Deep brain stimulation in the caudal zona incerta modulates the sensorimotor cerebello-cerebral circuit in essential tremor

NeuroImage 209 (2020) 116511

Contents lists available at ScienceDirect

NeuroImage journal homepage: www.elsevier.com/locate/neuroimage

Deep brain stimulation in the caudal zona incerta modulates the sensorimotor cerebello-cerebral circuit in essential tremor €ran Westling b, Johan Eriksson a, b Amar Awad a, b, *, Patric Blomstedt c, Go a b c

Umeå Center for Functional Brain Imaging (UFBI), Umeå University, Sweden Department of Integrative Medical Biology, Physiology Section, Umeå University, Sweden Department of Pharmacology and Clinical Neuroscience, Umeå University, Sweden

A R T I C L E I N F O

A B S T R A C T

Keywords: Essential tremor Deep brain stimulation Caudal zona incerta Functional MRI Cerebello-cerebral circuit

Essential tremor is effectively treated with deep brain stimulation (DBS), but the neural mechanisms underlying the treatment effect are poorly understood. Essential tremor is driven by a dysfunctional cerebello-thalamocerebral circuit resulting in pathological tremor oscillations. DBS is hypothesised to interfere with these oscillations at the stimulated target level, but it is unknown whether the stimulation modulates the activity of the cerebello-thalamo-cerebral circuit during different task states (with and without tremor) in awake essential tremor patients. To address this issue, we used functional MRI in 16 essential tremor patients chronically implanted with DBS in the caudal zona incerta. During scanning, the patients performed unilateral tremorinducing postural holding and pointing tasks as well as rest, with contralateral stimulation turned On and Off. We show that DBS exerts both task-dependent as well as task-independent modulation of the sensorimotor cerebello-cerebral regions (p
0.05, FWE cluster-corrected for multiple comparisons). Task-dependent modulation (DBS  task interaction) resulted in two patterns of stimulation effects. Firstly, activity decreases (blood oxygen level-dependent signal) during tremor-inducing postural holding in the primary sensorimotor cortex and cerebellar lobule VIII, and activity increases in the supplementary motor area and cerebellar lobule V during rest (p
0.05, post hoc two-tailed t-test). These effects represent differences at the effector level and may reflect DBSinduced tremor reduction since the primary sensorimotor cortex, cerebellum and supplementary motor area exhibit less motor task-activity as compared to the resting condition during On stimulation. Secondly, taskindependent modulation (main effect of DBS) was observed as activity increase in the lateral premotor cortex during all motor tasks, and also during rest (p
0.05, post hoc two-tailed t-test). This task-independent effect may mediate the therapeutic effects of DBS through the facilitation of the premotor control over the sensorimotor circuit, making it less susceptible to tremor entrainment. Our findings support the notion that DBS in essential tremor is modulating the sensorimotor cerebello-cerebral circuit, distant to the stimulated target, and illustrate the complexity of stimulation mechanisms by demonstrating task-dependent as well as task-independent actions in cerebello-cerebral regions.

1. Introduction Essential tremor, the most common movement disorder, can be disabling to the degree of necessitating deep brain stimulation (DBS). DBS is effective in alleviating the tremor, but the neural mechanisms underlying the treatment effect are poorly understood. Essential tremor is characterised by bilateral upper limb postural and/or kinetic tremor, and

occasionally tremor of the head and voice (Bhatia et al., 2018; Hopfner and Deuschl, 2018). In patients with disabling and medically intractable symptoms, DBS is effective in reducing the tremor (Deuschl et al., 2011). Traditionally, the ventral intermediate nucleus of the ventrolateral thalamus (Vim) has been targeted for stimulation, but recently stimulation in the caudal zona incerta (cZi) within the posterior subthalamic area is becoming more common and has been suggested to be more

Abbreviations: BOLD, blood oxygen level-dependent; cZi, caudal part of zona incerta; DBS, deep brain stimulation; PMC, premotor cortex; SMA, supplementary motor area; Vim, ventral intermediate nucleus of the thalamus. * Corresponding author. Umeå Center for Functional Brain Imaging (UFBI), Department of Integrative Medical Biology, Physiology section, Johan Bures v€ag 12, Biology building, Umeå University, 901 87, Umeå, Sweden. E-mail address: [email protected] (A. Awad). https://doi.org/10.1016/j.neuroimage.2019.116511 Received 18 September 2019; Received in revised form 9 November 2019; Accepted 30 December 2019 Available online 31 December 2019 1053-8119/© 2020 The Authors. Published by Elsevier Inc. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/).

A. Awad et al.

NeuroImage 209 (2020) 116511

in our unit at the University Hospital of Umeå (n ¼ 60) were considered for participation in the study. Of these, 35 were excluded due to cognitive impairment/dementia, claustrophobia, significant head tremor, or MRincompatible DBS system. Of the remaining 25 patients who were asked to participate in the study, 17 consented but one died from unrelated causes before the initiation of the study. The diagnosis was set by a movement-disorders specialist according to the ‘‘consensus statement of the Movement Disorder Society on Tremor’’ (Deuschl et al., 1998). A new consensus on the classification of tremor (Bhatia et al., 2018) was established after the diagnosis of our patients and the conduction of the study. However, none of the patients is likely to get a different tremor diagnosis in the light of the new consensus. See Table 1 for patient demographics. The patients had been receiving chronic DBS in the cZi with a stable clinical response for at least 1 year (range 1–5.8 years). The surgical procedure has been described in detail previously (Blomstedt et al., 2010). The target in the cZi was visually identified on stereotactic MRI slightly posteromedial to the posterior tip of the subthalamic nucleus at the level of the maximal diameter of the red nucleus (Fig. 1). The location of the electrodes was verified using an intraoperative CT fused with the preoperative MRI. The patients were implanted with electrode model 3389 Medtronic and a single “implanted pulse generator” (Activa, Medtronic). All patients gave written informed consent, and the study was approved by the local medical ethical board and was performed in accordance with the Declaration of Helsinki.

effective in alleviating the tremor (Barbe et al., 2018; Blomstedt et al., 2010; Deuschl et al., 2011; Fytagoridis et al., 2012; Plaha et al., 2011, 2008; Sandvik et al., 2012). Tremor generation is proposed to be centrally driven by a dysfunctional cerebello-thalamo-cerebral circuit resulting in pathological oscillations (Helmich et al., 2013; Hopfner and Deuschl, 2018). The abnormally oscillating circuit includes the cerebellum, brainstem, thalamus and the sensorimotor cortex (Raethjen and Deuschl, 2012), the activity of which is coherent with tremor frequency during motor tasks (Schnitzler et al., 2009). Regions within this circuit exhibit dysfunctional activity during active motor tasks and during rest (Bhalsing et al., 2013; Sharifi et al., 2014). Functional imaging studies have revealed abnormally increased activity in the primary sensorimotor cortex, premotor cortex (PMC), supplementary motor area (SMA), thalamus and cerebellum during motor tasks (Broersma et al., 2016; Bucher et al., 1997; Colebatch et al., 1990; Neely et al., 2015), and during rest (Boecker et al., 1996; Hallett and Dubinsky, 1993; Jenkins et al., 1993; Wills et al., 1994) as compared to normal subjects. How the disease-related activity is influenced by DBS has not yet been clarified. DBS, and other stereotactic procedures, have been proposed to interfere with the pathological tremor oscillations that presumably originate from the cerebellum and ascend, through the cZi, to the Vim and motor cortex (Blomstedt et al., 2009; Chazen et al., 2018; Groppa et al., 2014; Plaha et al., 2008). However, it has been suggested that the DBS effects extend beyond the stimulated target and influence the entire dysfunctional circuit (Fox et al., 2014). Indeed, DBS-related neural changes have been reported in cerebello-thalamo-cerebral regions both near and distant to the stimulated area in essential tremor patients with DBS in the Vim (Ceballos-Baumann et al., 2001; Gibson et al., 2016; Haslinger et al., 2003; Perlmutter et al., 2002). These previous studies were conducted in a small number of patients either intraoperatively during general anaesthesia or during resting conditions. DBS effects on the cerebello-thalamo-cerebral circuit during different motor tasks in awake and behaving essential tremor patients are unknown, especially in cZi DBS which, to date, has not been explored by functional neuroimaging. Since tremor in essential tremor is predominantly present during actions and not during rest (Bhatia et al., 2018), it is critical to investigate DBS effects during different motor tasks, in the presence and absence of tremor. We aimed, therefore, to investigate DBS effects on the activity of the cerebello-thalamo-cerebral circuit in essential tremor patients by means of task-based blood oxygen level-dependent (BOLD) functional MRI (fMRI). The first aim was to investigate whether DBS modulation of the sensorimotor cerebello-thalamo-cerebral circuit varies depending on the tasks, with and without tremor, i.e. task-dependent modulation. The second aim was to find out whether DBS in the cZi exerts general effects on the sensorimotor circuit regardless of the task at hand, i.e. taskindependent modulation. Essential tremor is characterised by postural tremor, but some patients may also exhibit intention tremor, i.e. tremor that increases in amplitude when approaching a target (Deuschl et al., 1998). Since intention tremor is only seen in a subgroup of patients and its presence has been specifically associated with cerebellar dysfunction (Deuschl et al., 2000; Zakaria et al., 2013), we aimed to investigate potentially specific DBS effects separately in patients exhibiting intention tremor. To address these questions, we used BOLD fMRI to measure brain activity during accelerometry-quantified postural and intention tremorprovoking tasks as well as during rest in awake essential tremor patients chronically implanted with DBS in the cZi.

