Abstracts

Rationale:
Electrical stimulation through implanted intracranial electrodes is a critical part of evaluation for epilepsy surgery in many patients. The interaction between electrical stimulation and activation of neural tissue is complex and impacted by many factors including properties of the electrodes and stimulus applied, properties of the tissue, and their interaction. We have previously used a computer model to study the relationship between depth electrodes, like those used in stereo-electroencephalography (SEEG), and the extent of neural activation and how this relationship is impacted by electrode design, orientation, and location. The passive electrical properties (conductivity) of tissue  are not invariable and likely have a range of values in both normal and pathological (e.g. tumor, dysplasia, sclerosis) circumstances. Here, we use a similar model to investigate the impact of varying passive tissue properties (electrical conductivity) on the volume of activated tissue.
Method:
A steady state finite element model of a slab of brain tissue (5x5x5 cm3), including areas with conductivities to mimic gray matter, white matter, and cerebrospinal fluid, with a depth electrode (0.8 mm diameter, 2 mm cylindrical contacts with center-to-center spacing of 3.5 mm) placed in the tissue was implemented in COMSOL Multiphysics Software (COMSOL, Inc.; Burlington, MA). A cylinder of tissue (2.5 cm diameter) centered within this slab was allowed to have electrical properties different from the rest of the model. Constant current, bipolar or monopolar stimulation was simulated and the resulting current density and electric potentials were calculated. The electric potential distribution was then imported into MATLAB R2020A (MathWorks, Natick, MA). The matrix of second spatial derivatives of the electric potentials (Hessian matrix) was calculated, and the maximum eigenvalue of the Hessian matrix at each point in space was used to estimate the volume of activated tissue, the region within which nerve cells are likely to be activated by the stimulus.
Results:
Several scenarios were modeled using the above procedure, including: (1) all conductivities of the cylinder 0.25-4X the surrounding tissue, (2) conductivity of just the gray matter or just the white matter in the cylinder 0.25-4X the surrounding tissue, and (3) radial conductivity of the white matter in the cylinder 0.25-4X the surrounding tissue. Compared to the control condition with the properties of the cylinder the same as the surrounding tissue), the volume of activated tissue was higher when the conductivity of the tissue was lower with the relative impact depending on the portion of tissue whose property was changed (all tissue > white matter only > gray matter only). The impact was also significantly higher for monopolar than bipolar stimulation.
Conclusion:
The passive electrical properties of neural tissue are likely to have an impact on the volume of tissue activated by a given electrical stimulus. In addition to the variations in excitability of epileptogenic vs. normal tissue, variations in passive electrical properties likely partially account for differences in electrical stimulation thresholds and responses.
Funding:
:This work was partially supported by a gift from Ms. Barbara Rosenblatt.