Abstract

The brain has a high energetic cost to support neuronal activity, requiring both oxygen and glucose supply from the cerebral vascular system. Additionally, the brain functions through complex patterns of interconnectivity between neuronal assemblies giving rise to functional network architectures that can be investigated across multiple spatial scales. Different brain regions have different roles and importance within these network architectures, with some regions exhibiting more global importance by being involved in cross-network communication while other being predominantly involved in local connections. There are indications that regions exhibiting a more global role in inter networks connectivity are characterized by a higher and more efficient metabolic profile, leading to differences in metabolic properties when compared to more locally connected regions. Understanding the link between oxygen/glucose metabolism and functional features of brain network architectures, across different spatial scales, is of primary importance.
This thesis consists of three original studies combining human brain resting-state multimodal neuroimaging and transcriptional data to investigate the glucose/oxygen metabolic costs of brain functional connectivity. We quantified glucose metabolism from positron emission tomography, and oxygen metabolism and functional connectivity from magnetic resonance imaging. In the first study, we highlight how the oxygen/glucose metabolism of brain regions can non-linearly relate to their functional hubness, within the resting-state networks of the brain across a nested hierarchy. We found that an increase in oxygen/glucose metabolism is associated with a non-linear increase in functional hubness where increase rates are both network- and scale-dependent. In the second study, we show specific transcriptional signatures that characterize the oxygen/glucose metabolic costs of regions involved in network global versus local centrality. This study highlights the different metabolic profiles of local and global regions, with gene expression related to oxidative metabolism and synaptic pathways being enriched in association with spatial patterns in common with resting blood flow and metabolism (oxygen and glucose) and globally-connected regions. In the third study, we demonstrate that there are oxygen/glucose metabolic costs to the functional integration and segregation of resting-state networks. We highlight that the metabolic costs of functional integration could reflect the hierarchical organization of the brain from unimodal to transmodal regions.