2.2. Imaging data acquisition All scans were performed with a Philips Achieva dStream 1.5 T MR scanner using a transmit-receive (T/R) head coil, and average headspecific absorption rate (SAR) below 0.1 W/kg. Two fMRI time-series were collected per patient, one for each stimulation condition (On and Off stimulation). The first five volumes were discarded prior to each session to allow fMRI signal equilibrium. For each acquisition, 213 vol per session were acquired. Functional echo-planar imaging (EPI) runs were performed with the following parameters: 31 interleaved axial slices at a TR 3000 ms, TE 50 ms, voxel size 3.44  3.49  4.4 mm, 0.5 mm inter-slice gap, field of view (FOV) 220  220 mm, and matrix size 64  63. Axial T1-weighted structural scan was collected after the first functional session with the following acquisition parameters: 180 slices, no inter-slice gap, 1  1  1 mm voxel size, TR 7.4 s, TE 3.4 ms, flip angle 8 , FOV 256  232 mm, matrix size 256  232, acquisition time 11.32 min. 2.3. Experimental design and procedure Prior to scanning, patients were trained to perform the motor tasks to ascertain proper task performance. While lying supine in the MR scanner, patients looked at a screen located vertically in front of them at reaching distance. The screen was visualised using a double mirror mounted on the head coil, in order to avoid lateral inversion produced if using a single mirror. The instructions and motor commands were implemented in Eprime 2 (Psychology Software Tools, Sharpsburg, PA, USA), and projected on the screen. The patients performed tasks by alternating between the following blocks guided by the instructions on the screen: 1. Postural holding during which the patients were instructed to raise the right forearm and keep it outstretched against gravity for 20 s. A preparatory image (an outstretched arm) appeared 1 s before the gosignal and remained visible during the whole block. This task was designed to elicit postural tremor. 2. Pointing task during which patients were instructed to perform reaching movements with the right forearm and point with the index finger at a target viewed on the screen. The block started with a 5 s preparatory instruction followed by a 20 s block to perform the movements. Five reaching and pointing movements were preformed (the onset was indicated by the appearance of the target and a gosignal). The task was designed to elicit intention tremor, if present.

2. Material and methods 2.1. Patients and surgical procedure We included 16 patients with essential tremor (9 male; average age 70 years, range 52–80 years). All essential tremor patients with cZi DBS 2

A. Awad et al.

NeuroImage 209 (2020) 116511

Table 1 Patient demographics and stimulation parameters. Patient

Sex

Age

Handed-ness

Months since surgery

Family history

Active DBS lead during fMRI

Stimulation parameters (amplitude, pulse width, frequency)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16

M F F F F M M F M F M F M M M M

75 78 78 80 59 67 78 75 67 69 68 70 75 57 52 77

L R R R R R R R R R R R R R R R

27 70 36 54 43 17 34 37 36 11 44 59 59 26 50 17

Yes No Yes Yes Yes Yes Yes Yes No Uncertain Yes Yes No Yes Yes No

Right Left Left Left Right Left Left Left Left Left Left Left Left Left Left Left

2.3 2.7 1.2 1.3 1.5 1.8 1.6 2.3 1.8 1.5 1.6 1.8 2.2 2.3 2.5 1.7

V, V, V, V, V, V, V, V, V, V, V, V, V, V, V, V,

60 μs, 60 μs, 60 μs, 60 μs, 60 μs, 60 μs, 60 μs, 60 μs, 60 μs, 60 μs, 60 μs, 60 μs, 60 μs, 60 μs, 60 μs, 60 μs,

130 Hz 150 Hz 160 Hz 140 Hz 130 Hz 160 Hz 140 Hz 140 Hz 160 Hz 140 Hz 140 Hz 150 Hz 160 Hz 160 Hz 140 Hz 140 Hz

maximal tremor reduction without eliciting side effects according. To minimize head movements due to motor tasks, a band was applied on the right upper arm restricting the mobility of the shoulder joint (12 patients). To further stabilise the patients’ heads, bite bars fixed on the head coil were used when tolerated by patients (seven patients). These bite bars were custom made for each patient to match the patient’s own teeth prior to the scanning session. Furthermore, foam padding was used between the head and head coil for additional head-motion restriction. To register tremor and hand movements, an MR compatible singleaxis optical accelerometer attached to the ulnar side of the hand was used during the study except in the first three patients. Accelerometer data were primarily used to gain precise task onsets and durations to be modelled in fMRI analysis, as well as to quantify the tremor elicited by different motor tasks during scanning.

2.4. Data analyses 2.4.1. Accelerometer data analysis Accelerometer data were visualised and analysed by using an inhouse software (Zoom). Onsets times and durations of the tasks were obtained individually for each patient. The time periods for each block were identified visually from the acceleration signal corresponding to the beginning of the first hand-movement to the end of the last movement. For the three patients without accelerometer registration, onsets times and durations were calculated from instruction presentation time points during scanning (logged by E-prime). Following extraction of onsets and durations, accelerometer data were processed as following: (i) a band-pass filter 0.7–13.9 Hz (6 dB) was applied to exclude gravitational acceleration, (ii) a notch filter 9.4–11.3 Hz was used to eliminate the artefacts from the gradient magnets of the scanner, and (iii) to quantify the tremor intensity, the root mean square (RMS) during each condition and action state was calculated. To calculate pure postural tremor values, larger movements corresponding with forearm flexion and extension were identified visually and they were excluded from the analysis. Such approach was not applicable during the pointing task due to multiple pointing trials. Thus, tremor values from the pointing task were contaminated with slow armmovements. To calculate the percentage improvement in postural tremor, the difference between RMS during On and Off stimulation was divided by the RMS during Off stimulation, and then multiplicated by 100. The median improvement was calculated for the 13 patients, and the statistical significance for the difference between On and Off was tested with a t-test (tested against improvement by 0%).

Fig. 1. The DBS target (cZi): A preoperative axial MR-image fused with a postoperative CT for a representative patient, demonstrating the localisation of the tips of the DBS electrodes in the caudal zona incerta (cZi); posteromedial to the subthalamic nucleus (STN) at the level of the maximal diameter of the red nucleus. R ¼ right hemisphere.

