A Hybrid Deep Learning model for predicting the early Alzheimer's Disease stages using MRI

The continuous evolution of technology in Biomedicine has given satisfactory answers for several complex diseases. Alzheimer's disease (AD), one of the major neurodegenerative diseases that cause dementia, belongs to this category. So far, no cure reverses or stops the biological changes that occur in the brains of patients; however, the early diagnosis and early intervention of Alzheimer's disease is a crucial step in reducing the burden on both the patient and the caregivers. One of the predominant ways to deal with this difficulty is by integrating artificial intelligence and large-scale biomedical data. In this direction, Magnetic resonance imaging (MRI) offers high-resolution data, which can be decrypted through artificial intelligence tools. In recent years, the research community has shifted to deep learning methods applied to medical images for the early diagnosis of Alzheimer's disease. In the present work, we propose a hybrid (called CNN-SVM) model based on Convolutional Neural Networks (CNN) and the Support Vector Machines (SVM) classifier to predict the early AD stages from MRI. Our results showed that the proposed CNN-SVM model outperforms other well-known algorithms supporting the more effective AD diagnosis.