Deep brain stimulation modulates pallidal and subthalamic neural oscillations in Tourette's syndrome

Abstract Introduction Previous studies found subthalamic nucleus deep brain stimulation (STN‐DBS) has clinical effect on Parkinson's disease, dystonia, and obsessive compulsive disorder. It is noteworthy that only a few studies report the STN‐DBS for Tourette's syndrome (TS). Globus pallidus interna (GPi)‐DBS is the one of the most common targets for TS. So, this paper aims to investigate the neural oscillations in STN and GPi as well as the DBS effect between these two targets in same patients. Methods The local field potentials (LFPs) were simultaneously recorded from the bilateral GPi and STN in four patients with TS. The LFPs were decomposed into neural oscillations, and the frequency and time–frequency characteristics of the neural oscillations were analyzed across the conditions of resting, poststimulation, and movement. Results No difference of resting LFP was found between the two targets. The poststimulation period spectral power revealed the high beta and gamma oscillations were recovered after GPi‐DBS but remained attenuated after STN‐DBS. The STN beta oscillation has fewer changes during tics than voluntary movement, and the gamma oscillation was elevated when the tics appeared. Conclusion The high beta and gamma oscillations in GPi restored after GPi‐DBS, but not STN‐DBS. High beta and gamma oscillations may have physiological function in resisting tics in TS. The cortex compensation effect might be interfered by the STN‐DBS due to the influence on the hyper‐direct pathway but not GPi‐DBS.


| INTRODUC TI ON
Tourette's syndrome (TS) is a neurological disorder manifested by motor and vocal or phonic tics. TS is often accompanied by obsessive compulsive disorder, attention-deficient hyperactive disorder, poor impulse control, and other comorbid behavioral problems (Albin, 2017;Albin & Mink, 2006). Patients with TS have impairments in impulse inhibition which might relate to the reduced inhibitory output of the basal ganglia (Ganos, 2016;Jackson, Draper, Dyke, Pépés, & Jackson, 2015;Jahanshahi & Rothwell, 2017). Deep brain stimulation (DBS) in the thalamus and basal ganglia has been introduced for the treatment of movement disorders (Zhao, Zhang, & Meng, 2016). The most common targets for TS include the center median thalamic region and the globus pallidus.
One study reported a Parkinson's disease (PD) patient who also has a history of TS in whom bilateral subthalamic nucleus (STN)-DBS improved both PD syndrome and tics. The curative effect of STN-DBS on dystonia and other hyperkinetic disease has been proved (Schjerling et al., 2013), and the STN-DBS also shows clinical effect on obsessive compulsive disorder (OCD; Mallet et al., 2008). Overall, STN is a promising target for TS, but the clinical effect and local field potential (LFP) of STN in TS have less been investigated.
Perioperative LFP recordings and intraoperative single-neuron recordings provide insights into neural activities. The power of low frequency over 1-10 Hz oscillations in the thalamus increases with the occurrence of frequent tics (Shute et al., 2016), and low-frequency oscillations over 3-12 Hz were found in the pallidus-thalamus circuit, which were correlated with the preoperative motor tic score (Neumann et al., 2018). Moreover, low-frequency oscillations were found in the thalamus of TS patients with severe compulsive behaviors . Microelectrode recording in the thalamus of patients with TS further confirmed that the firing patterns were within the low-frequency range . DBS at the thalamus tended to attenuate alpha oscillations in the contralateral thalamus and increase theta oscillations in the ipsilateral thalamus in a case study (Bour et al., 2015;Marceglia et al., 2017). Alternatively, in the globus pallidus interna (GPi), attenuated beta oscillation and elevated gamma and high-frequency oscillations were seen along with frequent tics (Jimenez-Shahed, Telkes, Viswanathan, & Ince, 2016).
Low-frequency (4-12 Hz) and high-frequency (>150 Hz) phaseamplitude couplings in the GPi were seen at rest (Ji et al., 2016). So, combining the evidence above, the low-frequency oscillation is pathological because it related to the clinical scores. The other frequency band might reflect the more complicated cross-effect when tics appeared.
What are the characteristic oscillatory activities in the GPi and STN of patients with TS? How do STN and GPi-DBS modulate these activities? This paper aims to answer these questions through an investigation of LFPs recorded simultaneously from the ipsilateral GPi and STN in patients with TS under the conditions of rest, movement, and after high-frequency stimulation to find explanations for the difference clinical effect of STN and GPi-DBS on TS.

