Communicability Systematically Explains Transmission Speed In A Cortical Macro-Connectome

Abstract

Global dynamics in the brain can be captured using fMRI, MEG, or electrocorticography (ECoG), but models are often restricted by anatomical constraints. Complementary single-/multi-unit recordings have described local fast temporal dynamics. However, because of anatomical constraints, global fast temporal dynamics remain incompletely understood. Therefore, we compared temporal aspects of cross-area propagations of single-unit recordings and ECoG, and investigated their anatomical bases. First, we demonstrated how both evoked and spontaneous ECoGs can accurately predict latencies of single-unit recordings. Next, we estimated the propagation velocity (1.0-1.5 m/s) from brain-wide data and found that it was fairly stable among different conscious levels. We also found that the anatomical topology strongly predicted the latencies. Finally, Communicability, a novel graph-theoretic measure, could systematically capture the balance between shorter or longer pathways. These results demonstrate that macro-connectomic perspective is essential for evaluating detailed temporal dynamics in the brain.\n\nAuthor SummaryThis study produced four main findings: First, we demonstrated that ECoG signals could predict the timing of evoked electrical spikes of neurons elicited by visual stimuli. Second, we showed that spontaneous ECoG recorded under a blindfold condition (without any stimuli) could also predict the timing of visually evoked neuronal spikes. We also clarified that performance predictions from blindfold data are essentially supported by the constraints of structural paths. Third, we quantified the propagation velocity (conductance velocity) as 1.0-1.5 m/s, and found that the velocity was stable among different conscious levels. Fourth, Communicability successfully characterized the relative contributions of shorter and longer paths. This study represents an important contribution to the theoretical understanding of the brain in terms of connectomics, dynamical propagations, and multi-scale architectures.