Back

Brain lesion locations associated with secondary seizure generalization

Nordberg, J.; Schaper, F. L.; Bucci, M.; Nummenmaa, L.; Joutsa, J.

2021-04-22 neurology
10.1101/2021.04.19.21255414 medRxiv
Show abstract

BackgroundStructural brain lesions are the most common cause of adult-onset epilepsy. The lesion location may contribute to the risk for epileptogenesis, but whether specific lesion locations are associated with a risk for secondary seizure generalization from focal to bilateral tonic-clonic seizures, is unknown. MethodsWe identified patients with a diagnosis of adult-onset epilepsy caused by an ischemic stroke or a tumor diagnosed at the Turku University Hospital in 2004-2017. Lesion locations were segmented on patient-specific MR imaging and transformed to a common brain atlas (MNI space). Both region-of-interest analyses (intersection with the cortex, hemisphere, and lobes) and voxel-wise analyses were conducted to identify the lesion locations associated with focal to bilateral tonic-clonic compared to focal seizures. ResultsWe included 170 patients with lesion-induced epilepsy (94 tumors, 76 strokes). Lesions predominantly localized in the cerebral cortex (OR 2.50, 95% C.I. 1.21-5.15, p = 0.01) and right hemisphere (OR 2.22, 95% C.I. 1.17-4.20, p = 0.01) were independently associated with focal to bilateral tonic-clonic seizures. At the lobar-level, focal to bilateral tonic-clonic seizures were associated with lesions in the right frontal cortex (OR 4.41, 95% C.I. 1.44-13.5, p = 0.009). No single voxels were significantly associated with seizure type. These effects were independent of lesion etiology. ConclusionsOur results demonstrate that lesion location is associated with the risk for secondary generalization of epileptic seizures independent of lesion etiology. These findings may contribute to identifying patients at risk for focal to bilateral tonic-clonic seizures.

Matching journals

The top 4 journals account for 50% of the predicted probability mass.

50% of probability mass above

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.