3. Resting task where patients were instructed to keep their arms by the body on the scanner table and relax completely during 20 s. The different tasks were performed sequentially in the order stated above, and repeated 10 times each. For patients implanted with bilateral DBS electrodes (n ¼ 5), the electrode ipsilateral to tested arm was switched off during the whole experiment. Two patients performed leftsided motor tasks. The first patient had unilateral DBS for a predominantly left-sided tremor. The second patient had bilateral DBS, but preferred performing left-sided motor tasks due to cramps in the right hand, unrelated to the stimulation. All tasks were performed with the DBS, contralateral to the tested arm, turned On and Off in two subsequent sessions, with the initial stimulation setting counterbalanced across patients, i.e. half of the patients started the first session with DBS Off and the other half with DBS On. During the experiment, therapeutic stimulation parameters were used, which were previously optimised in our clinic to produce a

2.4.2. Image processing Before pre-processing, image data (T1 and functional volumes) from 3

A. Awad et al.

NeuroImage 209 (2020) 116511

two patients with active right-sided electrodes during fMRI were flipped with respect to the mid-sagittal plane in order to achieve consistent group activation maps regarding lateralisation. The results of this approach were checked to ascertain accordant activations in motor areas with other patients’ activation maps. Data were pre-processed and analysed by using SPM12 (Wellcome Trust Centre for Neuroimaging, London, UK) implemented in MATLAB 7.11 (MathWorks). Batching of SPM routines and result visualisation were performed with a software developed in-house (DataZ). Data were slice-timing corrected, unwarped and spatially realigned to the first scan by rigid body transformations to correct for head movements, and coregistered to the T1-weighted anatomical image. Ultra-low frequency fluctuations were removed by applying a high-pass filter (cut-off ¼ 128 s). In all patients, maximal displacements due to head motion were <2.5 mm. A group-specific anatomical template was created from the individual 16 T1-weighted images using DARTEL (diffeomorphic anatomical registration through exponentiated lie algebra) for a more precise intersubject alignment (Ashburner, 2007). Individual functional data were normalised to the MNI anatomical space using the group-specific DARTEL template, and then spatially smoothed by using an 8-mm full-width half maximum (FWHM) Gaussian kernel. A group-specific anatomical image was created by averaging individual normalised T1-weighted images and then used to create a group-specific rendered brain image. Following the pre-processing routine, data were inspected to ascertain that artefacts induced by the DBS hardware did not impact normalisation to MNI space or distort the cortical areas of interest. The hardware-related artefacts resulted in signal loss surrounding the electrode, and the extension cables sited over the left parietal cortex (Supplementary Fig. 1).

DBS modulation in the case of pointing was calculated as DBS x motor task interaction effects [first factor: task (pointing, rest)], second factor [DBS setting (Off, On)]. This analysis was intended to be performed on a subgroup of patients exhibiting intention tremor, but was instead performed for the whole group since isolating patients with intention tremor was not possible (see below). To explore general DBS effects on the sensorimotor circuit, i.e. task-independent modulation, the main effects of DBS were calculated from the above-mentioned ANOVAs. The binary motor network mask was used (as described above) to focus the analyses to the cerebello-thalamo-cerebral circuit based on our a priori hypothesis regarding regions of importance for essential tremor. Importantly, the contrast used to create the mask is orthogonal to the test contrasts used later in ANOVA, which, combined with the balanced design matrix ensures that there is no statistical dependency between the generation of the motor mask and subsequent ANOVAs (Kriegeskorte et al., 2009). Regions were considered significant at the threshold of p
0.05, FWE cluster-corrected for multiple comparisons within the motor mask (cluster-defining threshold was set to an uncorrected voxel-based threshold of p
0.001). For completeness, we also report the results from whole-brain analyses. Percent signal changes were extracted from significant clusters from the ANOVAs for each region and condition, averaged within each cluster, and then used for post hoc repeated measures two-tailed t-tests of the directionality of effects and the difference between Off and On conditions. Effect sizes from each cluster were visualised as percent BOLD signal change benchmarked to BOLD signal during the resting condition Off stimulation according to the formula: [(βtask-βrest_Off)/βconstant] x 100, where βs are the regression coefficients for different experimental conditions (tasks) from the first-level analyses and βconstant is the mean intensity across the session.

2.4.3. fMRI data analyses fMRI data were analysed voxel by voxel by using a GLM (general linear model). Analyses were set up by including the experimental conditions for each task during DBS On as well as during DBS Off (postural, pointing and rest) as boxcar regressors of interest, convolved with the canonical haemodynamic response function. To account for residual head motion artefacts, six motion parameters estimated during realignment were included as nuisance regressors in the regression model. The linear regression analyses at the subject level produced betavalue images for each condition, which were used to calculate contrast images used as input in repeated-measures random-effects ANOVAs. As we were primarily interested in exploring DBS effects within the cerebello-thalamo-cerebral circuit, we created a binary mask based on a group-level network of brain regions engaged by the motor tasks (“motor network”). This approach increases the power of detecting effects. The mask was created from a one sample t-test in a random-effects model (calculated by contrasting postural holding and pointing with rest, regardless of DBS setting, p
0.0001, uncorrected with an extent threshold of 20 contiguous voxels). This threshold was chosen ad hoc to specifically include all parts of the cerebello-thalamo-cerebral network and little else. The mask consisted of all active regions within the sensorimotor cortices, the thalamus, and the cerebellum. The cerebral and thalamic clusters were extracted directly as two active separate clusters. Given the visual nature of the tasks, BOLD signal change was also observed in the occipital visual areas as one cluster continuous with cerebellar regions. To selectively extract cerebellar regions and exclude the occipital cortex, we used the SUIT cerebellum template (Diedrichsen, 2006) as an including mask, such that cerebellar regions, within the mask, were included and extracted, whereas occipital regions were not. The task-dependent DBS modulation in the case of postural holding was calculated as DBS x motor task interaction effects through randomeffects repeated-measures 2  2 ANOVA [first factor: task (postural holding, rest)], second factor [DBS setting (Off, On)]. A factorial design in SPM was used and the measurements were assumed to be dependent between levels but with unequal variance. Similarly, the task-dependent

2.5. Data availability statement All data supporting the conclusions of this manuscript will be made available by the authors upon request to any qualified researcher. 3. Results DBS resulted in contralateral reductions in postural tremor measured by the accelerometer in 13 patients during the fMRI (range: 31–98%, median: 87%), p < 0.05 (tested against improvement by 0%). The main tremor frequencies observed during Off stimulation covered 3–8 Hz according to a Fast Fourier Transform. Intention tremor is a tremor with increasing amplitude during movements towards a target (Deuschl et al., 1998). To identify a subgroup of patients exhibiting intention tremor, we attempted to measure a potential increase in the root mean square of the acceleration at the end of each trial in the pointing task. Due to several pointing trials (five) during each block (20 s), not allowing for sufficient time at the target, the pointing task was unfortunately not adequate to distinguish intention tremor clearly. Nevertheless, there was a group-wise reduction in the root mean square that ranged from 6 to 96% (median: 61%), p < 0.05, likely representing a mix of unspecific kinetic and intention tremor improvement. Performing the motor tasks relative to rest, regardless of DBS stimulation setting, was associated with an expected BOLD signal increase in the sensorimotor and visual regions (p
0.0001, uncorrected, k ¼ 20). The resultant motor network included the contralateral primary sensorimotor cortex, premotor cortices, SMA proper, thalamus and bilateral cerebellum (lobules right IV, V, VI, vermis, VIII and left VI) (Fig. 2). 3.1. Task-dependent DBS modulation Task-dependent modulation in the case of postural holding (DBS  task interactions) was observed in four regions: left primary sensorimotor 4

A. Awad et al.

NeuroImage 209 (2020) 116511

Fig. 2. The motor network mask as created from regions activated by right arm movements: A group level network of brain regions engaged by the motor tasks as compared to rest (postural holding and pointing task > Rest, p
0.0001, uncorrected, k ¼ 20) regardless of DBS setting, is visualised on a group-specific rendered anatomical image. R ¼ right hemisphere.

