Medicine and Health Sciences Commons^™

An Explainable Deep Learning Prediction Model For Severity Of Alzheimer's Disease From Brain Images, Godwin O. Ekuma

MSU Graduate Theses

Deep Convolutional Neural Networks (CNNs) have become the go-to method for medical imaging classification on various imaging modalities for binary and multiclass problems. Deep CNNs extract spatial features from image data hierarchically, with deeper layers learning more relevant features for the classification application. The effectiveness of deep learning models are hampered by limited data sets, skewed class distributions, and the undesirable "black box" of neural networks, which decreases their understandability and usability in precision medicine applications. This thesis addresses the challenge of building an explainable deep learning model for a clinical application: predicting the severity of Alzheimer's disease (AD). AD …

Comparison Of Longitudinal Changes In Resting State Functional Magnetic Resonance Imaging Between Alzheimer’S And Healthy Controls, Berk Can Yilmaz

Theses

Resting State Functional Magnetic Resonance Imaging (rs-fMRI) is a technique that is widely used for analyzing brain function using different approaches and methods. This study involves rs-fMRI analysis of Blood Oxygenation Level Dependent (BOLD) signals acquired from Alzheimer’s disease (AD) Patients and Healthy Controls (HC). Each subject in the study had both functional and anatomical images with at least one rs-fMRI scan with their Anatomical (T1) scans. Previous rs-fMRI studies have demonstrated that AD shows differences in Amplitude of Low Frequency (<0.1 Hz) Fluctuations (ALFF), and Regional Homogeneity (ReHo) measures according to HCs.

The aim of the study is to investigate individual and group level differences using ReHo and mALFF related …

Mixture Modeling With Applications In Alzheimer's Disease, Frank Appiah

Theses and Dissertations--Epidemiology and Biostatistics

This dissertation involves an application of mixture of regression models to 114 individuals who are cognitively intact (from the Alzheimer's Disease and Neuroimaging Initiative-ADNI, data). The correct number of components in the model were estimated with the Singular BIC (SBIC), marking the first time it has been applied to such a problem. The smallest true model in conjunction with the approximation of SBIC was fixed at 1. The resulting posterior probabilities from the model were used to estimate the probability of a person transitioning and risk plots were obtained that could in principle be used by clinicians to identify patients …