05/15/2026
By Danielle Fretwell

The Francis College of Engineering, Department of Biomedical Engineering, invites you to attend a Doctoral Dissertation Proposal defense by Oluwasola Okhuoya titled: "Biometric Feature Characterization for the Classification of Alzheimer’s Disease Using Machine Learning Models"

Candidate Name: Oluwasola Okhuoya
Degree: Doctoral
Defense Date: Thursday, May 28, 2026
Time: 11 a.m. - noon
Location: Falmouth 302 Conference Room or join the Teams meeting

Committee:

  • Advisor: Lara Thompson, Ph.D., Associate Chair Undergraduate program, Professor of Biomedical Engineering
  • Walfre Franco, Ph.D., Department Chair, Associate Professor of Biomedical Engineering
  • Kavitha Chandra, Ph.D., Professor of Electrical and Computer Engineering

Abstract:
Among individuals over 65 years old, Alzheimer’s Disease (AD) is the most prevalent form of dementia and remains incurable. Because AD progresses gradually over an extended timeframe, and some physical deficits appear before the onset of objective cognitive decline in both the preclinical and clinical dementia stages, it raises the need for timely and multimodal diagnostics tied to the disease. The overarching goal of our project is to examine gait, eye movement, and electric brain activity as biomarkers for AD. Although these measures are widely recognized as biomarkers for various patients suffering from mild cognitive impairment, a knowledge gap exists specifically for these markers tied to AD. Here, we seek to examine electric brain activity, movement, and ocular behavior of individuals with AD compared to those without; further, we will examine the effectiveness of machine learning (ML) algorithms for predicting AD in individuals in conjunction with the features linked to these measures. This offers an alternative approach to current neuroimaging and electrophysiological techniques used to diagnose AD; furthermore, the integration of diverse biomarkers can enhance specificity and prove more efficacious than one modality alone. Our approach involves the analysis of publicly available datasets: 1) to characterize key features associated with AD and 2) to utilize these features in conjunction with ML techniques to detect and predict AD; this may reveal patterns tied to brain dysfunction and defective eye movements, as well as gait dysfunction in AD patients. Data preprocessing will be implemented using MATLAB software due to its established signal processing toolboxes for filtering, artifact rejection, and signal segmentation. Feature extraction (time domain, time-frequency domain, and frequency domain features) and machine learning workflows will be developed in Python via Visual Studio CodeData. A comprehensive statistical analysis will be conducted in RStudio (version 2026.04.0; Posit team, 2026) to identify significant differences in feature values across AD datasets. By using ML techniques, such as support vector machine, XGBoost, and random forest, we will identify optimal performers for disease classification based on sensitivity, prediction accuracy, and computational efficiency. Together these patterns will help us build a ML pipeline that can help us predict AD. If successful, this will be a non-invasive and accessible way to characterize and diagnose AD. The expected outcome of this research is to advance scientific understanding of biometric features that characterize AD and determine diagnosis by employing ML techniques.