Abstract
Sudden unexpected death in epilepsy (SUDEP) affects more than 3000 individuals annually, yet objective and scalable biomarkers to assess risk remain limited. Postictal generalized electroencephalogram suppression (PGES) has been proposed as a potential biomarker, but its quantification is often subjective and variable. Here, we developed and validated an automated Cumulative Sum (CuSUM)-based algorithm to objectively quantify PGES duration in traumatic brain injury (TBI) mouse models of epilepsy. The algorithm was tested across three cohorts: 1 mm TBI, 2 mm TBI, and 2 mm TBI with HDAC inhibitor suberoylanilide hydroxamic acid (SAHA) treatment. Although SUDEP occurred in only a subset of animals in both TBI groups, injury severity influenced the timing of SUDEP onset. More severe injury was associated with earlier SUDEP events and shorter PGES durations, whereas less severe injury was characterized by longer PGES durations (1 mm TBI: ~110 s; 2 mm TBI: ~40 s) and delayed SUDEP onset. SAHA treatment further reduced PGES duration (~30 s), suggesting potential translational relevance for SUDEP. Bland-Altman analysis demonstrated strong agreement between automated and expert measurements. Together, these findings demonstrate that automated PGES quantification provides a reproducible and objective framework for assessing postictal EEG dynamics and their relationship to SUDEP timing in epilepsy models. This approach offered a scalable tool for mechanistic studies and treatment evaluation, supporting future efforts toward multimodal and clinically translatable SUDEP research. PLAIN LANGUAGE SUMMARY: This study developed a simple, automated method to measure postictal generalized electroencephalogram suppression (PGES) in mouse models of brain injury to estimate the risk of SUDEP. It matched expert ratings and showed that injury severity, PGES length, and SUDEP risk relate in unexpected ways. While an epigenetic inhibitor drug shortened PGES, further studies are needed to validate its efficacy in epilepsy models.