Georgia Institute of Technology
Standardizing entropy measures to characterize brain dyanmics in fMRI
Abstract
dc:description.abstractThis thesis investigates the reliability, standardization, and practical utility of entropy-based measures for characterizing brain dynamics in functional magnetic resonance imaging (fMRI). First, a systematic methodological review of 203 studies reveals that while entropy approaches are widely used across resting-state and task-based paradigms, their application is highly heterogeneous, with substantial variability in parameterization, analytical scope, and reporting practices, limiting reproducibility and cross-study comparability. Second, using a unified analysis framework applied to large-scale fMRI data, the study systematically evaluates major entropy measures, including Shannon entropy, Sample Entropy, Multiscale Entropy, Approximate Entropy, and Transfer Entropy, across criteria such as biological interpretability, cross-subject reproducibility, temporal stability, and parameter sensitivity, identifying Sample Entropy and multiscale entropy (Area Under Curve) as the most robust and reliable metrics. Finally, the predictive utility of entropy is demonstrated through machine-learning classification of eight brain states (rest and task conditions), where entropy-based features outperform conventional functional connectivity metrics and achieve highest performance when combined with them, reaching over 90% balanced accuracy. Overall, the findings establish entropy as a complementary and informative framework for capturing complex neural dynamics beyond traditional connectivity measures, while emphasizing that its effective use depends on careful methodological standardization and selection of appropriate entropy formulations.
Degree
thesis:*- Name thesis:degree_name
- Biomedical Engineering, MS
- Department thesis:degree_department
- Biomedical Engineering (Joint GT/Emory Department)
- Grantor
- Georgia Institute of Technology
- Year dc:date.issued
- 2026
Author and committee
dc:creator, dc:contributor.*- Author dc:creator
-
- Park, Jeonghoon
- Advisor dc:contributor.advisor
-
- Mason, Borzin
- Committee members dc:contributor.committeemember
-
- John Oshinski
- Keilholz, Shella
Rights
- Language dc:language.iso
- English
Identifiers
dc:identifier.*- Handle dc:identifier.uri
- https://hdl.handle.net/1853/81665