Impact Factor 1.821

The Frontiers in Neuroscience journal series is the 1st most cited in Neurosciences

Methods ARTICLE Provisionally accepted The full-text will be published soon. Notify me

Front. Comput. Neurosci. | doi: 10.3389/fncom.2018.00040

Modeling Current Sources for Neural Stimulation in COMSOL

  • 1Biomedical Engineering, Duke University, United States
  • 2Electrical and Computer Engineering, Duke University, United States
  • 3Neurobiology, Duke University, United States
  • 4Neurosurgery, School of Medicine, Duke University, United States

Background: Computational modeling provides an important toolset for designing and analyzing neural stimulation devices to treat neurological disorders and diseases. Modeling enables efficient exploration of large parameter spaces, where preclinical and clinical studies would be infeasible. Current commercial finite element method software packages enable straightforward calculation of the potential distributions, but it is not always clear how to implement boundary conditions to appropriately represent metal stimulating electrodes. By quantifying the effects of different electrode representations on activation thresholds for model axons, we provide recommendations for accurate and efficient modeling of neural stimulating electrodes. Methods: We quantified the effects of different representations of current sources for neural stimulation in COMSOL Multiphysics for monopolar, bipolar, and multipolar electrode designs. By examining the physics of the different approaches and the resulting activation thresholds for model axons, we provide recommendations for accurate and efficient modeling of neural stimulating electrodes. Results: We recommend modeling each electrode contact as a thin platinum domain, modeling the electrode substrate with the conductivity of silicone, and either using a point current source in the center of each electrode contact or using a boundary current source. Alternatively, to avoid possible numerical instabilities associated with a large range of conductivity values (i.e., platinum and silicone) and to eliminate the small mesh elements required for thin electrode contacts, the electrode substrate can be assigned the conductivity of platinum by using insulating boundaries between the substrate and surrounding medium, and within the substrate to isolate the contacts from each other. When modeling more than one contact, we recommend using superposition by solving the model once for each contact, leaving inactive contacts floating, and superposing the resulting potentials. We computed comparable errors in activation thresholds across the different implementations in the simplified model (electrode in a homogeneous, isotropic medium), and in a realistic models of rat spinal cord stimulation and human deep brain stimulation, indicating that the recommended approaches are applicable to different stimulation targets.

Keywords: Neural Engineering, Neuromodulation, Finite element method, Boundary conditions, computational modeling

Received: 23 Jan 2018; Accepted: 17 May 2018.

Edited by:

Anthony N. Burkitt, University of Melbourne, Australia

Reviewed by:

Jiang Wang, Tianjin University, China
Ahmed El Hady, Princeton University, United States  

Copyright: © 2018 Pelot, Thio and Grill. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

* Correspondence: Dr. Warren M. Grill, Duke University, Biomedical Engineering, Room 1427, Fitzpatrick CIEMAS, 101 Science Drive, Campus Box 90281, Durham, 27708-0281, NC, United States, warren.grill@duke.edu