Detection of neonatal asphyxia by analyzing the complexity of electroencephalography data

Introduction In neonates, the early detection of asphyxia improves survival rates and prevents long-term complications. In neonatal care, physiological signals, including heart rate and oxygen saturation, are routinely monitored. However, neonates with neurological conditions such as hypoxic-ischemic encephalopathy (HIE) require direct neural monitoring. Electroencephalography (EEG) is a non-invasive method for assessing neural activity and therefore can effectively detect early signs of asphyxia. Although studies on HIE have utilized clinical-grade EEG systems, the real-world application of wearable EEG devices in broader neonatal care remains underexplored. In this study, we aimed to investigate the effectiveness of wearable EEG devices in detecting asphyxia without restricting its progression to hypoxic-ischemic encephalopathy (HIE). Methods We used Fuzzy Entropy (FuzzyEn) to perform power spectral and complexity analyses on EEG signal data healthy neonates and those with asphyxia. Results We found that both delta band power and EEG signal complexity decrease in neonates with asphyxia, which is consistent with those of studies on HIE. Furthermore, FuzzyEn in combination with absolute power measurements captured complementary information that led to improved detection accuracy and enhanced identification performance. Discussion Wearable EEG devices are scalable and accessible for use in resource-constrained environments (such as rural and developing regions) and can be integrated into Internet of Things (IoT) systems. Our findings highlight the potential of wearable EEG devices in early detection of asphyxia, which may contribute to a more effective neonatal care and improved survival outcomes.