Detecting grey matter changes in preclinical phase of Alzheimer's disease by voxel-based morphometric and textural features: A preliminary study

Abstract

In the present study, a voxel-based morphometry (VBM) and three dimensional texture analysis were used to detect brain structural changes of AD subjects in preclinical phase. Three groups of T1-weighed MRI data were selected from Open Access Series of Imaging Studies (OASIS) database, including 13 converted subjects (CS) who were cognitively healthy at baseline, 13 non-demented subjects (NS) remained cognitively healthy, and 64 demented subjects (DS) for control study. Results showed grey matter (GM) atrophy in bilateral occipital lobe, superior temporal gyrus, and left inferior temporal lobe in CS (p< 0.001); textural features of the largest cluster significantly differed from that of NS, with the classification accuracy of 88.5% through cross validation, which may highlight the key role of these regions in prediction of the disease. Morphometric measurement plus texture analysis presents feasible in detecting preclinical AD.