Measuring directed functional connectivity in mouse fMRI networks using Granger Causality

Resting-state functional magnetic resonance imaging (rsfMRI) of the mouse brain has revealed the presence of robust functional connectivity networks, including an antero-posterior system reminiscent of the human default network (DMN) and correlations between anterior insular and cingulate cortices recapitulating features of the human “salience network”. However, rsfMRI networks are typically identified using symmetric measurements of correlation that do not provide a description of directional information flow within individual network nodes. Recent progress has allowed the measure of directed maps of functional connectivity in the human brain, providing a novel interpretative dimension that could advance our understanding of the brains’ functional organization. Here, we used Granger Causality (GC), a measure of directed causation, to investigate the direction of information flow within mouse rsfMRI networks characterized by unidirectional (i.e. frontal-hippocampal) as well as reciprocal (e.g. DMN) underlying connectional architecture. We observed robust hippocampal-prefrontal dominant connectivity along the direction of projecting ventro-subicular neurons both at single subject and population level. Analysis of key DMN nodes revealed the presence of directed functional connectivity from temporal associative cortical regions to prefrontal and retrosplenial cortex, reminiscent of directional connectivity patterns described for the human DMN. We also found robust directional connectivity from insular to prefrontal areas. In a separate study, we reproduced the same directional connectivity fingerprints and showed that mice recapitulating a mutation associated to autism spectrum disorder exhibited reduced or altered directional connectivity. Collectively, our results document converging directional connectivity towards retrosplenial and prefrontal cortical areas consistent with higher integrative functions subserved by these regions, and provide a first description of directional topology in resting-state connectivity networks that complements ongoing research in the macroscale organization of the mouse brain.