Abstracts

Rationale: Automated and reproducible metrics for scoring relative seizure severity and post-ictal duration can be used to assess changes in patient health over time. Based on published methods we have developed software capable of computing these summary data. The performance and output of the scoring methods are demonstrated on a dataset comprised of 49 patients with epilepsy undergoing continuous EEG monitoring. These methods are proposed to evaluate the impact of acute intervention on seizure characteristics.Methods: Patient data was collected at the Medical University of South Carolina (Charleston, SC) under the approval of Western IRB. All seizure annotations were verified through blinded review by a panel of three neurologists to adjudicate scoring of electrographic onset and end times. Relative seizure severity was computed as a function of intensity based on the median power of each 10-20 EEG channel relative to the interictal period, seizure duration, and seizure spread defined by the fraction of 10-20 EEG channels involved during ictus (Osorio 2010). Post-ictal duration was estimated based upon the time required for the EEG channels to return to the 95% confidence interval of their interictal power distributions (Tilz 2006, Lhatoo 2010, Poh 2012). EEG channels that were compromised during the course of the seizure(s) or during the interictal period(s) were identified on a per patient basis and excluded from the analysis.Results: 49 patients included in the analysis had a total of 276 seizures with clear annotation of electrographic seizure onset and end time. This included 135 focal seizures with or without impairment of consciousness, 4 secondarily generalized, and 137 seizures that were not classified during the original investigator review. Scoring for relative seizure severity provided a reliable method for rank ordering the magnitude of epileptic events when multiple were recorded for a patient. Most seizures (approximately 65%) did not exhibit a prolonged post-ictal recovery period, as assessed via monitoring of electrographic characteristics. Post-ictal recovery periods, however, increased as a function of seizure duration and spread (considering only events that returned to interictal state within 5 minutes of electrographic seizure end). Seizure duration and spread are aspects of the relative seizure severity score which may be generalized among a patient population. Altering parameters used to determine post-ictal duration enables further refinement of the results.Conclusions: Scoring relative seizure severity and evaluating post-ictal duration provide quantitative methods to evaluate seizures. When treatment strategies change, these EEG-based methods can be used to evaluate changes in seizures characteristics that may be indicative of clinical improvement.