| Subjects and surgery
All patients gave written informed consent to participate in this study, which was approved by the local ethics committees of Beijing Scale (Y-BOCS) were performed before and after the surgery (Table 1). The patients were followed up for 6 months.
All patients underwent stereotactic frame preoperative 3.0T magnetic resonance imaging (MRI). The electrodes were targeted to posterior GPi and the lateral area of STN. The targets and trajectory of the electrode implantation were calculated and determined using the Leksell SurgiPlan station (Elekta). The DBS electrodes were PINS L301 for STN with four platinum-iridium cylindrical surface contacts, and each contact was 1.27 mm in diameter and 1.5 mm in length and separated by 0.5 mm. The electrodes for GPi were PINS L302, which had contacts of the same size but separated by 1.5 mm (PINS). The subjects underwent surgery with local anesthesia to allow for intraoperative microelectrode recording and stimulation tests (Starr, 2002). The placement of DBS electrodes was confirmed by postoperative CT scans.

| Paradigm and recording
All patients underwent experiments on the 3-5th day after the surgery. The stimulation was turned off at least 4 hr before the recording. Three channels of bipolar LFPs were recorded from the adjacent four contacts (contact pairs: 0-1, 1-2, and 2-3) of each electrode. The LFPs of patient 1 were amplified, notch filtered to remove 50 Hz line noise, band-pass filtered over 1-500 Hz using a Digitimer amplifier (model D360; Digitimer Ltd.), and recorded with a sampling frequency of 1,000 Hz using a CED 1401 (Cambridge Electronic Design). The signal was down-sampled to 500 Hz for further analysis. The LFPs, electromyography (EMG) of symptominvolved muscles, and forearm flexors of patients 2-4 were amplified and band-pass filtered over 1-250 Hz using a custom-developed amplifier and recorded with a sampling frequency of 500 Hz. This amplifier is a wearable eight-channel wide-scale electrophysiological monitor with a wireless connection using Wi-Fi.  Table 1).
The stimulation was delivered to the active contact with 130 Hz, 60 µs, and amplitudes at 1, 2, and 2.5 V amplitudes. The stimulation was applied bilaterally in the GPi or STN. The duration of stimulation at each amplitude was 180 s with a 180-s interval between stimulations.

| Signal preprocessing
Signal processing and statistical analysis were performed with MATLAB scripts (MathWorks Inc.). The channels with the selected pair of contacts for the stimulation test were used for analysis (Table 1). For the resting state, a continuous 150-s signal was selected for case 1. Continuous 1,200-s signals were selected for cases 2-4. Three poststimulation segments form the pair of contacts selected for stimulation were chosen, and the stimulation period was excluded due to the noise and uncertain stimulation effect caused by relatively close distance. LFPs were band-pass filtered over 3-90 Hz and adaptively band-stop filtered to reject the 50 Hz line noise. EMG was first high-pass filtered at 20 Hz and then smoothed with a 5 Hz low-pass filter.

| Extracting oscillatory modulations by the tics and voluntary movement
The Thus, it is not reliable in methodology to calculate the synchronicity between the nucleus responses to the movements.

