Alzheimer’s Disease Detection using Machine Learning Techniques in 3D MR Images

Abstract

This study proposes a new method for the detection of Alzheimer’s Disease (AD) using first-order statistical features in 3D brain Magnetic Resonance(MR) images. Alzheimer’s disease is a neurodegenerative disorder that affects elderly people. This is a progressive disease and early detection and classification of AD can majorly help in controlling the disease. Recent studies use voxel-based brain MR image feature extraction techniques along with machine learning algorithms for this purpose. Grey and white matter of the brain gets affected and damaged due to AD and so studying these both prove to be more effective in predicting the disease. The proposed work uses 3D structural brain MR images to separate the white and grey matter MR images, extract 2D slices in the coronal, sagittal and axial directions and select the key slices from them for performing feature extraction on them. Feature extraction is applied on top of these slices to calculate the first-order statistical features and the prominent feature vectors generated by PCA are selected for further study. In the classification phase, different classifiers take the selected features as its input to predict the classes AD (Alzheimer’s Disease) or HC (Healthy Control) based on the observations in the validation set. Experimental results show that the accuracy of 90.9 % compared to other techniques.