Quantifying Effects of Model Edits on Neurostimulation Simulations

Presenter: Siddharth Simon

Research Category: Engineering and Technology
College: College of Engineering
Major(s): Computer Engineering, Computer Science
Graduation Date: 2023

This project was centered around studying the optimal amount of precision needed when creating a finite element model of the head, in order to minimize the amount of time spent manually editing the model, while also ensuring that any simulations of the electric field remain relatively unchanged. This is the first time that any quantification of this sort has been attempted in the context of optimizing transcranial direct current stimulation, and it has the potential to greatly benefit the Brain Stimulation and Simulation Lab at Northeastern, along with labs elsewhere, by reducing the time spent on creating finite element models of the head. This endeavor had its original version completed in Matlab to determine the difference in the electric fields between the simulations of different versions of a head model, and functionality to quantify the overall difference by tissue type between different versions of the same model; the latter can be defined by comparing the differences in voxels, which are essentially 3-dimensional pixels. However, this tool is being further refined to allow for different regions of interest to be targeted. Furthermore, more head models will have simulations run, in order to draw a statistically significant solution for the amount of precision needed when manually editing head models, as per the specific needs of the BSS Lab studies. Though statistical analysis on more models should be done to determine the needed amount of model precision for varying contexts, this novel tool could optimize data pipelines for labs at Northeastern and beyond.