The activity of the brain exhibits stereotypical patterns at various temporal and spatial scales. In particular, recent studies using fMRI and MEG have shown that the slow fluctuations (timescale of a second) exhibit correlations across the whole cortex. These macroscopic activity patterns, also termed functional connectivity (FC), suggest the existence of a (dynamic) resting state. How it relates to and interacts with cognitive tasks has raised recent interest within the neuroscience community. The recent dynamic mean-field (DMF) model developed by our group and collaborators has demonstrated that the structure of the resting-state FC patterns can be explained to a large extent by the long-range interregional cortical projections (or structural connectivity, SC). These have been mapped recently for the whole human cortex using tensor diffusion imaging techniques.
The focus of this presentation is on the relationship between SC and FC. In the DMF model, dynamical regime of the interconnected cortical areas depends on a global coupling strength applied to SC. Moreover it strong affects the FC patterns exhibited by the network. This leads to the concept of \\\"effective connectivity\\\" (EC), namely the feedback that drives the neural dynamics. Here we examine two specific aspects of EC.
First, we address the problem of inferring SC/EC when FC is known for a recurrently connected population. Second, we examine how the different procedures for optimizing EC (e.g., tuning global coupling versus individual weights) lead to distinct dynamical states and functional organization for the DMF cortical model. Together, these results are steps toward building a model that links neural dynamics to functions, in which synchrony between brain areas (FC) shapes or gates the computations performed by local neural circuits.
CIMH Mannheim, Therapy Building, big lecture hall, 15:00-16:00