Neural traffic as voxel-based measure of cerebral functional connectivity in fMRI

Abstract

To access functional connectivity by in vivo brain imaging voxel-by-voxel, we developed a novel approach named neural traffic (NT). NT depicts the intensity of functional connectivity on a voxel-by-voxel basis in the whole brain. Functional magnetic resonance imaging (fMRI) experiments were carried out on eight individuals during either hearing or viewing words. The blood oxygen level dependant (BOLD) signal was taken as measure of neural activity. For each voxel, functional connectivity with all other brain voxels was determined by calculating Pearson correlation coefficients at two connectivity thresholds ( r = 0.35 and 0.65). Then, NT images were derived by counting the number of suprathreshold connections for each individual voxel. Calculations based on random networks indicate that statistically reliable NT images can be derived in individuals. With regard to group analysis, at r = 0.35 NT images are similar though not identical with the first component of principal component analysis (PCA), displaying a widespread but not ubiquitous pattern of functionally connected cortical areas. At r = 0.65, NT group images display functional connectivity confined to circumscribed cortical regions which reach beyond the corresponding primary sensory areas, their known associated areas and the default network. In conclusion, NT goes beyond the approach of correlating the BOLD signal with the external stimulus-presentation time course by computing linear functional connectivity between all brain voxels based on any BOLD time course. First results demonstrate that the NT approach is likely – on an individual base – to reveal novel cortical and subcortical connectivities involved in stimulus processing.