Date of Award
Fall 2014
Document Type
Dissertation
Degree Name
Doctor of Philosophy (PhD)
Department
Biomedical Engineering
First Advisor
Butson, Christopher R.
Second Advisor
Schmit, Brian
Third Advisor
Prieto, Thomas
Abstract
Transcranial magnetic stimulation (TMS) is a neuromodulation technique used to treat a variety of neurological disorders. While many types of neuromodulation therapy are invasive, TMS is an attractive alternative because it is noninvasive and has a very strong safety record. However, clinical use of TMS has preceded a thorough scientific understanding: its mechanisms of action remain elusive, and the spatial extent of modulation is not well understood. We created a subject-specific, multiscale computational model to gain insights into the physiological response during motor cortex TMS. Specifically, we developed an approach that integrates three main components: 1) a high-resolution anatomical MR image of the whole head with diffusion weighted MRI data; 2) a subject-specific, electromagnetic, non-homogeneous, anisotropic, finite element model of the whole head with a novel time-dependent solver; 3) a population of multicompartmental pyramidal cell neuron models. We validated the model predictions by comparing them to motor evoked potentials (MEPs) immediately following single-pulse TMS of the human motor cortex. This modeling approach contains several novel components, which in turn allowed us to gain greater insights into the interactions of TMS with the brain. Using this approach we found that electric field magnitudes within gray matter and white matter vary substantially with coil orientation. Our results suggest that 1) without a time-dependent, subject-specific, non-homogeneous, anisotropic model, loci of stimulation cannot be accurately predicted; 2) loci of stimulation depend upon biophysical properties and morphologies of pyramidal cells in both gray and white matter relative to the induced electric field. These results indicate that the extent of neuromodulation is more widespread than originally thought. Through medical imaging and computational modeling, we provide insights into the effects of TMS at a multiscale level, which would be unachievable by either method alone. Finally, our approach is amenable to clinical implementation. As a result, it could provide the means by which TMS parameters can be prescribed for treatment and a foundation for improving coil design.