Neuroanatomical-Based Machine Learning Prediction of Alzheimer's Disease Across Sex and Age
Abstract: Alzheimer's Disease (AD) is a progressive neurodegenerative disorder marked by cognitive decline and memory loss. As of 2024, it affects about 1 in 9 Americans aged 65 and older, totaling 6.9 million individuals, with women comprising 4.2 million and men 2.7 million. Magnetic resonance imaging (MRI) is crucial for examining brain structure and identifying potential AD biomarkers, essential for early detection, accurate diagnosis, and effective disease management. This study employs machine learning techniques to analyze anatomical MRI scans and identify key brain regions associated with AD, focusing on the influence of sex and age on these biomarkers. Using the Random Forest Algorithm, the study achieved 92.87% accuracy in detecting AD from mild cognitive impairment and cognitive normals. The findings revealed consistent volume decreases in the hippocampus, amygdala, and entorhinal cortex across sexes and age groups, with specific variations such as decreased right amygdala volume in younger males and decreased left middle temporal cortex volume in females. These insights can guide clinical research and treatment strategies, helping to identify neuroanatomical markers and therapeutic targets for future interventions.
Advisor: Professor Loris Fichera
Committee: Professor Benjamin Nephew, Professor Haichong Zhang
Zoom Link: https://wpi.zoom.us/j/96604062037