Metabolic connectivity: methods and applications

Metabolic connectivity modelling aims to detect functionally interacting brain regions based on PET recordings with the glucose analogue [F]fluorodeoxyglucose (FDG). Here, we outline the most popular metabolic connectivity methods and summarize recent applications in clinical and basic neuroscience.

Metabolic connectivity is modelled by various methods including a seed correlation, sparse inverse covariance estimation, independent component analysis and graph theory. Given its multivariate nature, metabolic connectivity possess added value relative to conventional univariate analyses of FDG-PET data. As such, metabolic connectivity provides valuable insights into pathophysiology and diagnosis of dementing, movement disorders, and epilepsy. Metabolic connectivity can also identify resting state networks resembling patterns of functional connectivity as derived from functional MRI data.

Metabolic connectivity is a valuable concept in the fast-developing field of brain connectivity, at least as reasonable as functional connectivity of functional MRI. So far, the value of metabolic connectivity is best established in neurodegenerative disorders, but studies in other brain diseases as well as in the healthy state are emerging. Growing evidence indicates that metabolic connectivity may serve a marker of normal and pathological cognitive function. A relationship of metabolic connectivity with structural and functional connectivity is yet to be established.