What makes connections effective? An introduction to model‐based measures of brain connectivity (21.1)

Abstract

Human brain mapping has traditionally concentrated on the localization of functions, with the primary question being where in the brain, activity for a particular function can be mapped to or located. This focus on localization or functional segregation has been critical for demonstrating the correspondence between specific brain regions and functions such as the perception of color and visual motion among others. However, it has also been recognized that complex behaviors defy an easy one‐to‐one correspondence between functions and single brain regions; instead these behaviors emerge from the integrated activity of multiple cortical and subcortical areas or brain networks. Brain network connectivity can be examined on several levels. Structural connectivity defines the general organization of the brain, but it does not tell us, which connections are engaged for a particular brain function. Functional connectivity highlights regions with similar time courses of activity. However, since functional connectivity is based on measures of statistical dependency it cannot tell either about the direction of these interactions or whether the interactions represent direct or polysynaptic effects. Effective connectivity captures the dynamical nature of neuronal interactions together with structural models to show how one neuronal population influences another. This talk will focus on the concept of effective connectivity, its relationship to structural and functional connectivity, its response to experimental context and currently available methods for its quantification including Granger causality and dynamic causal modeling.