Abstract
The brain's resting-state energy consumption is expected to be driven by spontaneous activity. We previously used 50 resting-state fMRI (rs-fMRI) features to predict [(18)F]FDG SUVR as a proxy of glucose metabolism. Here, we expanded on our effort by estimating [(18)F]FDG kinetic parameters K(i) (irreversible uptake), K(1) (delivery), k(3) (phosphorylation) in a large healthy control group (n = 47). Describing the parameters' spatial distribution at high resolution (216 regions), we showed that K(1) is the least redundant (strong posteromedial pattern), and K(i) and k(3) have relevant differences (occipital cortices, cerebellum, thalamus). Using multilevel modeling, we investigated how much spatial variance of [(18)F]FDG parameters could be explained by a combination of a) rs-fMRI variables, b) cerebral blood flow (CBF) and metabolic rate of oxygen (CMRO(2)) from (15)O PET. Rs-fMRI-only models explained part of the individual variance in K(i) (35%), K(1) (14%), k(3) (21%), while combining rs-fMRI and CMRO(2) led to satisfactory description of K(i) (46%) especially. K(i) was sensitive to both local rs-fMRI variables (ReHo) and CMRO(2), k(3) to ReHo, K(1) to CMRO(2). This work represents a comprehensive assessment of the complex underpinnings of brain glucose consumption, and highlights links between 1) glucose phosphorylation and local brain activity, 2) glucose delivery and oxygen consumption.