Effective connectivity estimation from laminar profiles measured with sub-millimeter ultra-high field fMRI
Rainer Goebel (PI), Alessandra Pizzuti
Despite the availability of methods to estimate effective connectivity (EC) from fMRI data such as DCM and Granger Causality Mapping for more than 15 years, the sluggishness of the BOLD response and the unknown region-specific physiological (hemodynamic) response delays pose severe limitations to reliably estimate directed influences between brain regions from conventional fMRI data. Using sub-millimeter ultra-high field (7+ Tesla) fMRI to separate signals from laminar profiles may provide an indirect approach to differentiate bottom-up and top-down influences based on the generic anatomy of cortical connections showing that sensory input arrives in layer 4 whereas feedback connections target the deep and superficial layers. Using visual attention experiments targeting various levels of the visual hierarchy (V1, MT, LOC, FFA) we will investigate whether recorded laminar cortical profiles are able to inform EC models as well as detailed laminar spiking neural network models.