.6 cm, matrix 64  64; voxel size of 4  4  5 mm); a total of 130 volumes were acquired. For the sparse acquisition, each functional image volume consisted of thirty-two 4 mm T2*-weighted axial slices (TR/TE ¼ 2000/38 ms, FOV 24.0  24.0 cm, matrix 64  64; voxel size of 3.75  3.75  4 mm) for a total of 150 volumes. The sparse acquisition involved three volume acquisitions (6 s total) followed by 6 s of no acquisition for a total of 150 volumes. This allowed for phonic tics to be heard without the interference of the scanner noise but still captured the peak of the BOLD response from the phonic tic(s). MRI analysis Analysis of the functional images was performed similar to previous studies [32e34] using Cincinnati Children’s Hospital Image Processing Software (CCHIPSÓ), which was written in-house using Interactive Data Language (ITT Visual Information Solutions, Boulder, CO). The analysis of each subject’s fMRI data involved three main steps: 1) aligning each volume in the EPI dataset to a single reference volume using a pyramid iterative algorithm [35], with the reference volume chosen by minimizing a cost function based on the variation in image intensity; 2) implementing a generalized linear model (GLM) on the aligned dataset; 3) normalization of the data to the Talairach coordinate reference frame [36] prior to group comparison. The GLM included subject motion parameters as covariates as well as a set of cosine basis functions to account for drift, respiratory and cardiac signals. Due to the differences in the mean intensity among the three image volumes in the sparse acquisition, a result of T2* relaxation effects, an extra step was required to process these files. They must be split into three files, each with 50 image volumes, one for each volume in the three volume acquisition. Each of these files was then analyzed separately and the three resulting Z maps were combined into one. Regions of interest (ROI) analysis An ROI analysis was performed to determine the differences in activation among the active and sham in areas of neuronal activation elicited by the motor task. The areas of interest (right/left M1, SMA) were chosen by applying an uncorrected threshold of

P < 0.005 to the pre-cTBS group composite map (one sample t-test) of all 12 participants. These regions were then used to extract the positive mean Z value for each region from the individual participant’s pre-cTBS and post-cTBS Z maps. Single pulse TMS Single pulse TMS was performed with a MagstimÒ 200 stimulator and a figure-8, 70 mm coil (Magstim Co., Wales, UK). Surface electromyography (EMG) leads were placed over the first dorsal interosseous (FDI) muscle on the right hand. Participants were seated comfortably, with both arms fully supported on a pillow. Full muscle relaxation was maintained through visual and EMG monitoring. The coil was placed over the left primary motor cortex at the optimal site for obtaining a motor-evoked potential (MEP) in the right FDI, and the resting motor threshold (RMT) was quantified using standard methods [37]. MEP was recorded by using surface EMG electrodes taped to the skin. The signal was amplified and filtered (100/1000 Hz) (Coulbourn Instruments, Allentown, PA) before being digitized at 2 kHz and stored for analysis using SignalÒ software and a Micro1401 interface (Cambridge Electronic Design, Cambridge, UK). Neuro-navigated, targeted continuous theta burst stimulation CTBS was performed using Magstim SuperRapid2Ò (Magstim Co., Wales, UK). The cTBS magnetic pulsing pattern consists of 3 pulses per burst (30 Hz) with bursts repeating 5 times per second for a total of 600 pulses per train [38]. Thirty Hertz cTBS is different from the originally described cTBS [19] (50 Hz) but it can produce similar inhibitory cortical effects [22,23,38,39]. Stimulation intensity was set at 90% of the RMT. Using BrainSight2Ò, SMA targeting was individualized by reconstructing fMRI data into a 3dimensional brain image for neuro-navigation. SMA localization was based on BOLD signal acquired either during finger-tapping (n ¼ 9), continuous tic acquisition (n ¼ 2), or sparse acquisition (n ¼ 1) data. Four trains of cTBS were delivered per day for 2 consecutive days. The first and second cTBS trains were delivered 15 min apart. The third and fourth cTBS trains were delivered 60 and 75 min after the first cTBS. For each patient, cTBS treatment sessions were done at similar time of the day. Statistical analysis All statistical analyses were performed using the SASÒ9.3 (SAS Institute, Inc., Cary, NC). To evaluate the success of the randomization, age, years since tic onset, number of past medication trials, RMT, baseline YGTSS, GTS-QOL, RVTRS, PUTS, CY-BOCS and DuPaul ADHD Rating Scale scores were compared between treatment groups using exact WilcoxoneManneWhitney test. Clinical outcomes The primary clinical outcome for analysis was the YGTSS total tic score, from baseline to the one week visit. Treatment response is defined as a reduction of >6 points on YGTSS [40]. Analysis of covariance (ANCOVA) was used to analyze YGTSS, with baseline score as a covariate and cTBS type as a between-subject factor. ANCOVA was similarly used for secondary outcome ratings of GTSQOL, PUTS, CY-BOCS and DuPaul ADHD Rating Scale scores. Repeated measures analysis of variance (ANOVA) was used to compare the RVTRS scores from 3 time points (baseline, immediately after, and one week after cTBS stimulation), with baseline RVTRS score as a covariate. The level of statistical significance was set at 0.05 for all clinical ratings.

