Abstracts

Rationale: Understanding the multi-scale, connectomic reorganization of the epileptic brain will facilitate the development of new biomarkers to help prevent and cure epilepsy. Methods: The current study applied a rat model of chronic status epileptics (SE) induced by kainic acid (KA) injection into the left, CA3 hippocampal region. Deep brain electrophysiological recordings and ex-vivo diffusion tensor imaging (DTI) were combined to assess both functional and structural abnormalities after SE. Two groups of experimental animals were investigated that: (1) developed epilepsy after SE (E+, n=8), (2) did not develop epilepsy (E-, n=6), as well as sham-injected controls (n=6). Following KA injection, bipolar electrodes were positioned at 5 locations bilaterally in the brain (peri-limbic cortex, anterior cingulate, anterior thalamus, CA1, and Dentate Gyrus) and were used to record continuous EEG for 2 months. Local field potentials (LFPs) occurring during slow wave sleep were processed to quantify the gamma event coupling (GEC), pathological high-frequency oscillations (pHFO), and multi-unit discharges (MUDs). Following EEG data acquisition, ex-vivo DTI data were acquired at 2 months post-injection using a 7Tesla spectrometer.  Data were fitted for fiber orientation distributions using constrained spherical deconvolution, and a probabilistic streamline algorithm was used to conduct whole-brain fiber tractography. A structural connectome was constructed using a co-registered, parcellated atlas with 122 gray matter regions of interest using streamline counts. Graph theory analysis (GTA) was used to for analytical inference of changes in network topography. Results: We observed pHFOs in the hippocampus and in most of the brain areas recorded within E+ animals but no pHFOs were observed out of hippocampus in E- group within the first month. We also found a decrease in EEG-analyzed functional connectivity, as indicated by suppression of GEC between ipsilateral CA1 and prefrontal cortex, and this was restricted to the E+ group of animals. The imaging data revealed significant decreases of FA values in internal capsule (p=.003), fimbria (p=.001), fornix (p=.018) and corpus callosum (p=.02) in E+ compared to E- and sham control. Although we found no significant changes in global network metrics between E+, E- and sham groups, we did find modifications at the local network level that was indicated by an obvious reduction in the number of network hubs in ipsilateral prefrontal cortex, and a paradoxical increase in hub centrality in contralateral prefrontal cortex and bilateral hippocampal-thalamic regions in E+ animals. Combining imaging and EEG data, we observed a robust association between the occurrence of pHFO, reduction of GEC and change in hubness within the ipsilateral thalamic-hippocampal-prefrontal cortex network in E+ animals. Conclusions: The convergence of the data obtained through functional and structural network analysis to ultimately reveal an altered circuit, suggesting that combining imaging and EEG methods is a feasible approach for understanding brain reorganization that occurs during epileptogenesis. Funding: NIH Grant NS033310NIH Grant NS065877