Related Experiment Video
Updated: Jul 18, 2025
09:29
Continuous Video Electroencephalogram during Hypoxia-Ischemia in Neonatal Mice
Published on: June 11, 2020
3.4K
A Modified Aquila-Based Optimized XGBoost Framework for Detecting Probable Seizure Status in Neonates
Khondoker Mirazul Mumenin1, Prapti Biswas1, Md Al-Masrur Khan2
1Electronics and Communication Engineering (ECE) Discipline, Khulna University (KU), Khulna 9208, Bangladesh.
Sensors (Basel, Switzerland)
|August 26, 2023
Summary
This study introduces an optimized machine learning framework for faster, more accurate seizure detection in newborns using electroencephalography (EEG). The novel approach significantly improves diagnostic reliability for neonatal seizures.
Area of Science:
- Neurology
- Artificial Intelligence
- Biomedical Engineering
Background:
- Electroencephalography (EEG) is crucial for diagnosing pediatric neurological conditions, particularly neonatal seizures in Neonatal Intensive Care Units (NICUs).
- Traditional EEG interpretation is labor-intensive, requires specialized expertise, and can be challenging due to diverse seizure presentations.
- Machine Learning (ML) offers automated solutions for rapid and precise seizure detection.
Purpose of the Study:
- To develop and validate a novel, optimized ML framework for the automated detection of newborn seizures.
- To enhance the efficiency and robustness of seizure detection by employing a modified meta-heuristic optimization algorithm (Aquila Optimization - AO).
- To compare the performance of the proposed optimized model against established ML classifiers.
Main Methods:
- Development of an optimized ML framework incorporating the Aquila Optimization (AO) algorithm.
- Validation of the framework on a public dataset from Helsinki University Hospital, comprising EEG signals from 79 neonates.
- Comparative analysis of the proposed model against Decision Tree (DT), Random Forest (RF), and Gradient Boosting Classifier (GBC).
Main Results:
- The proposed optimized ML model achieved high performance metrics: 93.38% Accuracy, 93.9% Area Under the Curve (AUC), 92.72% F1 score, 93.38% sensitivity, and 77.52% specificity.
- The model demonstrated superior performance compared to existing shallow ML architectures, particularly in accuracy and AUC.
- The Aquila Optimization (AO) algorithm contributed to the model's efficiency and robustness in seizure detection.
Conclusions:
- The developed optimized ML framework represents a significant advancement in the reliable and automated detection of newborn seizures.
- This technology has the potential to benefit the medical community by improving diagnostic accuracy and efficiency in neonatal neurology.
- The findings support the integration of advanced AI techniques for critical care applications in pediatric neurology.
