SVR Analysis of Alzheimer's Disease relationship between MRI Metrics and Braak Score in the Hippocampus subregions

Abstract

Alzheimer's Disease (AD) plagues 1 and 9 seniors, with a higher relation to death of AD. MRI is a noninvasive imaging technique that can provide quantitative measurements of various physical properties of tissues within the body. The aim of this study was to explore the relationship between MRI modalities and the progression of AD through hippocampal subfields using different MRI techniques with the use of Machine learning techniques. A total of eleven samples were scanned with different types of MRI modalities. This study used Quantitative MRI (qMRI) metric values in different hippocampal subfields’ and Braak scoring with differing stages of AD using two SVR models to determine their relative predictive values. The first regressor model had shown a strong correlation between mean signal diffusion Mean Diffusivity (MD) in CA4 which was similar to what other studies have found. The second model had also identified MD as significant along with PA. While this study was unable to confirm other trends found in other studies, the SVR models were able to identify a certain trend with a small but high quality sample size. These findings have shed light on the potential of utilizing Machine Learning techniques, and MRI technologies to understand AD and its impact on the hippocampus