pointing conditions (Fig. 4B). For completeness, we also report results from whole-brain analyses (i.e. without using the cerebello-thalamo-cerebral mask). In the case of the postural task, task  DBS interaction effects were found in the right precuneus, bilateral cuneus and right superior and middle temporal gyrus (Supplementary Fig. 2A), in addition to the four sensorimotor clusters reported above. Although the interaction effect in these three regions was significant, there were no significant differences between DBS settings (Supplementary Fig. 2B). No additional results were observed in wholebrain analyses for the other conditions. There were no significant DBS-related effects in the thalamus. Since the thalamus is postulated to be of central importance in essential tremor, BOLD signal change in the thalamus was further explored. BOLD signal change from the thalamic cluster within the motor network was extracted (a region of interest was defined from the thalamic local maximum in the motor network defined above, MNI -16 -20 8 with a 5 mm radius) and plotted for each condition to visualise DBS effects in this cluster (Fig. 5). While BOLD signal in the thalamus increased during motor tasks (postural holding and pointing) relative to rest during Off as well as during On stimulation, there were no significant differences between DBS settings.

cortex, SMA proper, right anterior cerebellum (lobule IV and V) and posterior cerebellum (lobule VIII), (p
0.05, FWE cluster-corrected for multiple comparisons), see Table 2 and Fig. 3. The cerebral effects were located contralateral, whereas cerebellar effects were ipsilateral, to arm movements. Effect sizes from these regions, expressed as percent signal change benchmarked to BOLD during rest in DBS Off, are shown in Fig. 3. Post hoc t-tests (p
0.05, uncorrected) were used to determine the experimental condition during which DBS was driving the interaction effect. Two patterns emerged; BOLD signal decrease during the postural holding when DBS was turned on in the primary sensorimotor cortex (p ¼ 0.02) and cerebellar lobule VIII (p ¼ 0.04), and BOLD signal increase during rest when DBS was turned on in the SMA proper (p ¼ 0.009) and cerebellar lobule V (p ¼ 0.04). As we were not able to isolate a subset of patients exhibiting intention tremor, whole-group DBS x pointing-task interaction effects were explored. There were no statistically significant effects (p
0.05, FWE cluster-corrected).

3.2. Task-independent DBS modulation Task-independent DBS modulation, calculated as main effects of DBS, was observed in the left PMC (ipsilateral to DBS) both in the case of postural holding and pointing (postural holding: MNI coordinates -34 -8 52, F ¼ 30.9, k ¼ 5, p
0.05 FWE voxel-corrected; pointing: MNI coordinates -36 -8 52, F ¼ 25.7, k ¼ 40 p
0.05, FWE cluster-corrected) (Fig. 4A). The main effect of DBS in the case of postural holding did not survive correcting for multiple comparisons at the cluster level, but was statistically significant when correcting for multiple comparisons at the voxel level. The percent signal change in this region (PMC) was plotted for all conditions and showed that DBS exerted a similar effect in all tasks, i.e. BOLD signal increase during rest, postural holding and

4. Discussion In the present study, DBS in the cZi was demonstrated to exert both task-dependent as well as task-independent effects on the sensorimotor circuit. Task-dependent (DBS  task interaction) effects were seen in sensorimotor cerebello-cerebral regions: the primary sensorimotor cortex, SMA proper and cerebellum. Differential DBS effects were found depending on whether the patients performed tremor-inducing postural holding or rest. Specifically, BOLD signal in the primary sensorimotor cortex and cerebellar lobule VIII decreased when performing postural holding while DBS was turned on. In contrast, BOLD signal in the SMA proper and cerebellar lobule V increased during the resting condition when DBS was turned on. Task-independent effect was observed as activity increase in the lateral premotor cortex during all motor tasks and during rest. DBS modulation of cerebello-cerebral nodes distant from the stimulated target is consistent with the extensive afferent and efferent connections between the cZi and the cerebral cortex, thalamus, cerebellum, basal ganglia and brain stem (Blomstedt et al., 2009; Mitrofanis, 2005), and is in agreement with previous reports indicating distant, rather than only local, effects of DBS in the thalamus and subthalamic nucleus (Ceballos-Baumann et al., 2001; Gibson et al., 2016; Haslinger et al., 2003; Kahan et al., 2014, 2012; Perlmutter et al., 2002).

Table 2 DBS and task interaction effects in the case of postural holding. Region

Cluster size (voxels)

Local maxima

F

Coordinates in MNI space x, y, z (mm)

R anterior cerebellum

439

R cerebellum V R cerebellum IV SMA 1 SMA 2 Precentral gyrus Postcentral gyrus Postcentral gyrus –

31.9

20 -44 -20

24.6

6 -54 -18

29.8 17.6 26.6

2 -8, 72 18 -8 62 32 -22 64

25.5

30 -36 52

17.4

22 -36 70

21.7

22 -68 -54

Supplementary motor area (SMA proper) L primary sensorimotor cortex

R posterior cerebellum (lobule VIII)

210 426

89

4.1. Task-dependent DBS modulation In the case of postural holding, DBS had task-specific (interaction) effects in the cerebellum, the SMA and the primary sensorimotor cortex. Accumulating evidence, originating from structural (Bagepally et al.,

p
0.05, FWE-corrected; L ¼ left; R ¼ right. 5

A. Awad et al.

NeuroImage 209 (2020) 116511

Fig. 3. DBS and task interaction effects in the case of postural holding (task-dependent DBS modulation): Brain regions with significant interaction effects (DBS x motor task) on a group-specific rendered anatomical image (p
0.05, FWE-corrected), and bar plots with percent signal changes as benchmarked to the rest condition Off stimulation in each region. Brain regions and bar plots showing increased BOLD signal due to DBS are red-coloured, whereas regions and plots with decreased BOLD signal due to DBS are blue-coloured. R ¼ right hemisphere; * ¼ p
0.05; NS ¼ not significant (p > 0.05). Error bars represent SEM.

n et al., 2009; Cerasa and Quattrone, 2016; Gallea et al., 2012; Benito-Leo 2015) and functional (Boecker and Brooks, 1998; Broersma et al., 2016; Bucher et al., 1997) brain imaging studies as well as pathological examinations (Louis, 2016), indicates a possible primary pathological role of the cerebellum in essential tremor. These structural and functional abnormalities result in tremor oscillations which are transferred to the

sensorimotor cortex through the posterior subthalamic area and the thalamus (Helmich et al., 2013). DBS is proposed to interfere with these tremor oscillations at the level of the cZi or the part of the Vim with the highest connectivity to the contralateral dentate nucleus (Akram et al., 2018). The current cerebellar DBS effects lie within the sensorimotor 6

A. Awad et al.

NeuroImage 209 (2020) 116511

Fig. 4. Main effects of DBS (task-independent DBS modulation): (A) Task-independent DBS modulation, calculated as main effect of DBS, in the left premotor cortex (ipsilateral to DBS) in the case of postural holding in blue and pointing in red (p
0.05 FWE-corrected) and the anatomical overlap is in purple. For illustrative purposes, the main effect for postural holding is presented at p
0.001 uncorrected to more easily compare with effects from the pointing task. Percent signal change as benchmarked to the rest condition Off stimulation in the case of postural holding and pointing in (B). Coordinates (z) are in the MNI space; R ¼ right hemisphere. Error bars represent SEM.

activity increase during rest, no further recruitment of the SMA is evident when patients perform the postural holding during On stimulation (i.e. SMA is not a part of the motor network, defined as motor task > rest, during On stimulation). The DBS effect in the primary sensorimotor cortex (decrease during postural holding due to DBS) is compatible with results from previous functional imaging studies showing increased activity in the sensorimotor cortex during tremor-provoking motor tasks in essential tremor patients as compared to controls (Broersma et al., 2016; Bucher et al., 1997; Hallett and Dubinsky, 1993; Jenkins et al., 1993; Neely et al., 2015). In these studies, abnormal motor/tremor-related activity in primary sensorimotor and other regions was evident even when compared with passive tremor (passive wrist oscillations induced by an experimenter) in patients as well as with passive tremor and “mimicked” tremor in controls. The control tasks were designed to reproduce the proprioceptive input resulting from involuntary tremor. Thus, altered sensorimotor activity in patients is not only due to different proprioceptive tremor input but also represents underlying functional abnormality, and it seems that DBS normalises this increase in primary sensorimotor cortex activity. The overall effect of the task-dependent DBS modulation can be interpreted as differences at the effector level and may reflect DBSinduced tremor reduction since the cerebellum, SMA and primary sensorimotor cortex exhibit less motor task-activity as compared to the resting condition during On stimulation. This was achieved either through reduction during the active motor task (postural holding in the primary sensorimotor cortex and cerebellar lobule VIII) or through increases in resting/baseline activity (cerebellar lobule V and the SMA).

