Comparative Analysis of Skull-Stripped Versus Non-Stripped MRI Scans for CNN-Based Alzheimer's Disease Classification

Abstract

Alzheimer's disease (AD) is a debilitating neurodegenerative condition characterized by progressive cognitive decline. Early detection enables experts to initiate preventive treatment at the initial possible stage. The primary focus of the work is to develop an automated detection system using a Convolutional Neural Network (CNN), capable of analyzing the brain magnetic resonance imaging (MRI) scans acquired from the Alzheimer’s Disease Neuroimaging Initiative (ADNI). Moreover, skull removal was accomplished on the brain MR scans to create a new dataset. A comparative analysis was conducted on both datasets. Results showed impressive classification accuracies: 98.27% for skull-included scans and slightly higher for skull-stripped scans. The findings of the study demonstrate the successful application of deep learning (DL) techniques in neurological disease detection. The slight improvement in accuracy with the stripped version enables future preprocessing considerations in these areas.