Early Detection Of Alzheimer’s Disease By Leveraging Deep Learning Classification

Abstract

Alzheimer’s disease is an irreversible, progressive neurodegenerative disease that is very common in today’s elderly, and according to Centers for Disease Control and Prevention (CDC), Alzheimer’s disease is the sixth leading cause of death in the United States of America. Alzheimer’s gradually declines memory and cognitive abilities via deterioration to the hippocampus and amygdala areas of the brain. The loss of these cognitive skills is referred to as dementia, and dementia in patients suffering from Alzheimer’s disease can range in severity from very mildly demented to moderately demented. Early recognition of the severity of the dementia can help patients minimize damage and maintain the quality of their life. Magnetic Resonance Imaging (MRI) is the standard tool to study the structural changes in the brain by taking segmentations of the brain atimed intervals. In this study, we have used MRI segmentation images as the input for a machine learning model based on convolutional neural networks to detect the stage of Alzheimer’s dementia, if any. This model classifies the MRI segmentation data into four categories, being non-demented, very mildly demented, mildly demented, and moderately demented, using MRI data from a released Kaggle competition. This model is of low cost, easy to implement, and exhibits high performance and accuracy, boasting a 99.8% validation accuracy.