Bilateral Deep Brain Stimulation of the Nucleus Basalis of Meynert for Parkinson Disease Dementia: A Randomized Clinical Trial

Key Points

Question  Is deep brain stimulation of the nucleus basalis of Meynert safe in patients with Parkinson disease dementia, and which symptoms does it improve?

Findings  In this randomized, double-blind clinical trial of 6 patients, deep brain stimulation of the nucleus basalis of Meynert was found to be safe and well tolerated. Stimulation did not improve cognitive symptoms but may improve visual hallucinations.

Meaning  In patients with Parkinson disease dementia, deep brain stimulation of the nucleus basalis of Meynert can be safely performed, and its effects on troublesome neuropsychiatric symptoms require further exploration.

Importance  Deep brain stimulation of the nucleus basalis of Meynert (NBM DBS) has been proposed as a treatment option for Parkinson disease dementia.

Objective  To evaluate the safety and potential symptomatic effects of NBM DBS in patients with Parkinson disease dementia.

Design, Setting, and Participants  A randomized, double-blind, crossover clinical trial evaluated the results of 6 patients with Parkinson disease dementia who were treated with NBM DBS at a neurosurgical referral center in the United Kingdom from October 26, 2012, to July 31, 2015. Eligible patients met the diagnostic criteria for Parkinson disease dementia, had motor fluctuations, were appropriate surgical candidates aside from the coexistence of dementia, were age 35 to 80 years, were able to give informed consent, had a Mini-Mental State Examination score of 21 to 26, had minimal atrophy seen on results of brain magnetic resonance imaging, and lived at home with a caregiver-informant.

Interventions  After surgery, patients were assigned to receive either active stimulation (bilateral, low-frequency [20 Hz] NBM DBS) or sham stimulation for 6 weeks, followed by the opposite condition for 6 weeks.

Main Outcomes and Measures  The primary outcome was the difference in scores on each item of an abbreviated cognitive battery (California Verbal Learning Test-II, Wechsler Adult Intelligence Scale-III digit span, verbal fluency, Posner covert attention test, and simple and choice reaction times) between the 2 conditions. Secondary outcomes were exploratory and included differences in scores on standardized measurements of cognitive, psychiatric, and motor symptoms and resting state functional magnetic resonance imaging.

Results  Surgery and stimulation were well tolerated by all 6 patients (all men; mean [SD] age, 65.2 [10.7] years), with no serious adverse events during the trial. No consistent improvements were observed in the primary cognitive outcomes or in results of resting state functional magnetic resonance imaging. An improvement in scores on the Neuropsychiatric Inventory was observed with NBM DBS (8.5 points [range, 4-26 points]) compared with sham stimulation (12 points [range, 8-38 points]; median difference, 5 points; 95% CI, 2.5-8.5 points; P = .03) and the preoperative baseline (13 points [range, 5-25 points]; median difference, 2 points; 95% CI, −8 to 5.5 points; P = .69).

Conclusions and Relevance  Low-frequency NBM DBS was safely conducted in patients with Parkinson disease dementia; however, no improvements were observed in the primary cognitive outcomes. Further studies may be warranted to explore its potential to improve troublesome neuropsychiatric symptoms.

Trial Registration  clinicaltrials.gov Identifier: NCT01701544

Parkinson disease dementia (PDD) is characterized by impairment of executive and attentional functions, accompanied by episodic memory and visuoperceptual deficits, cognitive fluctuation, and neuropsychiatric disturbances, particularly visual hallucinations.^1-3 Parkinson disease dementia affects 75% to 90% of patients who have had Parkinson disease for more than a decade,^4,5 significantly reducing quality of life for both patients and caregivers.^6,7 With increasing life expectancy, PDD is set to become more prevalent.⁸ Current therapy with acetylcholinesterase inhibitors and N-methyl d-aspartate–receptor antagonists produces moderate improvement of symptoms at best.^9,10 Therefore, novel treatment options are needed.

It is recognized that the pathologic processes underlying PDD are not restricted to cortical Lewy bodies but are heterogeneous, with a variety of cellular neuropathologic features, genetic influences, and neurotransmitter deficits all contributing.¹ This fact may explain why single ligand–targeted drug therapy has so far proved to be limited. An alternative approach may be to directly modulate neural networks to relieve symptoms. Deep brain stimulation (DBS) has established safety and efficacy in ameliorating movement symptoms in Parkinson disease by modulating aberrant processing patterns in motor circuits.^11-13 Consequently, the therapeutic use of DBS is being investigated in dementia¹⁴ (or reinvestigated¹⁵).