Fig. 5. BOLD signal change in the thalamus: BOLD signal change from the thalamic cluster within the motor network (MNI -16 -20 8) is plotted for each condition. R ¼ right hemisphere; NS ¼ not significant (p > 0.05). Error bars represent SEM.

cerebellar lobules (IV/V and VIII) (Buckner et al., 2011; Kelly and Strick, 2003; Stoodley et al., 2012), that also have been reported to be structurally (Gallea et al., 2015) and functionally (Broersma et al., 2016; Buijink et al., 2015b; Muthuraman et al., 2018) disturbed in essential tremor. The DBS-related effects in the cerebellum might represent a modulation of primary pathological cerebellar activity, especially the BOLD signal decrease in cerebellar lobule VIII during postural holding when the DBS is turned on. Indeed, repetitive transcranial magnetic stimulation (rTMS) over the posterior cerebellum has been shown to supress tremor and the improvement was sustained for three weeks when the TMS was applied repeatedly for five sessions (Popa et al., 2013). The DBS effect in the SMA (increased activity during rest due to DBS) is consistent with the increased blood flow in the SMA due to thalamic stimulation during rest as reported in an early PET study (Perlmutter et al., 2002), and also the intraoperative fMRI study by Gibson et al. (2016). The SMA has gained more attention in recent studies dealing with essential tremor pathophysiology and it has been assigned a compensatory role in mitigating tremulous activity that originates from the cerebellum. This notion has emerged as the SMA has been shown to exhibit grey matter increase, as opposed to cerebellar atrophy, in essential tremor (Gallea et al., 2015). Our finding of DBS-associated BOLD signal increase in the SMA during rest cannot corroborate a compensatory function of the SMA. Interestingly though, as DBS induces

4.2. Task-independent DBS modulation The main effect of DBS was located in the left lateral PMC, ipsilateral to the stimulation. Common across all tasks, including rest, DBS resulted in an increase of BOLD in the PMC. Premotor regions have been considered to be part of the dysfunctional synchronised network involved in tremor generation. EEG/MEG studies reported coherence between central oscillations over the PMC and peripheral tremor oscillations (Muthuraman et al., 2012; Schnitzler et al., 2009). In functional neuroimaging studies, the PMC has been shown to exhibit abnormalities in blood flow and BOLD signal during rest and motor tasks (Buijink et al., 2015a; Colebatch et al., 1990; Jenkins et al., 1993), and decreased functional connectivity with the cerebellum (Lenka et al., 2017; Neely et al., 2015). Moreover, dysfunctional GABAergic neurotransmission, as shown by increased 11C-flunazenil binding to GABA-receptors, has been 7

A. Awad et al.

NeuroImage 209 (2020) 116511

Here, the main effect of DBS was an increase in the PMC activity regardless of the motor state of the patients, which we interpret as DBSinduced potentiation (restoration?) of premotor control on the sensorimotor circuit, making it less susceptible to tremor entrainment. This finding is consistent with the direction of the signals responsible for intermittent motor control, i.e. signals originating from the cerebellum and reaching the primary motor cortex after passing through the thalamus and the PMC (Gross et al., 2002). Furthermore, therapeutic modulation of the PMC is supported by a recent probabilistic tractography study reporting that stimulating the part of the thalamus with the highest connectivity to premotor regions is associated with better tremor control (Middlebrooks et al., 2018, but see Akram et al., 2019). On the other hand, the task-dependent DBS effects in the primary sensorimotor cortex, SMA and cerebellum represent differences at the effector level and may reflect DBS-induced tremor reduction since these cerebello-cerebral regions exhibit less motor task-activity as compared to the resting condition during On stimulation.

observed in the PMC, along with the cerebellum and the ventrolateral thalamus (Boecker et al., 2010). Although the role of the premotor regions in essential tremor is still unclear, it is reasonable for a therapeutic modulation to occur through or in orchestration with the PMC. The premotor regions are involved in higher-order integrative sensorimotor control such as motor planning, sensory guidance of movements and movement corrections though connections to the primary motor cortex and to the spinal cord (Chouinard and Paus, 2006; Dum and Strick 2004; Picard and Strick, 2001). Dysfunctional premotor control might be part of the pathophysiology in essential tremor manifesting as decreased/inadequate inhibitory control that makes the system more susceptible to oscillations. Speculatively, the DBS effects may be partly driven by modulating (restoring?) PMC activity which makes the sensorimotor system more resistant to perturbation by tremor oscillations. 4.3. Negative findings In the thalamus, there were no DBS-related effects. Two previous functional imaging studies have shown local increase in blood flow and BOLD signal change in the thalamus due to DBS (Gibson et al., 2016; Perlmutter et al., 2002). The failure to replicate these results may be due to different experimental designs and, more importantly, a different DBS target in our study (cZi). Another possibility is related to signal drop-out around the lead which is evident here and in previous fMRI studies. For example, it was not possible to record BOLD data from the subthalamic nucleus itself in a study by Kahan et al., where they explored DBS effects in Parkinson’s disease (Kahan et al., 2014). Potential DBS effects might, thus, be located within the area of signal drop-out. However, more posterior parts of the thalamus were not affected by signal dropout and showed clear signal change when comparing motor tasks with rest, as shown in Figs. 2 and 5. There were no DBS  task interaction effects during the pointing condition, perhaps because we could not isolate intention tremor, as we initially aimed to do. This was probably a result of a suboptimal design with too many pointing trials within each block, not allowing sufficient time for intention tremor to develop. There was, however, a group-wise reduction in the root mean square that likely represents a mix of kinetic and intention tremor improvement. Interaction effects are probably more dependent on optimal task design and performance as compared to main effects. Indeed, main effects of DBS (regardless of motor tasks) were evident despite this limitation.

4.5. Limitations Compared to other studies combining functional neuroimaging and DBS, we included a relatively large number of patients. Nevertheless, the sample size is small, and the study might be underpowered to detect other effects of interest. Underpowered studies might also overestimate detected effect sizes (Button et al., 2013). However, the analytical approach of restricting the search volume to brain regions known a priori to be involved in essential tremor, was aimed to increase the power of detecting treatment effects. Hardware-related artefacts from the electrode and the extension cables precluded the possibility of collecting high-quality data from the cZi/thalamus and parts from the left parietal cortex. BOLD signal from well-described sensorimotor-related regions were, however, still detectable. For safety reasons, we adhered to a strict experimental setup and used imaging parameters to keep the SAR values below 0.1 W/kg. As most previous fMRI-DBS studies, we used a 1.5T scanner and T/R head coil to reduce radiofrequency exposure to the DBS system. Low fieldstrength and T/R coil are not commonly used in fMRI studies nowadays as they offer lower signal-to-noise ratio. The spatial resolution was therefore relatively low (voxel size 3.44  3.49  4.4 mm) compared to commonly used fMRI parameters. While such spatial resolution does not match the precision needed for precise DBS electrode placement, it is sufficient to capture cortical and subcortical changes resulting from DBS. Also, the temporal resolution of fMRI precludes detecting fast effects, such as potentially relevant tremor oscillation-related activity. Despite these limitations, the study provides a novel approach to explore DBS effects with fMRI in awake patients during different task states, i.e. with and without tremor.

