Local neurodegeneration and global connectivity adaptation across the FTD‐AD spectrum

Background Neurodegenerative diseases involve weakened functional connectivity in disease‐targeted brain areas. Equally important but overlooked is the hyperconnectivity that appears in other brain areas (Hillary and Grafman, TICS 2017). Hyperconnectivity has been attributed to processes like disinhibition, imbalance, compensation, and reserve. It is critical to understand the neuroanatomical mechanism underlying hyperconnectivity because this functional process may accelerate subsequent disease progression. We performed structure‐function mapping for patients across subtypes and stages of the frontotemporal dementia‐Alzheimer’s disease atrophy spectrum. Our goal is to develop a comprehensive model relating diverse focal atrophy patterns to corresponding brain‐wide functional connectivity reconfigurations. Method We studied patients with Alzheimer’s disease (AD), behavioral variant FTD (bvFTD), semantic and nonfluent variant primary progressive aphasia (svPPA/nfvPPA), cortical basal syndrome (CBS), and healthy controls (HC). We included subjects who received a clinical diagnosis at the UCSF Memory and Aging Center and had structural and task‐free functional MRI scans (n=281). Each subject’s gray matter atrophy map was measured using voxel‐based morphometry volume loss in 273 ROIs. Functional connectivity matrices (273x273) were derived for each subject. We combined data for all patients and controls into a structural data matrix (281x273) and a functional data matrix (281x37128). We then performed partial least squares regression to find components that maximized the covariance between structure and function. Result The first PLS component captured the relationship between global atrophy burden and a distributed pattern of functional connectivity loss in unimodal cortical areas and enhancement in subcortical‐cortical pathways (r=0.64; Figure 1 ). The second PLS component showed svPPA‐like anterior temporal atrophy corresponding to atrophy‐proximal connectivity deficits, with enhancements in contralateral frontoparietal areas (r=0.67). The third PLS component revealed a spectrum from anterior (bvFTD) to posterior (AD), contrasting frontal‐insular atrophy, connectivity deficits, and parietal connectivity enhancements versus parietal atrophy, connectivity deficits, and frontal connectivity enhancements (r=0.51). Conclusion Specific atrophy subtypes across the FTD‐AD spectrum associate with proximal functional connectivity reductions. Intriguingly, these subtypes also exhibit concomitant functional connectivity enhancements in more distal areas. The enhancements are of the same magnitude as the deficits and may represent a general principle of functional “load‐shifting” (Jones et al, Brain 2016) away from disease‐targeted areas.