The human nucleus basalis of Meynert (NBM) is a discrete anatomical structure in the basal forebrain, located inferior to the posterior globus pallidus internus (GPi),¹⁶ and provides the major source of cholinergic innervation to the cortical mantle.¹⁷ Several lines of evidence support exploration of DBS to this structure as a potential therapy for PDD. First, NBM activity is strongly implicated in cognitive and behavioral functions, including arousal, attention, perception, and memory,^16,18,19 all of which are particularly impaired in patients with PDD.¹ Second, the NBM degenerates significantly in patients with PDD with up to 70% cell loss, which correlates closely with cortical cholinergic deficits and worsening cognitive impairment.^1,20-22 In comparable patients with Alzheimer disease, NBM cell loss and cholinergic deficits are less marked,¹⁶ suggesting that loss of NBM function may be a more important pathologic substrate in PDD and therefore a key circuit to modulate. In particular, low-frequency stimulation of the NBM may represent a way of enhancing cholinergic innervation of the cortex of patients with PDD, as has been shown in animal models.^18,23 In this randomized, double-blind clinical trial, we aimed to evaluate the safety and potential symptomatic effects of NBM DBS in patients with PDD.

This pilot trial used a double-blind, randomized crossover design to compare scores on a battery of standardized cognitive, psychiatric, and motor tests preoperatively and after 6 weeks of active bilateral NBM DBS and 6 weeks of sham DBS (Figure 1, Figure 2, and trial protocol in Supplement 1). Patients were treated from October 26, 2012, to July 31, 2015. Safety of the surgical procedure and tolerability of stimulation were evaluated throughout. Deep brain stimulation of the NBM was achieved using electrodes that straddled the GPi, providing the potential for subsequent conventional GPi DBS for coexisting motor impairments. The study was sponsored by University College London and performed at the National Hospital for Neurology and Neurosurgery, London, England. The trial conformed to the Seoul revision of the Declaration of Helsinki (2008)²⁴ and Good Clinical Practice guidelines and was approved by the East of England Research Ethics Committee. All enrolled participants were assessed for their capacity to provide informed consent by an independent neuropsychologist prior to providing written informed consent.

Patients with PDD who met the following eligibility criteria were recruited from a neurologic referral center in London, England: fulfilled criteria for the diagnosis of Parkinson disease²⁵ and for PDD,²⁶ had motor fluctuations likely to improve with GPi DBS²⁷ (thus, if no benefit was achieved with NBM DBS at trial end, they could switch to GPi DBS), were appropriate candidates for DBS surgery aside from the coexistence of dementia, were aged 35 to 80 years, were able to give written informed consent, had a Mini-Mental State Examination score of 21 to 26, had results of brain magnetic resonance imaging (MRI) showing minimal atrophy and no abnormality that would compromise compliance with the protocol, lived at home with a caregiver-informant, were willing to comply with the trial protocol, had no suspicion of other cause for parkinsonism or dementia, and had no prior intracerebral surgical intervention for Parkinson disease. In addition, all patients were already receiving a stable dose of acetylcholinesterase inhibitors at the time of recruitment, which was continued throughout the trial.

We randomly assigned patients to either the stimulation off-first group (sham stimulation for 6 weeks, followed by active stimulation for 6 weeks) or the stimulation on-first group (active stimulation for 6 weeks, followed by sham stimulation for 6 weeks) (Figure 2). We used computer-generated pairwise randomization according to order of enrollment so that equal numbers were recruited to each group in a counterbalanced order. Patients and assessing clinicians were blinded to stimulation condition. An unblinded clinician (T.F.) held the randomization list and spent the same time adjusting each patient’s stimulator at the start of both active and sham stimulation periods. Electrical parameters were selected to avoid any immediate or long-lasting adverse effect that could be perceived by the patient or blinded clinicians.

