Abstract
The ability to explore the architecture and functioning of the human brain in a non-invasive manner is an exciting development in modern medicine, psychology, and neuroscience. Current neuroimaging techniques are clinically useful and provide insights into the mechanisms underlying brain function and dysfunction. Alzheimer’s disease (AD) is a neurological disorder that causes memory loss and cognitive decline. It is the most common cause of these symptoms and impairs a person’s ability to function independently. Early detection of AD is crucial for early intervention and symptom management. In this study, we aimed to identify essential features in functional magnetic resonance imaging (fMRI) data using a patch-based approach (PBA) with the help of two 3D-Convolutional Neural Networks (CNNs) with fully connected layers. We then employed a decision tree and a k-nearest neighbor classifier to detect AD in three stages (AD, Mild cognitive impairment [MCI], and Normal Control [NC]) from the AD Neuroimaging Initiative (ADNI) public data set. Compared with results in the literature, our results show improved performance in categorizing AD.

This publication has 28 references indexed in Scilit: