Deep Learning Approach to Neonatal Seizure Detection Using Hybrid Conv-Bi-LSTM Model

Overview

Neonatal seizures cause significant morbidity and mortality, both acutely and in the long term, contributing to adverse neurodevelopmental outcomes. Traditional EEG seizure detection by human experts is constrained by limited efficiency, scalability, and objectivity, which can lead to delayed diagnosis and hinder optimal outcomes. Deep learning (DL) methods have shown promise in neonatal seizure detection, however, more recent DL architectures have not been robustly evaluated. Here, we evaluated a recent DL architecture, Conv-LSTM, in neonatal seizure detection in 20 infants from a publicly available dataset. The data were pre-processed to remove artifacts, segmented into 1-second epochs, then quantitative EEG features were extracted, including principal component analysis performed on entropy-based and wavelet decomposition features. In the classification of seizure epochs from non-seizure epochs, Conv-LSTM achieved an accuracy of 86% and an area under the curve (AUC) score of 0.91. This study demonstrates the potential of Conv-Bi-LSTMs in accurately detecting neonatal seizures based on EEG data, showing promise for reducing the mortality rates and negative long-term impacts on susceptible neonates.

Competition history

• AJAS 2026 Category not listed

Related projects

Closest projects by meaning, across every fair and year in the corpus.

Browse more like this

• All Computational Biology and Bioinformatics projects
• AJAS Computational Biology and Bioinformatics projects
• Every AJAS 2026 project