4.4. Integrated network view on DBS effects The dysfunctional cerebello-thalamo-cerebral circuit in essential tremor is normally responsible for sensorimotor control and has been shown to generate motor output as small movement discontinuities at a frequency of 8–10 Hz (Gross et al., 2002; Vallbo and Wessberg, 1993). The efferent drive of this normal physiological tremor originates from the primary motor cortex but is sustained by a cerebellar drive through the thalamus to the PMC before reaching the primary motor cortex (Gross et al., 2002). Motor output discontinuities (physiological tremor) represent alternating agonist and antagonist bursts, the timing and the amplitude of which are controlled by the cerebellum to produce smooth and coordinated movements (Filip et al., 2016; Schnitzler and Gross, 2005). Essential tremor is driven by dysfunction in the cerebello-thalamocerebral circuit resulting in pathological tremor oscillations that are probably transferred from the cerebellum to the motor cortex (Helmich et al., 2013; Hopfner and Deuschl, 2018; Plaha et al., 2008). These oscillations are modulated by DBS in the Vim or cZi which represent the bottleneck of the cerebello-thalamo-cerebral circuit. Why this circuit is susceptible to oscillations in essential tremor is not known but has been postulated to be rooted in the decreased GABAergic neurotransmission in the PMC, thalamus and cerebellum (Boecker et al., 2010).

5. Conclusions We show that DBS in the cZi exerts both task-dependent as well as task-independent effects on the sensorimotor circuit. Task-dependent modulation resulted in two patterns of DBS effects; activity decrease during tremor-inducing postural holding in the primary sensorimotor cortex and cerebellar lobule VIII, and activity increase in the SMA and cerebellar lobule V during rest. The task-dependent effects represent differences at the effector level and may reflect DBS-induced tremor reduction since the cerebellum, SMA and primary sensorimotor cortex exhibit less motor task-activity as compared to the resting condition during On stimulation. Task-independent DBS modulation was observed as increased activity during all motor tasks and during rest in the lateral PMC. This task-independent modulation may mediate the therapeutic effects of DBS through the facilitation of the premotor control over the sensorimotor circuit making it less susceptible to tremor entrainment. Our findings support the notion of DBS acting upon modulation of the sensorimotor cerebello-cerebral circuit distant to the stimulated target in 8

A. Awad et al.

NeuroImage 209 (2020) 116511

essential tremor and illustrate the complexity of DBS mechanisms by demonstrating task-dependent as well as task-independent actions in cerebello-cerebral regions.

and movement disorder society. Mov. Disord. 33, 75–87. https://doi.org/10.1002/ mds.27121. Blomstedt, P., Sandvik, U., Fytagoridis, A., Tisch, S., 2009. The posterior subthalamic area in the treatment of movement disorders: past, present, and future. Neurosurgery 64, 1029–1038. https://doi.org/10.1227/01.NEU.0000345643.69486.BC. Blomstedt, P., Sandvik, U., Tisch, S., 2010. Deep brain stimulation in the posterior subthalamic area in the treatment of essential tremor. Mov. Disord. 25, 1350–1356. https://doi.org/10.1002/mds.22758. Boecker, H., Brooks, D.J., 1998. Functional imaging of tremor. Mov. Disord. 13, 64–72. https://doi.org/10.1002/mds.870131311. Boecker, H., Weindl, A., Brooks, D.J., Ceballos-Baumann, A.O., Liedtke, C., Miederer, M., Sprenger, T., Wagner, K.J., Miederer, I., 2010. GABAergic dysfunction in essential tremor: an 11C-flumazenil PET study. J. Nucl. Med. 51, 1030–1035. https://doi.org/ 10.2967/jnumed.109.074120. Boecker, H., Wills, A.J., Ceballos-Baumann, A., Samuel, M., Thompson, P.D., Findley, L.J., Brooks, D.J., 1996. The effect of ethanol on alcohol-responsive essential tremor: a positron emission tomography study. Ann. Neurol. 39, 650–658. https://doi.org/ 10.1002/ana.410390515. Broersma, M., van der Stouwe, A.M.M., Buijink, A.W.G., de Jong, B.M., Groot, P.F.C., Speelman, J.D., Tijssen, M.A.J., van Rootselaar, A.-F., Maurits, N.M., 2016. Bilateral cerebellar activation in unilaterally challenged essential tremor. NeuroImage Clin 11, 1–9. https://doi.org/10.1016/j.nicl.2015.12.011. Bucher, S.F., Seelos, K.C., Dodel, R.C., Reiser, M., Oertel, W.H., 1997. Activation mapping in essential tremor with functional magnetic resonance imaging. Ann. Neurol. 41, 32–40. https://doi.org/10.1002/ana.410410108. Buckner, R.L., Krienen, F.M., Castellanos, A., Diaz, J.C., Yeo, B.T.T., 2011. The organization of the human cerebellum estimated by intrinsic functional connectivity. J. Neurophysiol. 106, 2322–2345. https://doi.org/10.1152/jn.00339.2011. Buijink, A.W.G., Broersma, M., van der Stouwe, A.M.M., van Wingen, G.A., Groot, P.F.C., Speelman, J.D., Maurits, N.M., van Rootselaar, A.F., 2015a. Rhythmic finger tapping reveals cerebellar dysfunction in essential tremor. Park. Relat. Disord. 21, 383–388. https://doi.org/10.1016/j.parkreldis.2015.02.003. Buijink, A.W.G., van der Stouwe, A.M.M., Broersma, M., Sharifi, S., Groot, P.F.C., Speelman, J.D., Maurits, N.M., van Rootselaar, A.-F., 2015b. Motor network disruption in essential tremor: a functional and effective connectivity study. Brain 138, 2934–2947. https://doi.org/10.1093/brain/awv225. Button, K.S., Ioannidis, J.P.A., Mokrysz, C., Nosek, B.A., Flint, J., Robinson, E.S.J., Munaf o, M.R., 2013. Power failure: why small sample size undermines the reliability of neuroscience. Nat. Rev. Neurosci. 14, 365–376. https://doi.org/10.1038/nrn3475. Ceballos-Baumann, A., Boecker, H., Fogel, W., Alesch, F., Bartenstein, P., Conrad, B., Diederich, N., von Falkenhayn, I., Moringlane, J.R., Schwaiger, M., Tronnier, V.M., 2001. Thalamic stimulation for essential tremor activates motor and deactivates vestibular cortex. Neurology 56, 1347–1354. https://doi.org/10.1212/ WNL.56.10.1347. Cerasa, A., Quattrone, A., 2016. Linking essential tremor to the cerebellum—neuroimaging evidence. Cerebellum 15, 263–275. https://doi.org/ 10.1007/s12311-015-0739-8. Chazen, J.L., Sarva, H., Stieg, P.E., Min, R.J., Ballon, D.J., Pryor, K.O., Riegelhaupt, P.M., Kaplitt, M.G., 2018. Clinical improvement associated with targeted interruption of the cerebellothalamic tract following MR-guided focused ultrasound for essential tremor. J. Neurosurg. 129, 315–323. https://doi.org/10.3171/2017.4.JNS162803. Chouinard, P.A., Paus, T., 2006. The primary motor and premotor areas of the human cerebral cortex. Neuroscience 12, 143–152. https://doi.org/10.1177/ 1073858405284255. Colebatch, J.G., Findley, L.J., Frackowiak, R.S., Marsden, C.D., Brooks, D.J., 1990. Preliminary report: activation of the cerebellum in essential tremor. Lancet 336, 1028–1030. https://doi.org/10.1016/0140-6736(90)92489-5. Deuschl, G., Bain, P., Brin, M., 1998. Consensus statement of the movement disorder society on tremor. Mov. Disord. 13 (Suppl. 3), 2–23. https://doi.org/10.1002/ mds.870131303. Deuschl, G., Raethjen, J., Hellriegel, H., Elble, R., 2011. Treatment of patients with essential tremor. Lancet Neurol. 10, 148–161. https://doi.org/10.1016/S14744422(10)70322-7. Deuschl, G., Wenzelburger, R., L€ offler, K., Raethjen, J., Stolze, H., 2000. Essential tremor and cerebellar dysfunction clinical and kinematic analysis of intention tremor. Brain 123, 1568–1580. https://doi.org/10.1093/brain/123.8.1568. Diedrichsen, J., 2006. A spatially unbiased atlas template of the human cerebellum. Neuroimage 33, 127–138. https://doi.org/10.1016/j.neuroimage.2006.05.056. Dum, R.P., Strick, P.L., 2004. Motor areas in the frontal lobe: the anatomical substrate for the central control of movement. In: Motor Cortex in Voluntary Movements: A Distributed System for Distributed Functions, pp. 1–46. Filip, P., Lungu, O.V., Manto, M.-U., Bares, M., 2016. Linking essential tremor to the cerebellum: physiological evidence. Cerebellum 15, 774–780. https://doi.org/ 10.1007/s12311-015-0740-2. Fox, M.D., Buckner, R.L., Liu, H., Chakravarty, M.M., Lozano, a.M., Pascual-Leone, A., 2014. Resting-state networks link invasive and noninvasive brain stimulation across diverse psychiatric and neurological diseases. Proc. Natl. Acad. Sci. 111, E4367–E4375. https://doi.org/10.1073/pnas.1405003111. Fytagoridis, A., Sandvik, U., Åstr€ om, M., Bergenheim, T., Blomstedt, P., 2012. Long term follow-up of deep brain stimulation of the caudal zona incerta for essential tremor. J. Neurol. Neurosurg. Psychiatry 83, 258–262. https://doi.org/10.1136/jnnp-2011300765. Gallea, C., Popa, T., García-Lorenzo, D., Valabregue, R., Legrand, A.-P., Marais, L., Degos, B., Hubsch, C., Fernandez-Vidal, S., Bardinet, E., Roze, E., Lehericy, S., Vidailhet, M., Meunier, S., 2015. Intrinsic signature of essential tremor in the cerebello-frontal network. Brain 138, 2920–2933. https://doi.org/10.1093/brain/awv171.

