Investigation of intraoperative accelerometer data recording for safer and improved target selection for deep brain stimulation

Background: Deep Brain Stimulation (DBS) is a well established surgical treatment for Parkinson’s Disease (PD) and Essential Tremor (ET). Electrical leads are surgically implanted in the deeply seated structures in the brain and chronically stimulated. The location of the lead with respect to the anatomy is very important for optimal treatment. Therefore, clinicians carefully plan the surgery, record electrophysiological signals from the region of interest and perform stimulation tests to identify the best location to permanently place the leads. Nevertheless, there are certain aspects of the surgery that can still be improved. Firstly, therapeutic effects of stimulation are estimated by visually evaluating changes in tremor or passively moving patient's limb to evaluate changes in rigidity. These methods are subjective and depend heavily on the experience of the evaluator. Secondly, a significant amount of patient data is collected before and during the surgery like various CT and MR images, surgical planning information, electrophysiological recordings and results of stimulation tests. These are not fully utilized at the time of choosing the position for lead placement as they are either not available or acquired on separate systems or in the form of paper notes only. Thirdly, studies have shown that the current target structures to implant the leads (Subthalamic Nucleus (STN) for PD and Ventral Intermediate Nucleus (VIM) for ET) may not be the only ones responsible for the therapeutic effects. The objective of this doctoral work is to develop new methods that help clinicians subdue the above limitations which could in the long term improve the DBS therapy.
Method: After a thorough review of the existing literature, specifically customized solutions were designed for the shortcomings described above. A new method to quantitatively evaluate tremor during DBS surgery using acceleration sensor was developed. The method was then adapted to measure acceleration of passive movements and to evaluate changes in rigidity through it. Data from 30 DBS surgeries was collected by applying these methods in two clinical studies: one in Centre Hospitalier Universitaire, Clermont-Ferrand, France and another multi-center study in Universitäspital Basel and Inselspital Bern in Switzerland. To study the role of different anatomical structures in the therapeutic and adverse effects of stimulation, the data collected during the study was analysed using two methods. The first classical approach was to classify the data based on the anatomical structure in which the stimulating contact of the electrode was located. The second advanced approach was to use patient-specific Finite Element Method (FEM) simulations of the Electric Field (EF) to estimate the spatial distribution of stimulation in the structures surrounding the electrode. Such simulations of the adverse effect inducing stimulation current amplitudes are used to visualize the boundaries of safe stimulation and identify structures that could be responsible for these effects. In addition, the patient-specific simulations are also used to develop a new method called "Improvement Maps" to generate 2D and 3D visualization of intraoperative stimulation test results with the patient images and surgical planning. This visualization summarized the stimulation test results by dividing the explored area into multiple regions based on the improvement in symptoms as measured by the accelerometric methods.
Results: The accelerometric method successfully measured changes in tremor and rigidity. Standard deviation, signal energy and spectral amplitude of dominant frequency correlated with changes in the symptoms. Symptom suppressing stimulation current amplitudes identified through quantitative methods were lower than those identified through the subjective methods. Comparison of anatomical targets using the accelerometric data showed that to suppress rigidity in PD patients, stimulation current needed was marginally higher for Fields of Forel (FF) and Zona Incerta (ZI) compared to STN. On the other hand, the adverse effect occurrence rate was significantly lower in ZI and FF, indicating them to be better targets compared to STN. Similarly, for ET patients, other thalamic nuclei like the Intermediolateral (InL) and Ventro-Oral (VO) as well as the Pre-Lemniscal Radiations (PLR) are as efficient in suppressing tremor as the VIM but have lower occurrence of adverse effects. Volumetric analysis of spatial distribution of stimulation agreed with these results suggesting that the structures other than the VIM could also play a role in therapeutic effects of stimulation. The visualization of the adverse effect simulations clearly show the structures which could be responsible for such effects e.g. stimulation in the internal capsula induced pyramidal effects. These findings concur with the published literature. With regard to the improvement maps, the clinicians found them intuitive and easy to use to identify the optimal position for lead placement. If the maps were available during the surgery, the clinicians' choice of lead placement would have been different.
Conclusion: This doctoral work has shown that modern techniques like quantitative symptom evaluation and electric field simulations can suppress the existing drawbacks of the DBS surgery. Furthermore, these methods along with 3D visualization of data can simplify tasks for clinicians of optimizing lead placement. Better placement of the DBS lead can potentially reduce adverse effects and increase battery life of implanted pulse generator, resulting in better therapy for patients.