Computer aided detection of imaging biomarkers for Alzheimer's disease

In this paper, we present a novel approach for the computer aided detection of imaging biomarkers responsible for Alzheimer's disease (AD) from magnetic resonance imaging (MRI) using meta-cognitive radial basis function network (McRBFN) classifier. The McRBFN classifier uses voxel-based morphometric features extracted from MRI and employs a sequential projection-based learning (PBL) algorithm for classification. We propose a recursive feature elimination approach (called PBL-McRBFN-RFE) to identify the most relevant and meaningful imaging biomarkers for AD detection. The study has been conducted using the well-known open access series of imaging studies dataset. The brain regions identified by the PBL-McRBFNRFE feature selection approach include hippocampus, parahippocampal gyrus, superior temporal gyrus, insula, precentral gyrus and extra nuclear, which have also been reported as critical regions in the medical literature. Further, we also conducted a study based on the age to identify the brain regions responsible for the onset of AD.