Back

Beamforming seizures from the temporal lobe

Garcia Dominguez, L.; Tarazi, A.; Valiante, T.; Wennberg, R.

2021-11-24 neuroscience
10.1101/2021.11.19.469291 bioRxiv
Show abstract

BackgroundSurgical treatment of drug-resistant temporal lobe epilepsy (TLE) depends on proper identification of the seizure onset zone (SOZ), and differentiation of mesial, temporolimbic seizure onsets from temporal neocortical seizure onsets. Non-invasive source imaging using electroencephalography (EEG) and magnetoencephalography (MEG) can provide accurate information on interictal spike localization; however, EEG and MEG have low sensitivity for epileptiform activity restricted to deep temporolimbic structures. Moreover, in mesial temporal lobe epilepsy (MTLE), interictal spikes frequently arise in neocortical foci distant from the SOZ, rendering interictal spike localization potentially misleading for presurgical planning. MethodsIn this study, we used two different beamformer techniques applied to the MEG signal of ictal events acquired during EEG-MEG recordings in six patients with TLE (three neocortical, three MTLE). The ictal source localization results were compared to the patients ground truth SOZ localizations determined from intracranial EEG and/or clinical, neuroimaging and postsurgical outcome evidence. ResultsBeamformer analysis proved to be highly accurate in all cases and able to reliably identify focal seizure onsets localized to mesial, temporolimbic structures. In three patients, interictal spikes were either absent, too complex for inverse dipole modeling, or localized to anterolateral temporal neocortex distant to a mesial temporal SOZ. ConclusionsThis report demonstrates the suitability of MEG beamformer analysis of ictal events in TLE, which can supersede or complement the traditional analysis of interictal spikes. The method outlined is applicable to any type of epileptiform event, greatly expanding the information value of MEG and broadening its utility for presurgical recording in epilepsy.

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.