All enrolled patients completed a baseline assessment and then underwent stereotactic implantation of bilateral DBS electrodes (model 3387 [patient A] or 3389 [patients B-F]; Medtronic). A more posterior entry point than that used for conventional subthalamic nucleus DBS was chosen (behind the coronal suture) to ensure a trajectory that was as vertical as possible, thereby enabling placement of the uppermost contacts of the lead in the posterior third of the ventral pallidum and the deepest contact(s) in the Ch4i subsector of the NBM. The average trajectory in the sagittal plane was 0° to 10° anterior to a line perpendicular to the anterior commissure–posterior commissure line and in the coronal plane was 0° to 5° lateral to the same. The Ch4i subsector was chosen as it is both the largest subsector of the NBM, giving the highest probability of successful electrode placement, and has the most widespread cortical projections, thus potentially influencing more cortical areas.¹⁶ Electrode implantation was guided by targeting the NBM on individual stereotactic axial and coronal proton density MRI scans on which the pallidum, optic tract, anterior commissure, and adjacent NBM were visible (1.5-T Siemens Espree, PDw Turbo Spin-echo; 1.0 × 1.0 × 2.0 mm; repetition time, 4000 milliseconds; echo time, 13 milliseconds). Operations for all patients were performed under general anesthesia using a Leksell stereotactic frame without microelectrode recording. The accuracy of DBS lead contact location was confirmed immediately with postoperative stereotactic MRI (Figure 3). Full details of our neurosurgical procedure have been published previously.^28,29 One week later, each patient received an Activa PC implantable pulse generator (Medtronic) in a second procedure under general anesthesia.

Patients completed their first postoperative assessment 1 week after implantation of the pulse generator (Figure 2). Three weeks later patients attended for 24 hours (local hotel accommodation rather than hospital admission) and were screened for the effects of stimulation in an open-label manner, using the Wechsler Adult Intelligence Scale-III digit span as an objective measure. Only low-frequency stimulation at 20 Hz was used, as this is the physiological discharge rate of NBM neurons in animals and stimulates acetylcholine release from their cortical terminals.^23,30 Only monopolar stimulation was used, with a pulse width of 60 μs. Optimum stimulation voltages were those producing the highest digit span scores with the lowest energy without adverse effects; these voltages were adopted for the blinded phase. Deep brain stimulation was subsequently turned off for 2 weeks, and then patients were randomized into the stimulation off-first or on-first group for the subsequent 6 weeks. After the first 6-week period, there was a 2-week washout period (DBS off); then patients were switched to the opposite condition for a further 6 weeks. All assessments performed at preoperative baseline were repeated at the end of each 6-week period except for measures of IQ (Figure 2).

At each visit, we prioritized an abbreviated cognitive battery of primary outcome measures to minimize the effect of patient fatigue, comprising the California Verbal Learning Test-II, Wechsler Adult Intelligence Scale-III digit span, verbal fluency, Posner covert attention test, and simple and choice reaction times.

At baseline and at the end of each blinded stimulation condition, additional exploratory outcomes included the following validated assessments: Mini-Mental State Examination, Mattis Dementia Rating Scale-2, short recognition memory for faces, Wechsler Adult Intelligence Scale-III arithmetic and letter-number sequencing, trail making test, Wechsler Adult Intelligence Scale-III symbol search and digit-symbol coding, Florida Apraxia Screening Test, Movement Disorder Society Unified Parkinson’s Disease Rating Scale, Freezing of Gait Questionnaire, Parkinson Disease Questionnaire, Scales for Outcomes in Parkinson Disease–Sleep, Scales for Outcomes in Parkinson Disease–Autonomic symptoms, nonmotor symptoms questionnaire, Starkstein Apathy Scale, EuroQol visual analogue scale, Neuropsychiatric Inventory, Blessed dementia scale, Hamilton Depression Scale, and Hamilton Anxiety Scale. All adverse events were systematically recorded.

All patients underwent functional MRI (fMRI) scanning at the end of each blinded 6-week condition to investigate the effects of NBM DBS on functional connectivity in the default mode network (DMN) (eAppendix in Supplement 2).

We planned to recruit 6 patients based on the fact that the safety of NBM DBS implantation in this vulnerable patient group is unknown, necessitating a cautious approach to recruitment, and based on an estimate of the number of eligible patients in our clinic. This study was therefore a pilot trial with a small sample size based on these practical considerations. While we chose to prioritize a specific battery of cognitive measures, all outcome measures are principally exploratory in nature, and the trial was not powered to judge efficacy. Statistical comparisons were performed solely as a means of screening which measures might be prioritized in future trials, and are not corrected for multiple comparisons. Two-tailed Wilcoxon signed rank tests were used to compare scores between the on-stimulation and off-stimulation periods when the distribution of differences was symmetrical, and 2-tailed sign tests otherwise. To inform on interindividual variability in outcomes, results data for all patients are presented in the eAppendix in Supplement 2. Data were analyzed using SPSS, version 24.0 (SPSS Inc). Differences in outcomes of P < .05 were considered priorities for future study.

