The BRAINMAP-DBS Study: BRain Network AnalysIs usiNg 7-Tesla MRI and MAgnetoencephalograPhy for Deep Brain Stimulation

Deep brain stimulation for Parkinson's disease and essential tremor The effect of deep brain stimulation (DBS) relies on the modulation of dysfunctional motor brain networks.1,2 On average, 50% motor improvement is achieved, using standardized motor assessments. However, approximately 20% of patients shows insufficient benefit, with less than 30% improvement.3,4 To improve these outcomes, better electrode placement and selection of DBS electrical parameter programming is needed.5 This necessitates more advanced visualization of the intricate motor networks; both anatomical (7-Tesla MRI) and functional (MEG).11,12,13 Current DBS implantations are based on 1.5- or 3- Tesla MR scans. The resolution of these scans however, is not sufficient enough to visualize brain networks, preventing precise electrode placement directly at the motor parts within the small (size of a coffee bean) brain nucleus. In addition to the 7-Tesla MRI guided electrode placement, programming will be directed at influencing the cortical motor areas, by applying MEG, resulting in an overall decrease in dysfunctional network activity; instead of the "one-brain" model for everyone, DBS thereby becomes patient-specific, focused on one's own brain network.

Brain network analysis using 7-Tesla MRI and MEG The 7-Tesla MRI and MEG protocols already developed by our group and reported in (five) studies will be applied.6,7,8,9,10 In these studies we showed undistorted visualization of brain networks using 7-Tesla MRI and MEG in Parkinson disease patients that underwent DBS, and the potential to increase improvement in motor functioning by at least 10%. The 7-Tesla MRI network analysis, generated by combining T2 and diffusion weighted MRI, shows the (coloured) subdivisions of the brain nucleus and associated cortical projections. The MEG network analysis, generated from the MEG recordings, shows which cortical regions are affected (increase or decrease in activity, resulting in different colours) for each of the four individual electrode-contacts. Co-registration of the 7-Tesla MRI network, the MEG network and the CT scan will be performed for precise electrode localization. The result will show in which subdivision the DBS electrode-contact is located (placement), to which cortical area the subdivision projects and what activity patterns and changes therein are induced on the cortical areas (activation).16 Clinical test scores of the electrode-contact will then be correlated (regression analyses) with the network analysis. Since the entire network analysis will be visualized integrated on MRI, both individual analyses and group analyses (co-registration of networks of multiple patients) are possible, as well as application for the variable autoencoder. DBS location and programming based on 7-Tesla MRI and MEG network analysis will be used to optimize the outcome for beneficial clinical effects (suppression of tremor, bradykinesia, rigidity) with minimizing side effects (apathy, speech).15

For each patient, the total MEG recording will take 55 minutes. Nine short periods of unilateral stimulation will be performed with one of eight individual electrode-contacts, and one recording during DBS off in a randomized sequence. The effect of DBS on motor function (tremor, bradykinesia, rigidity, speech) is usually immediate and the assessment time is short (minutes). DBS settings can be temporarily adjusted during MEG recordings along with clinical assessment.

Upon completion of this process for all eight contacts of the electrode, the network analysis is performed and will show the DBS electrode-contact (left and right) which (solely) modulates the motor brain network. When the electrode-contact (DBS program) that the patient had before enrolment differs from the new 'brain network guided program' as determined by the network analysis, this new program is added to the DBS system. In the new program, margins for adjusting stimulation amplitude are set-up for later adjustments. The program that the patient had before enrolment in the study is kept as one of the programs on the device, which can be used as an 'escape option' by the patient if the newly programmed configurations have a worse clinical effect. After three months of network guided DBS standardized assessments of motor skills (UPDRS-III, TETRAS) will be performed as part of standard clinical care.

This project was set up in close collaboration with five DBS patients (patient-researchers). They identified regaining independence as the most important effect of DBS. They cited the possible adverse impact on cognition and mood (memory, speech, apathy) by DBS as their main concern. The discussed concerns were directly incorporated into the project.

Only few centers worldwide have a 7-Tesla MRI and MEG scan and therefore we will make this knowledge available through an open access database. We will apply the network analyses in the variable autoencoder (machine learning algorithm, a type of artificial intelligence) to generate prediction models with the help of the Department of Biomedical Engineering \& Physics (Amsterdam UMC).14 The application of machine learning to large numbers of high-resolution scans allows for the generation of network models (connectomes) to indicate the most optimal location and programming for the DBS electrode; directly applicable for all DBS groups.

Feasibility Every patient undergoing DBS at the Amsterdam UMC undergoes a preoperative 7-Tesla MRI as part of standard care. In the 450 7-Tesla MR-scans we have performed to date, we rarely encountered non-compatible implants or severe claustrophobia. Every patient undergoes screening with a MRI safety questionnaire and MRI metal detector (preventing taking ferromagnetic materials into the MRI). 7-Tesla MRI is a non-invasive technique which causes no pain and, importantly, the electromagnetic fields produce no known tissue damage of any kind. The MR system may make loud tapping, knocking, or other noises at times during the procedure. Earplugs are provided to prevent problems that may be associated with noise generated by the scanner. At all times, the patient will be (visually) monitored and will be able to communicate with the 7-Tesla MRI technologist using an intercom system. The patient may (request to) stop the acquisition at any time by using the push button (hold by the patient continuously).

MEG recordings are non-invasive, pain free and have negligible risks. Based on literature and our own experience in 40 patients, MEG recordings during DBS are a safe procedure. Participants will make one extra trip to the MEG center located at location VUmc. Changing DBS settings within the limits stated in the individual DBS passport of the participants is a risk-free procedure. As participants are in supine position on a comfortable bed during MEG recordings, we do not expect them to experience unacceptable motor symptoms.

For the first time 7-Tesla MRI and MEG network analysis will be combined in in a large-scale prospective fashion; introducing patient specific network guided DBS electrode placement and programming.