S.W. Wu et al. / Brain Stimulation 7 (2014) 212e218

ambidextrous), six patients received active while six others received sham cTBS. Baseline clinical and physiological variables did not differ (Table 1). Comorbidities and concurrent medications are presented in Table 2. All twelve patients who received cTBS completed the study without serious adverse events. Three participants complained of mild adverse events (abdominal pain, headaches, dry eyes) which resolved without medical intervention.

Table 1 Baseline characteristics. Active Age YGTSS Tic years Past medication trials RMT RVTRS PUTS GTS-QOL CYBOCS DuPaul ADHD Rating Scale

13.5 27.5 7.5 3.3 52 10.2 25.7 35.8 8.5 28.8

         

Sham 3.9 7.4 4.2 2.7 6.2 4.8 6.7 27.9 8.7 12.0

15.5 26.8 9.5 2.8 48.7 7.5 23.7 27.5 5.7 27

         

4.0 4.8 5.5 1.6 14.5 2.9 8.2 13.2 3.8 12.4

Test statistic

P value

33 40 35.5 39.5 48 47 41 38.5 40.5 28.5

0.36 0.92 0.62 1.0 0.17 0.22 0.79 0.97 0.85 0.88

215

Clinical outcome

ADHD, Attention Deficit Hyperactivity Disorder; CY-BOCS, Children’s Yale-Brown Obsessive-Compulsive Scale; GTS-QOL, Gilles de la Tourette Syndrome Quality of Life Scale; PUTS, Premonitory Urge for Tics Scale; RMT, resting motor threshold; RVTRS, Rush Videotaped Tic Rating Scale; Tic years, years since initial tic onset; YGTSS, Yale Global Tic Severity Scale total tic score.

Pathophysiological changes The primary brain-based outcomes for analysis were the BOLD signal changes in bilateral SMA and each M1 from baseline to immediately after the 8th cTBS session. Given the presence of outliers in the fMRI data, a robust linear modeling [41,42] was chosen for analysis. This has several advantages over the traditional model based on Least Squares. Statistically robust methods are able to resist the adverse effect of outliers if they are present, while being reasonably efficient if outliers are absent. For this dataset, we used a robust ANCOVA model in order to make best use of the data, given the limited sample size. The resulting linear model for comparing finger tapping-related fMRI BOLD signals before and after cTBS treatment included baseline fMRI BOLD signal as a covariate, cTBS type as between-subject factor and the interaction between baseline fMRI BOLD signal and cTBS type. Results Patients Thirteen participants were recruited for the study. One child (10 year-old male) did not receive cTBS because his motor threshold was too high. Of the remaining twelve (11 right handed and 1

