| Description |
The goal of this work is to construct a simulation toolset for studying and improving neuroprosthetic devices for restoring neural functionality to patients with neural disorders or diseases. This involves the construction and validation of coupled electromagnetic-neural computational models of retina and hippocampus, compiling knowledge from a broad multidisciplinary background into a single computational platform, with features specific to implant electronics, bulk tissue, cellular and neural network behavior, and diseased tissue. The application of a retina prosthetic device for restoring partial vision to patients blinded by degenerative diseases was first considered. This began with the conceptualization of the retina model, translating features of a connectome, implant electronics, and medical images into a computational model that was "degenerated." It was then applied to the design of novel electrode geometries towards increasing the resolution of induced visual percept, and of stimulation waveform shapes for increasing control of induced neural activity in diseased retina. Throughout this process, features of the simulation toolset itself were modified to increase the precision of the results, leading to a novel method for computing effective bulk resistivity for use in such multiscale modeling. This simulation strategy was then extended to the application of a hippocampus prosthetic device, which has been proposed to restore and/or enhance memory in patients with memory disorders such as Alzheimer's disease or dementia. Using this multiscale modeling approach, we are able to provide recommendations for electrode geometry, placement, and stimulation magnitude for increased safety and efficacy in future experimental trials. In attempt to model neural activity in dense hippocampal tissue, a simulation platform for considering the effects the electrical activity of neural networks have on the extracellular electric field, and therefore have on their neighboring cells, was constructed, further increasing the predictive ability of the proposed methodology for modeling electrical stimulation of neural tissue. |
