Computational Modeling of Deep Brain Stimulation of the Ventral Striatum

Widge, Alik; Dougherty, Darin

Abstract

This project will transfer a computational modeling technology, StimVision, from the McIntyre Lab at Case Western Reserve University to the deep brain stimulation (DBS) group at Massachusetts General Hospital (MGH). DBS is currently used to treat severe obsessive-compulsive disorder (OCD) through high-frequency electrical stimulation of the ventral internal capsule/ventral striatum (VC/VS). VC/VS is a complex brain region containing white matter tracts projecting to multiple frontal cortical regions. It is not understood which VC/VS fibers matter most to DBS' clinical effects on OCD, in part because we do not know how the stimulator's electrical field propagates through and activates this electrically irregular tissue. This leads to very inconsistent clinical effects (about 50% of patients respond poorly or not at all) for this invasive and expensive procedure. StimVision is designed to model those electric fields, but has not been adapted to or deployed in this specific application. As part of the BRAIN Initiative's efforts to transfer methods between laboratories, the McIntyre group will adapt StimVision for VC/VS modeling and train the MGH team in its use. We will then deploy it as part of our ongoing analysis of neurophysiologic and behavioral data in our existing patient cohort. We expect to show that the use of these advanced modeling techniques clarifies neural mechanisms underlying known behavioral effects, by compensating for heterogeneous VC/VS anatomy. Our Objective is to achieve that technology transfer and demonstrate MGH's successful uptake of the methods to the point that we can use them independently going forward. We approach this through two Aims.
Aim 1 represents the Case group's alterations to the user interface and algorithms of StimVision to be compatible with MGH's imaging and analysis workflows. This includes improvements to the electrical models that may be necessary for properly capturing VC/VS effects and incorporating novel tractography algorithms originally developed by MGH/Brigham collaborators.
Aim 2 then transfers initial and revised versions of StimVision to MGH and demonstrates their successful use for data analysis. We will re-analyze two MGH datasets, one concerning hypomania (a major DBS side effect) and one concerning behavioral effects from DBS manipulation during a psychophysical task. In both cases, we hypothesize that use of StimVision modeling will further clarify the cortical circuits underpinning these effects, and further hypothesize that this will primarily implicate the anterior cingulate. The goal of this Aim, however, is not necessarily to prove these hypotheses. It is demonstration of technology transfer, the core success criterion of this BRAIN RFA.

Public Health Relevance

This project will develop software that models how a novel treatment, deep brain stimulation (DBS) for obsessive-compulsive disorder (OCD), might affect specific brain structures. It will then transfer that software to a clinical DBS group that has multiple ongoing studies of DBS' mechanisms of action in OCD. By bringing this modeling capability into a clinical trials and human neuroscience group, we hope to advance the treatment of OCD through DBS and other circuit-oriented therapies.