Structural MRI based Alzheimer’s Disease prognosis using 3D Convolutional Neural Network and Support Vector Machine

Abstract

Alzheimer’s Disease (AD) is a prevalent, irreversible, chronic, and progressive disease leading to structural changes in the brain, causing the cognitive decline of brain function. Early detection of AD before clinical manifestation is crucial for patient care, effective therapeutic measures, and cost-saving. To address the challenge of timely diagnosis, in this paper, we designed a 3D-CNN framework with SVM as a classifier to harness the advantages of both Deep learning (DL) and Machine learning (ML). Experiments on AD neuroimaging initiative (ADNI) dataset yields fair 91.85% accuracy, 95.56% sensitivity, 90% specificity, 82.69% precision, and 88.66% f-score exhibiting the SVM outperformance over other ML classifiers. Thus, the proposed model is effective for the prognosis of AD and can be incorporated in healthcare.