Subcortical (thalamic) automated seizure detection: A new option for contingent therapy delivery

Abstract

The feasibility of automated detection of cortical-onset epileptic seizures from subcortical structures such as the thalamus was investigated via simultaneous recording of electroencephalography (EEG) and anterior and centromedian thalamic nuclei electrical signals (electrothalamography) in nine subjects with pharmacoresistant seizures admitted to an epilepsy monitoring unit after deep brain stimulating electrode implantation. Thalamic electrical signals were analyzed using a validated seizure detection algorithm, and times of seizure onset and termination were compared to those determined through visual analysis of video-EEG. Ictal activity was recorded from the scalp and thalamic nuclei in three subjects who had seizures during the 3-4-day recording period. In the majority of seizures, ictal activity in the thalamic nuclei preceded electrographic onset as determined from the EEG or clinical onset as determined from behavioral observations. Interictal epileptiform discharges were also recorded from the thalamus and in certain instances had no scalp representation. Subcortical/thalamic detection of cortical-onset seizures is feasible. This approach would enable contingent therapy delivery and may be particularly valuable for subjects with multiple or difficult-to-localize epileptogenic regions.

Keywords: Computerized electrical current delivery; Computerized seizure detection; EEG; Electrothalamography; Subcortical.