An accurate Prediction and Classification of early Alzheimer’s Diseases using Machine Learning Algorithm

Abstract

Alzheimer’s disease (AD) is a progressive neurodegenerative disease that severely impairs memory, cognition and everyday functions. Early and accurate diagnosis of AD is important for timely clinical management but the conventional diagnostic protocols still suffer from subjectivity and high cost. In this study we propose a deep learning architecture for early AD prediction and classification based on Convolutional Neural Networks (CNNs). We trained the model using medical imaging data, specifically Magnetic Resonance Imaging (MRI), to automatically learn hierarchical spatial features indicative of the early AD brain subtly changing structure. Unlike traditional machine learning, the CNN uses convolutional layers, pooling, and fully connected layers to replace the need for handcrafted feature extraction to derive end-to-end learning. The proposed method is integrated with regularization and data augmentation to prevent overfitting and obtain robustness to different patient samples. Experimental evaluation results demonstrate that the CNN-based model possesses greater accuracy, sensitivity, and specificity than conventional classifiers and is able to accurately distinguish normal controls, patients with mild cognitive impairment, and patients with Alzheimer’s. This work shows the potential of CNNs as a valuable tool in computer Aided diagnosis (CAD) in the field of neurodegenerative diseases, contributing to the system of earlier diagnosis and clinical decision-making in their management.