Comparative Analysis of Deep Learning Models for Alzheimer’s Disease Classification Using MRI Scans

Alzheimer’s disease (AD) is a progressive neurological disorder characterized by memory loss, cognitive decline, and behavioral changes, representing a major source of disability in the aging population. Accurate and early diagnosis is essential for enabling timely interventions that may help slow the progression of the disease. Traditional diagnostic procedures typically involve clinical assessments and MRI imaging to detect early structural brain alterations. However, manual interpretation of MRI data is often subjective, labor-intensive, and susceptible to inter-observer variability. Recent developments in deep learning (DL), particularly through Convolutional Neural Networks (CNNs), offer automated and efficient methods for analyzing medical images. This study presents a comparative evaluation of five advanced CNN architectures—ResNet-50, MobileNet, VGG16, InceptionV3, and Xception—applied to AD classification using transfer learning strategies. To address dataset imbalance, the Synthetic Minority Over-Sampling Technique (SMOTE) is employed, which generates new synthetic samples by interpolating between existing instances and their nearest neighbors. Experimental analysis shows that InceptionV3 achieves the highest performance with 98% accuracy and an F1-score of 98%, followed closely by MobileNet and VGG16, each with 97% accuracy. This approach can support healthcare practitioners by minimizing dependence on subjective analysis, leading to more consistent, scalable, and precise diagnostic practices for Alzheimer’s disease.