Alzheimer's Disease (AD) is a progressive neurodegenerative disorder that primarily affects memory, cognitive function, and behavior, leading to severe impairment in daily activities. It is one of the most common causes of dementia among the elderly population worldwide and poses a significant burden on healthcare systems and society. Early diagnosis of Alzheimer's disease is essential for timely intervention, effective treatment planning, and slowing disease progression. However, conventional diagnostic techniques rely heavily on neuroimaging interpretation and neuropsychological assessments, which are often time-consuming, expensive, and dependent on clinical expertise. Recent advances in deep learning (DL) have demonstrated remarkable potential in automating the diagnosis of Alzheimer's disease using medical imaging data. This paper presents a de- tailed analysis of deep learning-based techniques for Alzheimer's disease detection, with a particular focus on convolutional neural network (CNN) architectures applied to magnetic resonance imaging (MRI) and non-MRI modalities. In addition to the analytical review, this study implements a CNN-based Alzheimer's disease detection system capable of classifying brain MRI images into four clinical stages: Non-Demented, Very Mild Demented, Mild Demented, and Moderate Demented. The proposed end-to- end framework enables automatic feature extraction and multi- class classification without the need for handcrafted features. Experimental observations demonstrate the feasibility and scalability of CNN-based approaches for early Alzheimer's disease detection and their potential application in clinical decision support systems.
T et al. (2026) studied this question.