Mean diffusivity (MD) derived from diffusion tensor imaging has shown its ability to assess the microscopic structural integrity damage of gray matter in amnestic mild cognitive impairment (aMCI), a prodromal stage of Alzheimer's disease (AD). However, little is known about the small world topology networks constructed by cortical MD in cognitive disease. In this work, we measured the cortical MD in the entire brain in patients with aMCI (n = 30) and AD (n = 30) compared with cognitive-normal (CNs) controls (n = 30), and then constructed the cortical diffusivity network by using graph-theoretical analysis. Compared with CNs, patients with aMCI and AD showed abnormal small-world property of cortical diffusivity networks (higher degree of clustering and longer path length), reflecting a less optimal topological organization. Moreover, the mean degree of connections of network in aMCI patients was characterized by lower than CNs but higher than AD. In addition, 11 hub regions were identified by negative correlations between MD and the score of Montreal Cognitive Assessment after multiple regression analysis, including bilateral hippocampi and related limbic system. Among those hub regions, the connectivity of the right olfactory cortex and middle orbital gyrus to the rest of brain regions were disrupted earlier than the other 9 regions in aMCI when compared to CN. In conclusion, the change of cortical diffusivity in topological network organization, mean degree of connections, and disrupted hub regions in aMCI may serve to identify patients in the prodromal stage of AD and reflect microstructural deterioration of neurodegeneration.
Keywords: Alzheimer's disease; amnestic mild cognitive impairment; mean diffusivity; small world networks.