Date of Award
3-2023
Document Type
Thesis
Degree Name
Master of Science
Department
Department of Operational Sciences
First Advisor
Daniel A. Pamplona, PhD
Abstract
Military members in combat situations are at a high risk for experiencing a Traumatic Brain Injury (TBI), often leaving them with Post-Traumatic Headache (PTH). There is currently a poor understanding of what factors predict if a patient will endure acute (short-term) PTH or Persistent PTH (PPTH). The Concussion Assessment, Research, and Education (CARE) Consortium dataset contains a variety of pre-and-post-injury concussion assessments taken by NCAA athletes and service academy cadets between 2014 and 2018. Participants in this study (n = 1769) suffered a TBI. Using this longitudinal data, three models were constructed: an LSTM model using the raw assessment scores, a feed-forward Artificial Neural Network (ANN) model using extracted features, and an LSTM model using sequential extracted features. These models distinguished patients with PPTH from patients with acute PTH with Precision-Recall Area Under the Curve (PR-AUC) scores of 0.44, 0.47, and 0.46, respectively. This research has set the foundation for future studies of the CARE Consortium dataset or other TBI datasets to improve the prognosis of PPTH.
AFIT Designator
AFIT-ENS-MS-23-M-113
Recommended Citation
Campo, Todd C., "Prediction of Persistent Post-Traumatic Headache with Long Short-Term Memory Networks" (2023). Theses and Dissertations. 6988.
https://scholar.afit.edu/etd/6988
Comments
A 12-month embargo was observed.
Approved for public release. Case number on file.