| Clinical evaluation
All patients had prodromal symptoms related to motor or vocal tics and comorbidities including obsessions, anxiety, and depression. One patient had stimulation-induced dyskinesia during STN- F I G U R E 1 Signals and the power spectral density. The electromyography (EMGs) recorded from the symptom-involved muscle and local field potentials (LFPs) recorded from ipsilateral globus pallidus interna (GPi) and subthalamic nucleus (STN) in case 2 are shown, respectively, in the time range of 0-150 and 117-119 s in panel (a). The percentage power spectra over 0-90 Hz of LFPs recorded at resting state and averaged across all cases are shown in panels (b and c). The alpha power peaks and high beta power peaks were found in both GPi and STN Figure 1a shows the surface EMGs of the symptom-involved muscle and the pallidal and subthalamic LFPs simultaneously recorded from case 2. The percentage power spectra over 0-90 Hz showed distinct power peaks around low frequency and high beta frequency (Figure 1b,c). Specifically, the power peaks over low frequency were found in six GPi with 13 ± 1.5 Hz (mean ± SD) and six STN with 12 ± 2.5 Hz (mean ± SD). The high beta power peaks were found in five GPi with 26 ± 2 Hz (mean ± SD) and six STN with 26 ± 2 Hz (mean ± SD; see Figure S2). Moreover, the high beta power peaks were clearer in cases 1 and 3 than in cases 2 and 4.

| Oscillatory activities at rest
No significant LFP difference was found between the two nuclei at rest. No significant change in the other frequency bands after stimulation in either nucleus can be found in this study (All p > .05). All individual LFPs are shown in Figure S2.

| Comparison of tic and voluntary movementrelated oscillations
The   neurons and suppresses the synchronized low-frequency oscillation (McCairn, Iriki, & Isoda, 2015). Excessive stimulation in the STN may induce dyskinesia together with the elevated low-frequency power . However, after STN or GPi-DBS, we did not observe the low-frequency band power was inhibited. The phenomenon might result from the tics do not like rigidity and bradykinesia which is status and DBS have aftereffect on the status, we did not observe the tics in the short poststimulation period.

| High beta neural oscillation in TS
High beta oscillations were found in the GPi and STN in this study.
Beta oscillation is widely explored in the cortical-basal ganglia network and is considered related to hypoactivity. These oscillations desynchronize before and during the movement onset and reflect effort, velocity, and force of the movement (Anzak et al., 2012;Joundi et al., 2012;Kühn et al., 2004). The power of beta oscillations over 13-35 Hz in the GPi in TS is not correlated with the severity of motor tics but contributes to the prediction of preoperative YGTSS scores (Neumann et al., 2018). Here in this study, the 15-30 Hz beta band power has fewer changes in tics than in voluntary movement, which might also imply that this frequency band is involved in TS.
During the occurrence of tics, broadband attenuations over the beta frequency were seen in the STN (Figure 3a). The frequencyintegrated dynamics showed that tics and voluntary movement-related beta ERD existed in both the GPi and STN, while the degree of power attenuation was lower around tics than around voluntary movement. Considering that facial tics are much smaller in size than a voluntary wrist-folding movement, this result confirms that beta oscillation is physiologically related to motor behavior and may not be pathological in TS.
Beta oscillation can be subdivided into low beta (<20 Hz) and high beta (>20 Hz) oscillations. In PD, the low beta oscillation is most likely correlated with motor severity and suppressed on levodopa (Little & Brown, 2014;Ray et al., 2008). Alternatively, high beta oscillation is less related to parkinsonian bradykinesia and rarely correlated with the amount of dopamine. High beta oscillation may result from the hyper-direct pathway (Neumann et al., 2016(Neumann et al., , 2018Oswal et al., 2016) and is predominantly driven by the cortex (Hirschmann et al., 2011;Litvak et al., 2010;Williams et al., 2002). The hyper-direct pathway transfers cortical action in shorter latencies via the STN with high beta activity (Botvinick, Cohen, & Carter, 2004; and therefore assists in situations that require rapid stop responses to external and internal triggers (Aron et al., 2007;Eagle et al., 2007). In this regard, relative beta increases are found in multiple cortical areas and STN during reactive and proactive inhibition.
Successful impulse control shows higher beta power than unsuccessful impulse control in the STN and downstream nuclei, the GPi (Alegre et al., 2013;Brittain et al., 2012;Brücke et al., 2008;Kühn et al., 2004). Thus, the high beta oscillation in TS may reflect successful impulse control of the internal urge to express tics. Interestingly, more high beta oscillations were seen in cases 1 and 3, which also had fewer tics after the surgery and better improvements during the acute phase, while cases 2 and 4 had fewer high beta oscillations and rapid occurrence of motor and vocal tics. In fact, the high beta oscillation in the basal ganglia might be a compensation that stabilizes the physiological function of the cortical-basal ganglia network (Leventhal et al., 2012;Martin-Rodríguez & Mir, 2018).  -Shahed et al., 2016). In this study, the subthalamic gamma oscillation can be observed during the onset of tic. In contrast to the movement-related modulations over the beta band, the wrist-folding voluntary movement, which is stronger than facial tics, was not accompanied by more gamma synchronization. So, we hypothesize that the high gamma oscillation is also involved in the compensation effect of brain under TS. While the high beta does not change when the tics appeared, this might result from the highly complexity of the modulation effect of the cortex; the high beta might reflect a sustained status maintain which is elevated in TS. The high gamma is related to a sudden inhibition attempt.
The stimulation effects of DBS on the oscillations of basal ganglia in TS have not been widely investigated. The signals in brain is not always "pathology"; it has many physiology functions, like the beta is related to the motor control and gamma is related to the movement speed and vigilance, and the signals help people stop fast to the sudden danger, which is an inhibition effect (Brown & Williams, 2005). Compared to purely inhibition theory, the filter effect of deep brain stimulation is more and more recognized. A moderate function is good (Ashkan, Rogers, Bergman, & Ughratdar, 2017;Chiken & Nambu, 2016;McIntyre, Savasta, Kerkerian-Le Goff, & Vitek, 2004). In this study, high beta and gamma oscillations restore in the circuit after GPi-DBS but exhib- control dysfunction such as gambling and hypersexuality in PD (Merola et al., 2017), which might also relate to the compromised physiology inhibition function.

