Authors:

Rodrigo Romero

Date Created:

Not specified

Course Title:

Professor:

Not specified

About Paper:

Transcranial Magnetic Stimulation (TMS) is a non-invasive brain stimulation approach using coils driven by low-frequency currents to stimulate a targeted region. Common applications are brain function studies and clinical purposes. Electric field (E-field) generated during TMS has been shown to be the effective ingredient, and computational E-field dosimetry has been accepted to improve TMS cortical mapping and TMS targeting. Finite element method (FEM) is one of the popular numerical techniques applied to evaluate the E-field during TMS, where a scalar potential determining the E-field in the brain is obtained by solving Poisson's equation derived from the current continuity condition. However, due to a large number of unknowns, the induced linear equation in FEM is usually solved with iterative solvers, where the convergence tolerance needs to be judiciously chosen to balance the accuracy and the efficiency of the solver. In this study, different convergent tolerances are first tested on the linear equations derived from the 1st-order FEM applying to different head models (i.e., spherical head model, lower-resolution MRI-derived head model, and high-resolution MRI-derived head model). Then they are tested on the linear equations derived from the FEM of different orders (1st and 2nd order). The results show that an optimal trade-off tolerance can be found in different cases, and this tolerance could improve the computational speed by at least 1/3 without losing the necessary accuracy.

Source:

Purdue University / 2023

Topics:

No topics listed

Co-authors:

Rodrigo Romero