Fig. 1: Identifying information-related pathways in macroscale brain networks. | Nature Communications

Fig. 1: Identifying information-related pathways in macroscale brain networks.

From: Evidence for increased parallel information transmission in human brain networks compared to macaques and male mice

Fig. 1

a A weighted and symmetric structural connectivity matrix summarizes the white matter wiring of the brain for each species. For every pair of brain regions (i, j), the 5 shortest structural paths (light blue) connecting the two regions are identified using the k-shortest path algorithm18. b For every (human or non-human) subject, the mutual information between region pairs is computed from z-scored regional time series obtained from fMRI recordings. c By analyzing the mutual information values along each structural path, the data processing inequality (DPI) is used to assess whether the specific structural path represents an information-related pathway between regions i and j. Left panel: two brain regions i, j are connected by a structural path crossing regions x1, x2; green lines represent direct structural connections (white matter fibers). Each region is associated with a neural activity-related time series; the amount of information shared by two regions is quantified by their mutual information I (darker and thicker arcs indicate stronger I). Right panel: a structural path (i, x1, x2, j) is labeled as relay information-related pathways if the pairwise mutual information values do not increase along the (undirected) path (first row, red shading); it is not an information-related pathway otherwise (second row, gray shading: Ij,i > Ix2,i). d A parallel communication score (PCS) is computed at the individual level (i.e., for every subject n) and for every pair of brain regions i, j by counting the number of structural paths that serve as information-related pathways between the two regions. e Parallel communication scores are investigated across mammalian species, highlighting a spectrum of communication strategies from selective information transmission (light yellow; low PCS), to parallel information transmission (dark brown; high PCS). Particularly, our work highlights a parallel relay communication gap between humans and non-human mammals (macaques, male mice), with humans’ brain network communication tailored towards parallel transmission, and macaques and mice towards selective transmission. Brain schematic from scidraw.io97,98,99.

Back to article page