Discriminating Parkinson's disease using functional connectivity and brain network analysis

Author

Daniel Gellerup

Graduation Semester and Year

2016

Language

English

Document Type

Thesis

Degree Name

Master of Science in Industrial Engineering

Department

Industrial and Manufacturing Systems Engineering

First Advisor

Shouyi Wang

Abstract

In this study, we explored the use of functional connectivity patterns in fMRI data to classify subjects on the basis of Parkinson's disease. We explore various brain networks and features. We partition our fMRI data in 5 filtered frequency ranges. We use a proximal support vector machine paired with a minimum-redundancy and maximum-relevance feature selection method on each frequency range. We use a majority voting ensemble classification method on the results of the proximal support vector machine classification results. We use a double 5-fold cross validation scheme for model validation. We achieve 84% accuracy 74% sensitivity, and 93% specificity. Our results indicate that the ensemble method is effective compared to a single broad frequency range, and that Bonferroni correction may enhance classification results. We produce brain graphs to illustrate the brain networks of Parkinson's and control subjects.

Keywords

Parkinson's disease

Disciplines

Engineering | Operations Research, Systems Engineering and Industrial Engineering

License

This work is licensed under a Creative Commons Attribution-NonCommercial-Share Alike 4.0 International License.

Comments

Degree granted by The University of Texas at Arlington

Recommended Citation

Gellerup, Daniel, "Discriminating Parkinson's disease using functional connectivity and brain network analysis" (2016). Industrial, Manufacturing, and Systems Theses. 22.
https://mavmatrix.uta.edu/industrialmanusys_theses/22