Masters Applied Projects

Project 20

Utilizing Machine Learning to Determine Most Predictive Tests for Mild Cognitive Impairment and Alzheimer’s Disease

Student: Tyler Rose

Mentors: Dr. Sydney Schaefer – SBHSE
Dr. Scott Beeman – SBHSE
Dr. Mike Malek-Ahmadi – Banner Alzheimer’s Institute

YouTube Link: View the video link below before joining the zoom meeting

Zoom link: https://asu.zoom.us/j/85070618717

Time: 10am – 2pm

Abstract

Alzheimer’s disease (AD) and Mild Cognitive Impairment (MCI) are inherently troubling due to the inability to cure/treat the cause. Current treatment plans are able to slow the progression of the pathologies, but are unable to revert any of the damage the patient has already suffered. This has led to a search for predictive tests, or combinations thereof, which can accurately predict a future progression to MCI or conversion to AD. In this retrospective look at the dataset published by the National Alzheimer’s Coordinating Center (NACC) machine learning algorithms were optimized for prognosis predictions of patients with known results. Through this optimization, the predictability of currently used tests were weighed against one another in order to determine the most useful testing battery set. Hopefully, this information can lead to the generation of a standardized specific set of tests with known predictive effectiveness. This would in turn benefit patients by slowing the progression of their respective condition and allow them to maintain cognitive normalcy and independence for years longer than previously capable.