Metabolic neural networks of the marmoset brain in awake and isoflurane-anesthetized resting states.

The common marmoset is a key model in translational neuroscience, yet the reliance on anesthesia for neuroimaging introduces significant confounds. To provide a physiological basis for future investigations we characterized cerebral glucose metabolism during naturally awake and isoflurane anesthetized states using [18F]FDG PET. Four adult male marmosets underwent repeated PET/CT sessions in a crossover design, alternating tracer uptake between awake and anesthetized conditions. Following normalization of imaging data to the Marmoset Brain Mapping atlas, standardized uptake values (SUV) and SUV ratios (SUVR) were calculated. Metabolic topology was then assessed by combining non-parametric permutation-based voxel-wise statistical mapping with network-level correlation analysis. We observed that anesthesia induces a fundamental reorganization of metabolic topology rather than a uniform global suppression. Voxel-wise analysis revealed significantly elevated metabolism in auditory, inferior temporal, and prefrontal cortices during the awake state, whereas anesthesia induced a redistribution toward subcortical regions, with preserved activity in the hypothalamus and substantia nigra. Furthermore, network analysis revealed a significant increase in mean cortical correlation strength under anesthesia (awake: r = 0.162; anesthetized: r = 0.209), reflecting paradoxical hypersynchrony driven by the compression of metabolic signal variance. In contrast, subcortical connectivity strength declined (awake: r = 0.291; anesthetized: r = 0.220), indicating that subcortical regulatory centers displayed a disruption of cohesive functional coupling. By systematically defining these state-dependent alterations, our findings offer a physiological reference for distinguishing biological baselines from anesthetic artifacts, thereby reinforcing the translational validity of future marmoset investigations.