Abstract

Info

Title:: Functional magnetic resonance imaging transformation for classification of Alzheimers disease

Author(s):: JIA Hongfei; WANG Yu; XIAO Hongbing; XING Suxia; School of Artificial Intelligence, Beijing Technology and Business University, Beijing 100048, China

Keywords:: Keywords: Alzheimers disease functional magnetic resonance imaging 3DPCANet support vector machine regional homogeneit

Abstract:: Abstract: To identify the early symptoms of Alzheimers disease (AD), an improved 3DPCANet combined with functional magnetic resonance imaging (fMRI) transformation is proposed for classifying patients at different stages of AD. After the fMRI image is preprocessed, regional homogeneity image transformation is carried out for the preprocessed fMRI images. Then, the features of the transformed images are extracted using the improved 3DPCANet. Finally, support vector machine is used for AD classification. The experimental results show that the improved 3DPCANet model can be used to extract the effective classification features from the images after transformation. The classification accuracies reach 90.00%, 88.89%, and 88.00% for late mild cognitive impairment vs AD, subjective memory decline vs AD, and subjective memory decline vs early mild cognitive impairment, respectively, which proves the feasibility and effectiveness of the proposed method.