Although 3 participants in each group had YGTSS score reduction of >6, there was no difference comparing active and sham cTBS (ANCOVA; cTBS type F1,8 ¼ 0.68, P ¼ 0.43; cTBS type*baseline YGTSS F1,8 ¼ 0.63, P ¼ 0.45). Secondary outcome RVTRS changes on day 2 and day 9 also did not show a group difference (repeated measures ANOVA; time*group interaction F2,18 ¼ 2.01; P ¼ 0.16). Analysis of secondary clinical outcomes did not detect any statistical significant difference in GTS-QOL, PUTS, CY-BOCS and DuPaul ADHD Rating Scale scores (Table 2). Neuroimaging Compared to sham treatment, fMRI activation during finger tapping was significantly decreased over SMA, left M1 and right M1 after active cTBS (Table 3). Treatment group effects were significant (Fig. 2). Baseline was not statistically significant, but there was a significant baseline*treatment group interaction, suggesting that cTBS effects on movement-related BOLD signal were partly influenced by baseline BOLD activation. Discussion This is the first sham-controlled cTBS study for TS/chronic tic disorder patients using SMA as the stimulation target. Although, compared to sham stimulation, the 2-day repeated, cTBS inhibitory protocol did not demonstrate a difference in tic reduction, it did significantly decrease movement-related BOLD signals in SMA and bilateral M1. Both children and adults tolerated eight trains of cTBS over 2 consecutive days without any serious adverse event.

Table 2 Clinical data. Age/sex Co-morbidity Current medication(s)

RMT YGTSS

RVTRS

GTS-QOL

PUTS

CY-BOCS

DARS

Day1 Day9 Day1 Day2 Day9 Day1 Day9 Day1 Day9 Day1 Day9 Day1 Day9 Active 10M cTBS 10M 12M

ADHD ADHD, OCD ADHD, OCD

13M 16M 20M

ADHD, OCD ADHD, OCD

Average Sham 10M ADHD, OCD cTBS 13F ADHD, OCD 15M ADHD, OCD 16M 17F 22F OCD Average Average (both groups)

topiramate methylphenidate, sertraline clonazepam, guanfacine, risperidone, sertraline guanfacine, topiramate aripiprazole, atomoxetine, sertraline baclofen, clonazepam, gabapentin, venlafaxine baclofen, clonidine, dextroamphetamine, guanfacine, risperidone haloperidol, methylphenidate methylphenidate

54 51 63

31 22 37

21 28 32

11 11 15

11 5 10

10 11 8

14 53 52

14 20 60

29 28 29

13 21 32

0 19 19

0 18 16

11 39 38

18 29 35

51 48 45

21 34 20

14 34 10

4 15 5

4 5 5

4 14 0

2 20 74

0 19 6

15 20 33

13 21 15

9 3 1

0 0 1

22 34 e

21 39 e

77

27.5 27

23.2 27

10.2 2

6.7 2

7.8 6

35.8 21

19.8 25

25.7 10

19.2 16

8.5 8

5.8 9

28.8 34

28.4 36

23 27 31 33 20 26.8 27.2

14 22 25 31 11 21.7 22.4

7 9 9 10 8 7.5 8.8

0 6 9 10 8 5.8 6.3

4 6 9 10 8 7.2 7.5

23 17 30 21 53 27.5 31.7

43 10 20 9 51 26.3 23.1

30 17 27 29 29 23.7 24.7

25 21 27 25 28 23.7 21.4

5 7 1 2 11 5.7 7.1

4 4 1 0 11 4.8 5.3

26 44 18 13 e 27 27.9

16 43 23 12 e 26 27.2

43 49 40 46 37

ADHD, attention deficit hyperactivity disorder; CY-BOCS, Children’s Yale-Brown Obsessive-Compulsive Scale; DARS, DuPaul ADHD Rating Scale, GTS-QOL, Gilles de la Tourette Syndrome Quality of Life Scale; OCD, obsessive compulsive disorder; PUTS, Premonitory Urge for Tics Scale; RMT, baseline resting motor threshold; RVTRS, Rush Videotaped Tic Rating Scale; YGTSS, Yale Global Tic Severity Scale total tic score.

216

S.W. Wu et al. / Brain Stimulation 7 (2014) 212e218

Table 3 fMRI data comparing active vs. sham cTBS. Location

r2

SMA

0.41

Left M1

Right M1

Estimate

P value

Baseline Group effect (cTBS type) Interaction effect (cTBS*baseline)

1.14 4.17 1.62

0.12 0.02 0.05

Baseline Group effect (cTBS type) Interaction effect (cTBS*baseline)

0.26 5.22 1.48

0.28 0.0004 0.001

Baseline Group effect (cTBS type) Interaction effect (cTBS*baseline)

0.31 4.96 1.45

0.17