Abstracts

Rationale: Circadian and multidian probabilities of seizure occurrence have gained interest as they may offer options for the identification of proictal phases. Recently, a number of reports have suggested that such cyclical patterns of seizure occurrence contribute more to seizure prediction performance than objective measures, e.g., derived from multimodal wearable data. Information on seizure occurrence is mostly based on electronic or classical seizure diaries. Patients, however, are known to underreport seizures and to have a particular underreporting during sleep. We here assessed how well objective seizure time distribution based on simultaneous video and EEG assessment during in-hospital monitoring correlated with subjective seizure documentation as based on patient-triggered seizure alarms.

Methods: Retrospective analysis of 2003 patients aged 10-85 years undergoing video-EEG monitoring at the Epilepsy Center of Freiburg University Hospital from January 2010 to February 2022. Cyclical timing of patient alarms (button presses whenever a seizure was noted) was compared to objective timings of seizure onsets. Correspondence of patient-triggered activations with objective seizure timing was considered if it occurred within a time window of 3 minutes before to 7 minutes after seizure onset. Cyclical occurrence of subjectively noted vs. objectively documented seizures was compared using Watson’s two-sample test of homogeneity. Comparisons with circular uniform distributions were done using Watson’s goodness of fit test.

Results: Out of the total of 29631 seizures, 7313 (24.7%) were reported by button presses. Associated button press occurred in 67.2% of focal aware seizures, 26.7% of focal unaware seizures, and 22.3% of bilateral tonic-clonic seizures. The circular distributions of the hour of the day of seizure-occurrence differed significantly from a uniform distribution, demonstrating an uneven circadian distrbution for both patient-based and objective seizure documentation (p < 0.01 in either case). There was a significant difference between the circadian distributions of seizure-associated button presses and objectively documented seizures, respectively (p < 0.01). The maximum of patient-alarms was at daytime between 1 pm and 2:30 pm (circular mean: 2-3 pm), the maximal numbers of objective seizures documented was during night-time between 11 pm and 9 am (circular mean: 2-3 a.m.). Thus, the fraction of seizures missed in patient-based documentation during nighttime far exceeded that ovbserved during datyime.