Research Associate Rush University Medical Center, Illinois
Rationale: A significant number of elements must be considered in the clinical response to stimulation therapy delivered directly to epileptogenic brain networks. These elements include stimulation parameter settings, the number and interdependence of anatomical targets, electrode number, geometry or shape of the electrode contacts, electrode location and biophysical properties of the stimulated medium, distribution of cathodic and anodic contacts, and trajectory of axonal bundles adjacent to the stimulation site. Stereotactic coordinates are important in planning depth lead placement for Deep Brain Stimulation (DBS: Medtronic) and the Responsive Neurostimulation System (RNS; NeuroPace). However, patient-specific lead placement modeling incorporating the above elements can provide crucial information regarding likelihood of targeting epileptogenic networks identified in the presurgical evaluation. Such a system becomes particularly important in candidates with previous resective surgeries and/or neuronal migration abnormalities. Our laboratory incorporated a compartmental cable modeling algorithm with a novel patient-specific lead implantation planning system to predict propagation pathways for DBS and RNS. This platform simulated, pre-implantation, how varying the above elements propagated neuromodulation therapy throughout ictogenic networks.
Methods: The platform was developed using Java and VTK interfacing with COMSOL (Burlington, MA) for optimizing computationally-intensive modeling. Simulating propagation pathways complemented the conventional pre-surgical planning workflow of three candidates implanted with depth leads in the anterior nucleus of the thalamus (ANT), and two candidates (one parahippocampal, one extra-temporal neocortical) implanted with RNS depth leads. 1) Virtual electrode lead placement used a 3D model of the patient-specific gapless MRI. 2) Leveraging finite element method analysis, the calculated electric potential and electric field around each active cylindrical depth contact while accounting for axonal bundle orientation generated a so-called activation function (AF). The AF was derived from the cable equation for axons. 3) Predicting distant cortical activation in each patient was performed by strategically placing the AF volume seeds associated with the activated depth contacts. This technique created a modulated circuit tractography (MCT) map. The pre-implant MCT map was used as a targeting template for positioning up to two depth leads in either the ANT for DBS or in white matter near the cortical grey-white matter junction for RNS. Results: Our platform generated: 1) planned positioning and activation of depth lead contacts, 2) computation and dynamic visualization of tract responses (depolarization and hyperpolarization) to either constant voltage or constant current-based pulse waveforms (Figure 1), 3) visualization of predicted tractography maps relative to the electrode contact positions, 4) multi-modal image registration and 3D-visualization of pathophysiological neural circuits, and 5) a DICOM-compliant interface to communicate with image archiving and intraoperative imaging technologies (O-Arm; Medtronic). Conclusions: Our depth electrode planning and navigation system provides an ability to predict the effect of stimulation parameters propagating modulation therapy through patient-specific axonal pathways involved in epileptogenic networks. Additionally, the platform can be used as a tool to simulate activation pathways, ‘on-demand’, during implantation of depth leads in the operating room. Funding: Please list any funding that was received in support of this abstract.: Foglia Family Foundation
