Related Experiment Video
Updated: Jul 11, 2025
05:58
Application of an Amplitude-integrated EEG Monitor Cerebral Function Monitor to Neonates
Published on: September 6, 2017
38.9K
Detection and Severity Identification of Neonatal Seizure Using Deep Convolutional Neural Networks from Multichannel
Biniam Seifu Debelo1, Bheema Lingaiah Thamineni2, Hanumesh Kumar Dasari3
1Department of Biomedical Engineering, Nigist Eleni Mohamed Memorial Compressive Specialized Hospital, Wachamo University, Hosanna, Ethiopia.
Pediatric Health, Medicine and Therapeutics
|November 7, 2023
Summary
A new deep convolutional neural network system accurately detects neonatal seizures from EEG data. This AI tool aids clinicians in diagnosing seizures, especially where expert neurologists are scarce.
Area of Science:
- Neurology
- Artificial Intelligence
- Medical Diagnostics
Background:
- Neonatal seizures are common neurological issues in newborns, often presenting subtle clinical signs.
- Identifying neonatal seizures through clinical observation alone is challenging and prone to errors.
- Early and accurate diagnosis is crucial for managing potential severe neurological dysfunction.
Purpose of the Study:
- To develop a deep convolutional neural network (CNN) based diagnostic system.
- To accurately determine and classify the severity of neonatal seizures using multichannel neonatal EEG data.
- To improve the diagnostic accuracy of neonatal seizures, overcoming limitations of clinical observation.
Main Methods:
- Utilized publicly available multichannel neonatal EEG datasets.
- Preprocessed 2D time series EEG data into waveform images.
- Trained and evaluated deep convolutional neural network models for seizure detection and classification.
Main Results:
- Achieved 92.6% accuracy in binary classification for seizure detection.
- Attained 88.6% accuracy in multiclassification for seizure severity.
- Demonstrated high performance metrics including F1-score, specificity, and precision for both classification tasks.
Conclusions:
- The developed CNN system effectively detects neonatal seizures.
- The system shows potential as a decision-making tool for clinicians.
- It can assist in resource-limited settings lacking expert neurologists.
