Abnormal metabolic connectivity in the pilocarpine-induced epilepsy rat model: a multiscale network analysis based on persistent homology

Abstract

Temporal lobe epilepsy is associated with dysfunctional brain networks. Here we investigated metabolic connectivity in the pilocarpine-induced epilepsy rat model and applied a new multiscale framework to the analysis of metabolic networks of small-animal brains. [(18)F]fluorodeoxyglucose PET was acquired in pilocarpine-induced chronic epilepsy rats and controls to yield interregional metabolic correlation by inter-subject manner. When interregional correlation of epilepsy rats and controls was compared directly, the epilepsy rats showed reduced connectivity involving the left amygdala and left entorhinal cortex. When regional graph properties were calculated to characterize abnormal nodes in the epileptic brain network, the epilepsy rats showed reduced nodal and local efficiencies in the left amygdala. Then, a new multiscale framework, persistent brain network homology, was used to examine metabolic connectivity with a threshold-free approach and the difference between two networks was analyzed using single linkage distances (SLDs) of all pairwise nodes. We found a tendency for longer SLDs between the left insula/left amygdala and bilateral cortical/subcortical structures in the epilepsy rats. Persistent brain network homology analysis as well as interregional correlation study implied the abnormal left limbic-paralimbic-neocortical network in the pilocarpine-induced epilepsy rat models. In conclusion, we found a globally disrupted network in the epileptic brain in rats, particularly in the limbic and paralimbic structures by direct comparison, graph properties and multiscale network analysis. These results demonstrate that the multiscale and threshold-free network analysis can be used to find the network abnormality in small-animal brains as a preclinical research.

Keywords: Epilepsy; FDG PET; Graph theory; Metabolic connectivity; Persistent homology.