Abstract

Previous work employing graph theory and nonlinear analysis has found increased spatial and temporal disorder, respectively, of functional brain connectivity in schizophrenia. We present a new method combining graph theory and nonlinear techniques that measures the temporal disorder of functional brain connections. Multichannel electroencephalographic data were windowed and functional networks were reconstructed using the minimum spanning trees of correlation matrices. Using a method based on Shannon entropy, we found elevated connection entropy in gamma activity of patients with schizophrenia; however, gamma connection entropy remained elevated in patients with schizophrenia even after a reduction in symptoms due to treatment with antipsychotics. Our results are consistent with several possibilities: aberrant functional connectivity is epiphenomenal to schizophrenia, aberrant functional connectivity is a central feature but antipsychotics reduce symptoms by an independent mechanism, or connection entropy is not an appropriately sensitive measure of brain abnormalities in schizophrenia