| Limitations
Due to the small sample size of this study, it is not reliable to evaluate the relationship between the oscillatory activities and the severity of tics. Hence, it is difficult to determine whether stimulation-modulated high beta and gamma oscillations are correlated with symptom.

| CON CLUS ION
The

ACK N OWLED G M ENTS
The authors express great appreciation to Prof. Edmund T Rolls at Institute of Science and Technology for Brain-Inspired Intelligence, Fudan University, for his valuable comments and language revisions.

CO N FLI C T O F I NTE R E S T
We wish to confirm that there are no known conflicts of interest associated with this publication and there has been no significant financial support for this work that could have influenced its outcome.
We confirm that the manuscript has been read and approved by all named authors and that there are no other persons who satisfied the criteria for authorship is missed. We further confirm that the order of authors listed in the manuscript has been approved by all of us.
We confirm that we have given due consideration to the protection of intellectual property associated with this work and that there are no impediments to publication, including the timing of publication, with respect to intellectual property. In so doing, we confirm that we have followed the regulations of our institutions concerning intellectual property. We further confirm that any aspect of the work covered in this manuscript that has involved either experiment animals or human patients has been conducted with the ethical approval of all relevant bodies and that such approvals are acknowledged within the manuscript. We understand that the corresponding author is the sole contact for the editorial process (including editorial manager and direct communications with the office). He/she is responsible for communicating with the other authors about progress, submissions of revisions, and final approval of proofs. We confirm that we F I G U R E 4 Schematic illustration of the possible mechanism of the different effect of globus pallidus interna-deep brain stimulation (GPi-DBS) and subthalamic nucleus (STN)-DBS for TS. The low-frequency oscillation in STN and GPi implicates the pathological state of tics. The high beta oscillation underlines a compensatory sustained inhibition effect, and the gamma oscillation underlines a sudden inhibition effect. The high beta and gamma oscillations in STN and GPi remained attenuated after STN-DBS, while recovering after GPi-DBS have provided a current, correct email address which is accessible by the corresponding author and which has been configured to accept email from zjguo65@163.com.

DATA AVA I L A B I L I T Y S TAT E M E N T
Data were made available to all interested researchers upon request.