Analytical Approaches for Enhancing fMRI Brain Mapping Accuracy in Individuals

Author

• David G. Ellis, University of Nebraska Medical CenterFollow

ORCID ID

0000-0002-3718-6836

Graduation Date

Summer 8-14-2026

Document Type

Dissertation

Degree Name

Doctor of Philosophy (PhD)

Programs

Biomedical Informatics

First Advisor

Michele Aizenberg

Abstract

Individual brain mapping focuses on accurately and reliably mapping the brain of the individual rather than a group of individuals. A primary use case for individual brain mapping is in preoperative planning for brain tumor resection patients. When preparing for and performing surgery, the surgeon needs to know which regions of the brain, if lesioned, would critically impact patient well-being and survival. Because individual brain mapping deviates from the goals of traditional brain mapping, special attention is needed to optimize its utility given its unique objectives and constraints. This dissertation proposed and evaluated three different approaches for enhancing the accuracy, reliability, and utility of fMRI for individual brain mapping. The first approach implemented an anatomically constrained smoothing methodology that prevents smoothing across cortical boundaries. The analysis showed that unconstrained smoothing spreads activation across cortical boundaries, increases white matter activation, and biases graph theory connectivity metrics. In contrast, the proposed anatomically constrained smoothing mitigated these artifacts by preventing smoothing across cortical boundaries, while still increasing reliability. The second approach utilized the group-average mapping results from previously analyzed individuals to inform new individual brain mapping cases using GPU-accelerated Bayesian inference. In both healthy individuals and brain tumor patients, utilizing group-average informed Bayesian inference improved the accuracy of the mapping results. The third approach trained a deep learning model to provide individual-specific predictions that were then used to inform the individual brain mapping results. Overall, this dissertation showed that methods developed for analyzing fMRI data in individual subjects can improve results in ways that may meaningfully enhance the clinical utility of task-fMRI. More broadly, this work provides a framework for thinking about, evaluating, and optimizing techniques for clinical brain mapping in individual patients.

Rights

The author holds the copyright to this work and any reuse or permissions must be obtained from the author directly.

Recommended Citation

Ellis, David G., "Analytical Approaches for Enhancing fMRI Brain Mapping Accuracy in Individuals" (2026). Theses & Dissertations. 1107.
https://digitalcommons.unmc.edu/etd/1107