Disease-specific hypometabolic epicenters in dementia: Neural networks and clinical correlations

BackgroundIn dementia, it remains unclear how the hypometabolism pattern relates to the intrinsic network, where the disease-specific hypometabolism epicenters are located, and what the epicenter's clinical effects are.ObjectiveIn this cross-sectional study, we aim to identify disease-specific hypometabolism epicenters in Alzheimer's disease (AD), posterior cortical atrophy (PCA), and frontotemporal dementia (FTD) using a data-driven network-based approach; to determine whether intrinsic structural and functional brain networks shape the spatial distribution of hypometabolism; and to examine the clinical relevance of epicenter hypometabolism by assessing its association with network disruption and disease-related clinical features.MethodsDisease-specific epicenters were identified using data-driven analysis in AD (n = 98), PCA (n = 40), and FTD (n = 71). Hypometabolism was measured by [¹⁸F]-fluorodeoxyglucose positron emission tomography (FDG-PET). Brain regions with reduced network connectivity to the epicenter were identified, and the correlation between changes in the epicenter and clinical features was examined.ResultsRegional hypometabolism was significantly correlated with that of the network-connected neighbors. The disease-specific hypometabolism epicenters were the bilateral angular gyrus in AD, the angular and middle occipital gyri in PCA, and the rectus and anterior cingulate gyri in FTD. The hypometabolism of the epicenters contributed to the disruption of the network and was correlated with clinical features.ConclusionsThe hypometabolism epicenters are disease-specific and correlated with clinical features in dementia, providing insight into neuromodulation targets.