Status epilepticus (SE) is a severe, protracted seizure often resistant to antiseizure medications. Its pathophysiology is unclear, with seizures potentially arising from various brain regions. SE is frequently accompanied with peri-ictal MRI abnormalities (PMA), which could reveal new insights into its implicated brain network and lead to new therapeutic targets.

We identified 73 published cases of SE-related diffusion restriction and used lesion network mapping to identify a common brain network. Locations of diffusion restriction from ischemic stroke (n=490) served as controls. We identified an SE network using a voxel-wise t-test with 10,000 permutations and tested whether lesion overlap with this SE network could predict SE in an independent prospective cohort. The validation cohort consisted of 47 patients with SE-related diffusion restriction versus 120 of another etiology (stroke, tumor, infection etc.). To contextualize our findings, we compared the locations of diffusion restriction and  SE network to maps of metabolic demand (FDG-PET) and distributions of gene expression derived from the Allen Human Brain Atlas. 

SE-related PMAs were connected to the medial pulvinar thalamus and precuneus cortex (>85% sensitive, P_FWE < 0.05 specific). We derived an SE network using these regions, which showed functional connectivity to the limbic system, default mode network, septum, claustrum, tectum, and cerebellum. Lesion overlap with the identified SE network predicted SE in an independent cohort (AUC = 0.883, p < 0.001), improving significantly (D = 2.931, df = 306.86, p = 0.004) over an FDG-PET map of metabolic demand (AUC=0.703, p < 0.001). The SE network spatially correlated with gene distributions for HOMER2 and metabotropic glutamate receptors 5 and 7.

We identified an SE network that discriminates SE-related diffusion restriction from other etiologies and enhances our understanding of seizure circuits, which could lead to novel treatment targets for ongoing seizures.