EARLY DETECTION OF ALZHEIMER’S DISEASE USING MAGNETIC RESONANCE IMAGING AND DIFFUSION TENSOR IMAGING

Abstract

Recently, classification and prediction of several diseases can be performed via machine learning methodologies. Of particular importance comes the neurodegenerative diseases, those related to losing neurons and brain cognitive functions, which encompasses Alzheimer’s Disease (AD). The large amount of data being readily-available and the increasing computer powers help boost the unleashed growing usage of these machine learning algorithms. The objectives of this work were 1) to find out the class activation maps (CAMs) deriving the network decision, and 2) to detect AD and its earlier pathology; namely, the mild cognitive impairment (MCI), from healthy controls (HC) in robust and low-cost network design. Both tasks were implemented using convolutional neural networks (CNNs).