Amplitude integrated electroencephalography for neonatal seizure detection: an alternative opinion to the 2025 Cochrane review.

Abstract

Abstract: Amplitude-integrated electroencephalography (aEEG) is widely used in neonatal intensive care units for bedside neuromonitoring, seizure detection, prognostication, and guidance of therapeutic hypothermia. A recent Cochrane systematic review comparing aEEG with conventional video-electroencephalography (cEEG) questioned the value of aEEG for neonatal seizure detection. However, careful interpretation of these findings is warranted. This commentary highlights key methodological limitations of the review, including the inclusion of studies without raw EEG display, heterogeneity in electrode configurations, inconsistent reporting of interpreter expertise, and variable seizure definitions. These factors likely bias pooled estimates toward underestimating the performance of modern aEEG systems when used according to contemporary standards. We emphasize that aEEG should not be evaluated as a substitute for cEEG, but as a complementary bedside modality. Beyond seizure detection, aEEG provides clinically meaningful information through background assessment, early prognostication in hypoxic-ischemic encephalopathy, and longitudinal trend monitoring, particularly in settings where cEEG is unavailable. We conclude that dismissing aEEG technology based on methodologically limited comparisons risks undervaluing an important and pragmatic neuromonitoring tool. Future research and guidelines should prioritize optimized application, standardized protocols, training, and integration with emerging analytic approaches, including artificial intelligence, to support equitable neonatal neurocritical care worldwide. Impact: aEEG has been widely used in Neonatal Intensive Care Units (NICUs) for bedside neuromonitoring of critically ill newborns, focusing on seizure detection, prognostication, and eligibility for therapeutic hypothermia. A Cochrane Review published in August 2025 compared aEEG to conventional video-electroencephalography and questioned the value of aEEG for seizure detection. Numerous methodologic limitations were present in the review, including the use of studies without a raw EEG signal display, heterogeneity in electrode configurations, inconsistent interpreter expertise, and different seizure definitions. In sensitivity analyses presented in the Cochrane review, studies that incorporated raw EEG tracings showed higher sensitivity and specificity for seizure detection compared to those relying solely on the compressed aEEG. When used with raw EEG signal display, standardized protocols, and trained readers, aEEG provides reliable and clinically meaningful information and is critical in settings where cEEG is not available. Future studies should optimize its application and integrate aEEG into multimodal neuroprotection strategies.Copyright © The Author(s) 2026.