Between October 26, 2012, and July 31, 2015, we assessed 25 patients with PDD and enrolled 8 into the study. Two patients met eligibility criteria but were subsequently excluded prior to surgery: 1 opted for GPi DBS for motor symptoms only, and the other was determined to be unsafe for anesthesia owing to postural hypotension. Six patients (all men; mean [SD] age, 65.2 [10.7] years) proceeded to NBM DBS implantation (Figure 1). Table 1³¹ summarizes their characteristics, stereotactic coordinates of their active NBM contacts, and individual stimulation parameters during the blinded period. The most ventral active contact was successfully placed in the Ch4i subsector of the NBM in each patient (eFigure 1 in Supplement 2). Surgery was well tolerated, and all patients were ambulatory within 24 hours and fully oriented within 48 hours. All 6 patients completed the blinded crossover phase and were included in the analysis (Figure 1).

No serious adverse events occurred during the trial period. In patient C, erosion of the right electrode cap through the scalp occurred 15 months after the start date of the trial (during open-label follow-up), necessitating surgical removal, which did not result in harmful sequelae. eTable 1 in Supplement 2 lists all adverse events. The surgical procedure itself did not have a negative effect on cognition (eTable 2 in Supplement 2).

There was no significant improvement in any of the scales in the abbreviated cognitive battery (Table 2). Individual results for all patients are presented in eTables 3 to 8 in Supplement 2.

The most encouraging finding was an improvement at group level in Neuropsychiatric Inventory total scores with NBM DBS (Table 2); at baseline, patients had a median Neuropsychiatric Inventory total score of 13 points (range, 5-25 points), while scores were 12 points (range, 8-38 points) at the end of the sham stimulation period and 8.5 points (range, 4-26 points) at the end of the active stimulation period. Pairwise comparisons between baseline and active stimulation showed a median improvement of 2 points (95% CI, −8 to 5.5 points) that was not statistically significant (P = .69), whereas comparison between sham and active stimulation showed a significant median improvement of 5 points (95% CI, 2.5 to 8.5, P = .03 points). There were no significant between-group differences in Neuropsychiatric Inventory total scores according to randomization sequence. This improvement was driven primarily by a reduction in hallucinations subscale scores in patients A and D (eTable 5 in Supplement 2): at baseline, patients A, C, and D reported daily complex visual hallucinations, while patient B reported regular extracampine hallucinations. Patients A and D both experienced near-complete cessation of visual hallucinations after surgery when NBM stimulation was turned on, followed by a resurgence of hallucinations when stimulation was subsequently turned off (eFigure 2 in Supplement 2).

Patients A, D, and E reported subjective improvement in their health-related quality of life (EuroQol visual analogue scale; eTable 6 in Supplement 2) during on-stimulation compared with both off-stimulation and baseline. Patients B and F reported no change and patient C reported a decline. This pattern was mirrored by the subjective reporting of distress levels by the patients’ caregivers on the Neuropsychiatric Inventory (eTable 5 in Supplement 2).

Patients A, C, and D showed improvement in Movement Disorder Society Unified Parkinson’s Disease Rating Scale part IV scores during on-stimulation compared with both off-stimulation and baseline (eTable 6 in Supplement 2). These patients experienced levodopa-induced dyskinesias, and their improvements in Movement Disorder Society Unified Parkinson’s Disease Rating Scale part IV scores were owing to a reduction in the frequency and severity of these dyskinesias during on-stimulation.

Activity throughout the DMN was readily identifiable in all patients with both active and sham NBM DBS. However, there was no significant difference in DMN activity between stimulation conditions at the group level (eFigure 3 in Supplement 2).

In this double-blind, randomized trial of 6 patients with established PDD, low-frequency (20 Hz) stimulation of the NBM was safe and well tolerated. There was no significant change in cognitive outcomes across the group with stimulation, but there was a suggestion of improvement in neuropsychiatric symptoms. There were no functional connectivity changes in the DMN.

The fact that both surgery and stimulation were well tolerated in this vulnerable patient group is important. The presence of dementia is considered a contraindication to DBS therapy,³² yet our study suggests that well-selected patients with cognitive and psychiatric symptoms can consent to and tolerate such treatment without serious adverse events or cognitive deterioration. This finding is in line with other studies in Alzheimer disease^14,33 and should support further exploratory studies of NBM DBS in patients with PDD.

