Control Theory Measures for Dynamic Analyses of Human Functional Connectome Data

Control theory represents a new frontier in the study of the human connectome as it offers the possibility of introducing new explainable features with the potential of binding neural connectivity to its pathological alterations. Starting from resting-state functional MRI (rs-fMRI) data sets, we estimate the functional connectivity for a given brain parcellation, thus deriving the system matrix that can be studied with network control theory. We assess the energy efficiency of large-scale brain network dynamic reconfigurations occurring at rest as well as the extent to which the transitions between and across distinct dynamic states are not just random but preferentially follow energetically efficient trajectories. In addition, we derive novel dynamic features of system controllability, distinguishing between average and modal controllability. The application of these concepts is illustrated on real rs-fMRI data sets acquired in four adult subjects, highlighting the range of variability of these features, in relation to the analysis settings, and discussing their potentials for possible use in future studies.