Degeneration of Subcortical White Matter in Alzheimer's Disease: Atlas-Based Automated Mapping and Its Diagnostic Utility Based on Multi-Variate Model

Takashi Yoshiura¹, Akio Hiwatashi¹, Koji Yamashita¹, Hironori Kamano¹, Yukihisa Takayama¹, Eiki Nagao¹, Tuvshinjargal Dashjamts¹, Hiroshi Honda¹

We measured mean diffusivities (MD) in subcortical white matter (WM) in 78 different cortical regions using an atlas-based mapping method in 33 patients with Alzheimerfs disease (AD) and 28 healthy control subjects to determine whether the topographical pattern of the diffusion abnormalities can be used to diagnose AD. Uni-variate analysis in which discrimination was attempted based on MD in the single region resulted in the accuracy of 88.5 %. Multi-variate analysis in which a linear discriminant function based on MDs from multiple cortical regions increased the accuracy up to 96.7 %.