Classification for Alzheimer's disease based on SVM using a spatial texture feature of cortical thickness

Features defined on the cortical surface derived from magnetic resonance imaging provide important information to diagnosis the Alzheimer's disease (AD) and its premonitory symptoms Mild Cognitive Impairment (MCI). In general, the methods based on region-wise features poorly reflect the detailed spatial variation of cortical thickness, and those based on vertex-wise features are sensitive to noise. In this paper, we propose an innovative method to describe the spatial texture information of cortical thickness named Spatial Multi-scale Block Local Binary Patterns. And combine this information with the region-wise characters of cortical thickness as the final classification features. The performance based on the proposed method is significantly higher for the AD/NC, AD/MCI and sMCI/cMCI groups than using PCA or taking the average of cortical thickness zones as classification features.