Brain Topological Changes in Subjective Cognitive Decline and Associations with Amyloid Stages

Background Amyloid spreads throughout the cortex in Alzheimer's disease (AD), which is thought to cause disturbances in brain networks. It is not clear how structural and functional connectivity change as amyloid accumulates progressively across cortical stages in individuals with subjective cognitive decline (SCD), a risk condition for AD. Methods Graph-theoretic analyses of functional and structural brain networks were performed using resting-state functional magnetic resonance imaging and diffusion tensor imaging in 100 individuals with SCD and 86 normal controls (NC) derived from an ongoing large cohort study. Topological properties at node and network levels were characterized as indicators of information exchange efficiency and network robustness. Amyloid burden was quantified in a subset of 55 individuals with SCD by amyloid positron emission tomography using a frequency-based staging method, yielding mean regional standardized uptake value ratios (SUVr) for four anatomical divisions and global SUVr for the entire cortex, which were further correlated with topological property measures. Results Individuals with SCD showed increased functional nodal efficiency and structural nodal betweenness in the left anterior and median cingulate gyri compared to NC. No group differences in network-level properties were found. Frequency-based amyloid staging revealed four anatomical divisions, including the fusiform and lateral temporal gyri (> 50%), occipital areas (30%-50%), default mode network, the midline brain and lateral frontotemporal areas (10%-30%), and the remaining cortex (< 10%). Global and regional SUVr of the four amyloid stages were positively associated with the node-level properties of a set of default mode network hubs, with the left anterior and posterior cingulate gyri being congruently associated with all amyloid stages. Conclusion Individuals with SCD showed increased information exchange efficiency in the left cingulate, with unchanged network-level effectiveness or robustness of a complex network. In this at-risk population, progressive amyloid accumulation across cortical stages continuously influences the adaptation of functional and structural networks via default mode network hubs, irrespective of the local physical proximity of the amyloid pathology. The positive associations between node-level topological properties and amyloid burden suggest an increasing burden of information processing in the cortical hubs during early cortical amyloid deposition in this risk population for Alzheimer's disease.