Author contribution statement Amar Awad: conceptualization, methodology, data collection, data analysis, writing the initial draft, reviewing and revising manuscript. Patric Blomstedt: conceptualization, reviewing and revising manu€ ran Westling: methodology, data analysis, script, funding acquisition. Go reviewing and revising manuscript. Johan Eriksson: conceptualization, methodology, reviewing and revising the manuscript, funding acquisition. Funding This study was supported by grants from the county of V€asterbotten to P.B and from the medical faculty at Umeå University to J.E. Declaration of competing interests The authors declare the following financial interests/personal relationships which may be considered as potential competing interests: Patric Blomstedt is a consultant for Abbott, Medtronic and Boston Scientific, and a shareholder in Mithridaticum AB. Amar Awad, G€ oran Westling and Johan Eriksson declare no conflicts of interest. Acknowledgements We would like to thank all the patients who have kindly taken part in this study; AnnaKarin Kronhamn for recruitment of patients and assistance during the experiments; Lars Nyberg for comments on the experimental design and the manuscript; Roland S Johansson for comments on the manuscript; Anders B€ackstr€ om and Micael Andersson for technical support and Anders Lundquist for statistical support. Appendix A. Supplementary data Supplementary data to this article can be found online at https://doi. org/10.1016/j.neuroimage.2019.116511. References Akram, H., Dayal, V., Mahlknecht, P., Georgiev, D., Hyam, J., Foltynie, T., Limousin, P., De Vita, E., Jahanshahi, M., Ashburner, J., Behrens, T., Hariz, M., Zrinzo, L., 2018. Connectivity derived thalamic segmentation in deep brain stimulation for tremor. NeuroImage Clin 18, 130–142. https://doi.org/10.1016/j.nicl.2018.01.008. Akram, H., Hariz, M., Zrinzo, L., 2019. Connectivity derived thalamic segmentation: separating myth from reality. NeuroImage Clin 101758. https://doi.org/10.1016/ j.nicl.2019.101758. Ashburner, J., 2007. A fast diffeomorphic image registration algorithm. Neuroimage 38, 95–113. https://doi.org/10.1016/j.neuroimage.2007.07.007. Bagepally, B.S., Bhatt, M.D., Chandran, V., Saini, J., Bharath, R.D., Vasudev, M., Prasad, C., Yadav, R., Pal, P.K., 2012. Decrease in cerebral and cerebellar gray matter in essential tremor: a voxel-based morphometric analysis under 3T MRI. J. Neuroimaging 22, 275–278. https://doi.org/10.1111/j.1552-6569.2011.00598.x. Barbe, M.T., Reker, P., Hamacher, S., Franklin, J., Kraus, D., Dembek, T.A., Becker, J., Steffen, J.K., Allert, N., Wirths, J., Dafsari, H.S., Voges, J., Fink, G.R., VisserVandewalle, V., Timmermann, L., 2018. DBS of the PSA and the VIM in essential tremor. Neurology 91, e543–e550. https://doi.org/10.1212/ WNL.0000000000005956. Benito-Le on, J., Alvarez-Linera, J., Hernandez-Tamames, J.A., Alonso-Navarro, H., Jim enez-Jim enez, F.J., Louis, E.D., 2009. Brain structural changes in essential tremor: voxel-based morphometry at 3-Tesla. J. Neurol. Sci. 287, 138–142. https://doi.org/ 10.1016/j.jns.2009.08.037. Bhalsing, K.S., Saini, J., Pal, P.K., 2013. Understanding the pathophysiology of essential tremor through advanced neuroimaging: a review. J. Neurol. Sci. 335, 9–13. https:// doi.org/10.1016/j.jns.2013.09.003. Bhatia, K.P., Bain, P., Bajaj, N., Elble, R.J., Hallett, M., Louis, E.D., Raethjen, J., Stamelou, M., Testa, C.M., Deuschl, G., 2018. Consensus statement on the classification of tremors. From the task force on tremor of the international Parkinson 9