The change in neuropsychiatric symptoms with NBM DBS was driven by a reduction in complex visual hallucinations, most dramatically in patient D. These observations could not be explained by medication changes, which remained constant throughout the trial. The generation of visual hallucinations in PDD is poorly understood but is linked to cholinergic deficiency,¹ and strong associations have been shown between the integrity of NBM cholinergic projections and development of hallucinations.³⁴ Acetylcholinesterase inhibitors significantly improve visual hallucinations in patients with PDD^35,36; thus, the possibility that hallucinations may also be reduced by NBM DBS, even in individual patients, provides support to the hypothesis that stimulation might modulate cholinergic transmission.¹⁶ Visual hallucinations have a particularly negative effect on quality of life for both patients with PDD and caregivers^6,37,38; therefore, the reduction in these symptoms may underlie the subjective improvements in quality of life reported by patients A and D (and their caregivers) during active stimulation compared with both sham stimulation and baseline (eTables 5 and 6 in Supplement 2).

Another core feature of PDD is fluctuations in alertness and cognition.^1-3 We attempted to measure this factor in the blinded, caregiver-rated Daytime Sleepiness score of the Scales for Outcomes in Parkinson Disease–Sleep scale (eTable 7 in Supplement 2): a group-level improvement in median Daytime Sleepiness scores was observed with NBM DBS (10 points [range, 4-16 points]) compared with sham stimulation (13.5 points [range, 7-15 points]) and baseline (10.5 points [range, 4-13 points]), but this improvement was not significant (median difference between NBM DBS and sham stimulation, –1.0 points; 95% CI, −8.0 to 3.0 points; P = .42; and median difference between NBM DBS and baseline, 1.0 points; 95% CI, −5.0 to 6.0 points; P = .75). We did not observe a clinical change in vigilance with NBM DBS.

We did not observe the dramatic cognitive improvements reported in the single published case of low-frequency NBM DBS in PDD.³⁹ Our patients’ performance on many cognitive tests remained unchanged, and in some instances worsened, with stimulation. This finding may have occurred because the previous case was largely conducted as open-label or because simultaneous subthalamic nucleus and NBM stimulation was used, which may have synergistic effects.¹

We found no significant changes in resting state functional connectivity of the DMN between active and sham NBM DBS. We restricted our analysis to the DMN, as connectivity changes in this network are associated with cognitive decline in advanced PD^40-42 and to avoid embarking on multiple comparisons of our imaging data. Therefore, we cannot exclude the possibility that significant changes occurred in other brain networks owing to NBM DBS.

An unexpected finding was the improvement in Movement Disorder Society Unified Parkinson’s Disease Rating Scale Part IV scores during on-stimulation compared with both off-stimulation and baseline in patients A, C, and D, driven by reduction in their dyskinesias. These results are likely explicable by the current spread from NBM to the overlying GPi, in addition to any possible microlesion effect of the surgery. However, conventional GPi DBS for dyskinesia control in Parkinson disease is delivered at high frequency (130 Hz), so the finding that low-frequency (20 Hz) stimulation directed toward the NBM also attenuated dyskinesias warrants further study.

A direct comparison between our study and the 2 previous double-blind trials of DBS in dementia is difficult since both trials were conducted in patients with Alzheimer disease, and only 1 of them targeted the NBM.^14,33 Changes in visual hallucinations were either not seen³³ or not assessed,¹⁴ although such symptoms are less common in Alzheimer disease.⁴³

The major limitation to our study was the small sample size, which, although practical in context and appropriate for exploratory data collection, was not powered to judge efficacy. Our results must therefore be viewed as hypothesis generating to assist the design and planning of future trials. Another relevant issue is that all patients continued acetylcholinesterase inhibitor therapy during the trial; although there were no dose alterations, this fact still means that potential physiological effects of NBM DBS on the cholinergic system, and consequent clinical effects, may have been partially masked. However, given the relative safety of acetylcholinesterase inhibitors in comparison with DBS, we did not think it was appropriate to expose patients to surgical risks who might gain sufficient cognitive benefits from the use of medications alone. In addition, we did not include a randomized control group of patients with PDD who did not undergo surgery, and so we cannot objectively determine whether NBM DBS made a difference to the natural progression of cognitive deficits during the trial period. We only investigated the effects of low-frequency (20 Hz) NBM stimulation; however, the scientific rationale for this frequency is limited,^16,39 and stimulation at a different frequency might produce different results. Finally, had we not decided to place contacts in the posteroventral GPi to allow fallback treatment of motor symptoms, we may have been able to locate the DBS lead in a more optimal region of the NBM, such as closer to its outflow path into the external capsule.

One aim of this pilot study was to inform sample size calculations for future trials. In accordance with our results, detailed measurements of the frequency and severity of neuropsychiatric symptoms, particularly visual hallucinations, should comprise future primary outcome measures. However, whether potential beneficial effects restricted to these symptoms ultimately justify a surgical intervention in patients with dementia will require individual patient-centered discussions.

Our results suggest that NBM DBS is both safe and well tolerated in carefully selected patients with PDD. Although low-frequency (20 Hz) stimulation in the Ch4i subregion of the NBM did not consistently improve the range of cognitive deficits seen in our patients, it may improve refractory neuropsychiatric symptoms, particularly visual hallucinations. Given that we have demonstrated the safety of this procedure, we propose that longer-term follow-up trials can be justified to explore symptomatic effects and allow comparisons against patients who did not undergo surgery to judge any effect of NBM DBS on the natural history of PDD. Such studies should also explore the effects of stimulation among NBM subregions and formally test whether stimulation at other frequencies can reproducibly effect cognitive and neuropsychiatric symptoms in Lewy body–related dementias.

Corresponding Author: Thomas Foltynie, MD, PhD, Unit of Functional Neurosurgery (Box 146), The National Hospital for Neurology and Neurosurgery, Queen Square, London WC1N 3BG, England (t.foltynie@ucl.ac.uk).

Accepted for Publication: October 5, 2017.

Published Online: December 18, 2017. doi:10.1001/jamaneurol.2017.3762

Author Contributions: Drs Gratwicke and Foltynie had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis. Drs Jahanshahi and Foltynie share senior authorship.

Study concept and design: Gratwicke, Zrinzo, Akram, Yousry, Limousin, Jahanshahi, Foltynie.

Acquisition, analysis, or interpretation of data: Gratwicke, Zrinzo, Kahan, Peters, Beigi, Akram, Hyam, Oswal, Day, Mancini, Thornton, Limousin, Hariz, Jahanshahi, Foltynie.

Drafting of the manuscript: Gratwicke, Kahan, Hyam, Limousin, Jahanshahi, Foltynie.

Critical revision of the manuscript for important intellectual content: Gratwicke, Zrinzo, Kahan, Peters, Beigi, Akram, Hyam, Oswal, Day, Mancini, Thornton, Yousry, Limousin, Hariz, Jahanshahi.

Statistical analysis: Gratwicke, Kahan, Day.

Obtained funding: Gratwicke, Jahanshahi, Foltynie.

Administrative, technical, or material support: Gratwicke, Zrinzo, Kahan, Peters, Beigi, Akram, Hyam, Oswal, Mancini, Thornton, Yousry, Jahanshahi.

Study supervision: Zrinzo, Akram, Day, Limousin, Jahanshahi, Foltynie.

Conflict of Interest Disclosures: Dr Gratwicke reported receiving honoraria from Medtronic, UCB Pharmaceuticals, Britannia Pharmaceuticals, and Bial. Dr Zrinzo reported receiving honoraria from Medtronic and Boston Scientific. Dr Hyam reported receiving honoraria from Medtronic and St Jude Medical. Dr Day reported receiving personal fees from Takeda Development Centre Europe Ltd. Dr Limousin reported receiving honoraria from Medtronic, St Jude Medical, and Boston Scientific. Dr Hariz reported receiving honoraria from Medtronic and Boston Scientific. Dr Jahanshahi reported receiving honoraria and travelling expenses from Medtronic. Dr Foltynie reported receiving honoraria from Medtronic, St Jude Medical, Profile Pharma, Bial, Abbvie Pharmaceuticals, UCB Pharmaceuticals, and Oxford Biomedica. No other disclosures were reported.

Funding/Support: This study was funded by grant PAR11141 from the Brain Research Trust and was sponsored by University College London. The work was undertaken at University College London/University College London Hospital and was funded in part by the Department of Health National Institute for Health Research Biomedical Research Centres funding scheme. The Unit of Functional Neurosurgery, University College London Institute of Neurology is supported by the Parkinson’s Appeal and the Monument Trust.

Role of the Funder/Sponsor: The funding sources had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.

Additional Contributions: We thank all the patients who participated in this trial and their partners, as well as their referring clinicians.