A. Awad et al.

NeuroImage 209 (2020) 116511 Muthuraman, M., Heute, U., Arning, K., Anwar, A.R., Elble, R., Deuschl, G., Raethjen, J., 2012. Oscillating central motor networks in pathological tremors and voluntary movements. What makes the difference? Neuroimage 60, 1331–1339. https:// doi.org/10.1016/j.neuroimage.2012.01.088. Muthuraman, M., Raethjen, J., Koirala, N., Anwar, A.R., Mideksa, K.G., Elble, R., Groppa, S., Deuschl, G., 2018. Cerebello-cortical network fingerprints differ between essential, Parkinson’s and mimicked tremors. Brain 141, 1770–1781. https:// doi.org/10.1093/brain/awy098. Neely, K. a, Kurani, A.S., Shukla, P., Planetta, P.J., Wagle Shukla, A., Goldman, J.G., Corcos, D.M., Okun, M.S., Vaillancourt, D.E., 2015. Functional brain activity relates to 0–3 and 3–8 Hz force oscillations in essential tremor. Cerebr. Cortex 25, 4191–4202. https://doi.org/10.1093/cercor/bhu142. Perlmutter, J.S., Mink, J.W., Bastian, a J., Zackowski, K., Hershey, T., Miyawaki, E., Koller, W., Videen, T.O., 2002. Blood flow responses to deep brain stimulation of thalamus. Neurology 58, 1388–1394. https://doi.org/10.1212/WNL.58.9.1388. Picard, N., Strick, P.L., 2001. Imaging the premotor areas. Curr. Opin. Neurobiol. 11, 663–672. https://doi.org/10.1016/S0959-4388(01)00266-5. Plaha, P., Javed, S., Agombar, D., O’ Farrell, G., Khan, S., Whone, A., Gill, S., 2011. Bilateral caudal zona incerta nucleus stimulation for essential tremor: outcome and quality of life. J. Neurol. Neurosurg. Psychiatry 82, 899–904. https://doi.org/ 10.1136/jnnp.2010.222992. Plaha, P., Khan, S., Gill, S.S., 2008. Bilateral stimulation of the caudal zona incerta nucleus for tremor control. J. Neurol. Neurosurg. Psychiatry 79, 504–513. https:// doi.org/10.1136/jnnp.2006.112334. Popa, T., Russo, M., Vidailhet, M., Roze, E., Lehericy, S., Bonnet, C., Apartis, E., Legrand, A.P., Marais, L., Meunier, S., Gallea, C., 2013. Cerebellar rTMS stimulation may induce prolonged clinical benefits in essential tremor, and subjacent changes in functional connectivity: an open label trial. Brain Stimul 6, 175–179. https://doi.org/ 10.1016/j.brs.2012.04.009. Raethjen, J., Deuschl, G., 2012. The oscillating central network of Essential tremor. Clin. Neurophysiol. 123, 61–64. https://doi.org/10.1016/j.clinph.2011.09.024. Sandvik, U., Koskinen, L.-O., Lundquist, A., Blomstedt, P., 2012. Thalamic and subthalamic deep brain stimulation for essential tremor. Neurosurgery 70, 840–846. https://doi.org/10.1227/NEU.0b013e318236a809. Schnitzler, A., Gross, J., 2005. Normal and pathological oscillatory communication in the brain. Nat. Rev. Neurosci. 6, 285–296. https://doi.org/10.1038/nrn1650. Schnitzler, A., Münks, C., Butz, M., Timmermann, L., Gross, J., 2009. Synchronized brain network associated with essential tremor as revealed by magnetoencephalography. Mov. Disord. 24, 1629–1635. https://doi.org/10.1002/mds.22633. Sharifi, S., Nederveen, A.J., Booij, J., van Rootselaar, A.-F., 2014. Neuroimaging essentials in essential tremor: a systematic review. NeuroImage Clin 5, 217–231. https://doi.org/10.1016/j.nicl.2014.05.003. Stoodley, C.J., Valera, E.M., Schmahmann, J.D., 2012. Functional topography of the cerebellum for motor and cognitive tasks: an fMRI study. Neuroimage 59, 1560–1570. https://doi.org/10.1016/j.neuroimage.2011.08.065. Vallbo, a B., Wessberg, J., 1993. Organization of motor output in slow finger movements in man. J. Physiol. 469, 673–691. https://doi.org/10.1113/jphysiol.1993.sp019837. Wills, A.J., Jenkins, I.H., Thompson, P.D., Findley, L.J., Brooks, D.J., 1994. Red nuclear and cerebellar but no olivary activation associated with essential tremor: a positron emission tomoraphic study. Ann. Neurol. 36, 636–642. https://doi.org/10.1002/ ana.410360413. Zakaria, R., Lenz, F.A., Hua, S., Avin, B.H., Liu, C.C., Mari, Z., 2013. Thalamic physiology of intentional essential tremor is more like cerebellar tremor than postural essential tremor. Brain Res. 1529, 188–199. https://doi.org/10.1016/j.brainres.2013.07.011.

Gibson, W.S., Jo, H.J., Testini, P., Cho, S., Felmlee, J.P., Welker, K.M., Klassen, B.T., Min, H.-K., Lee, K.H., 2016. Functional correlates of the therapeutic and adverse effects evoked by thalamic stimulation for essential tremor. Brain 27, aww145. https://doi.org/10.1093/brain/aww145. Groppa, S., Herzog, J., Falk, D., Riedel, C., Deuschl, G., Volkmann, J., 2014. Physiological and anatomical decomposition of subthalamic neurostimulation effects in essential tremor. Brain 137, 109–121. https://doi.org/10.1093/brain/awt304. Gross, J., Timmermann, L., Kujala, J., Dirks, M., Schmitz, F., Salmelin, R., Schnitzler, A., 2002. The neural basis of intermittent motor control in humans. Proc. Natl. Acad. Sci. 99, 2299–2302. https://doi.org/10.1073/pnas.032682099. Hallett, M., Dubinsky, R.M., 1993. Glucose metabolism in the brain of patients with essential tremor. J. Neurol. Sci. 114, 45–48. Haslinger, B., Boecker, H., Büchel, C., Vesper, J., Tronnier, V., Pfister, R., Alesch, F., Moringlane, J., Krauss, J., Conrad, B., Schwaiger, M., Ceballos-Baumann, A., 2003. Differential modulation of subcortical target and cortex during deep brain stimulation. Neuroimage 18, 517–524. https://doi.org/10.1016/S1053-8119(02) 00043-5. Helmich, R.C., Toni, I., Deuschl, G., Bloem, B.R., 2013. The pathophysiology of essential tremor and Parkinson’s tremor. Curr. Neurol. Neurosci. Rep. 13, 378. https:// doi.org/10.1007/s11910-013-0378-8. Hopfner, F., Deuschl, G., 2018. Is essential tremor a single entity? Eur. J. Neurol. 25, 71–82. https://doi.org/10.1111/ene.13454. Jenkins, I.H., Bain, P.G., Colebatch, J.G., Thompson, P.D., Findley, L.J., Frackowiak, R.S.J., Marsden, C.D., Brooks, D.J., 1993. A positron emission tomography study of essential tremor: evidence for overactivity of cerebellar connections. Ann. Neurol. 34, 82–90. https://doi.org/10.1002/ana.410340115. Kahan, J., Mancini, L., Urner, M., Friston, K., Hariz, M., Holl, E., White, M., Ruge, D., Jahanshahi, M., Boertien, T., Yousry, T., Thornton, J.S., Limousin, P., Zrinzo, L., Foltynie, T., 2012. Therapeutic subthalamic nucleus deep brain stimulation reverses cortico-thalamic coupling during voluntary movements in Parkinson’s disease. PLoS One 7, e50270. https://doi.org/10.1371/journal.pone.0050270. Kahan, J., Urner, M., Moran, R., Flandin, G., Marreiros, A., Mancini, L., White, M., Thornton, J., Yousry, T., Zrinzo, L., Hariz, M., Limousin, P., Friston, K., Foltynie, T., 2014. Resting state functional MRI in Parkinson’s disease: the impact of deep brain stimulation on ‘effective’ connectivity. Brain 137, 1130–1144. https://doi.org/ 10.1093/brain/awu027. Kelly, R.M., Strick, P.L., 2003. Cerebellar loops with motor cortex and prefrontal cortex of a Nonhuman primate. J. Neurosci. 23, 8432–8444. https://doi.org/10.1523/ JNEUROSCI.23-23-08432.2003. Kriegeskorte, N., Simmons, W.K., Bellgowan, P.S., Baker, C.I., 2009. Circular analysis in systems neuroscience: the dangers of double dipping. Nat. Neurosci. 12, 535–540. https://doi.org/10.1038/nn.2303. Lenka, A., Bhalsing, K.S., Panda, R., Jhunjhunwala, K., Naduthota, R.M., Saini, J., Bharath, R.D., Yadav, R., Pal, P.K., 2017. Role of altered cerebello-thalamo-cortical network in the neurobiology of essential tremor. Neuroradiology 59, 157–168. https://doi.org/10.1007/s00234-016-1771-1. Louis, E.D., 2016. Linking essential tremor to the cerebellum: Neuropathological evidence. Cerebellum 15, 235–242. https://doi.org/10.1007/s12311-015-0692-6. Middlebrooks, E.H., Tuna, I.S., Almeida, L., Grewal, S.S., Wong, J., Heckman, M.G., Lesser, E.R., Bredel, M., Foote, K.D., Okun, M.S., Holanda, V.M., 2018. Structural connectivity–based segmentation of the thalamus and prediction of tremor improvement following thalamic deep brain stimulation of the ventral intermediate nucleus. NeuroImage Clin 20, 1266–1273. https://doi.org/10.1016/ j.nicl.2018.10.009. Mitrofanis, J., 2005. Some certainty for the “zone of uncertainty”? Exploring the function of the zona incerta. Neuroscience 130, 1–15. https://doi.org/10.1016/ j.neuroscience.2004.08.017.

10