Related Experiment Video
Updated: Sep 18, 2025
Application of an Amplitude-integrated EEG Monitor Cerebral Function Monitor to Neonates
Published on: September 6, 2017
Detection of pre-ictal epileptic events using a self-attention based neural network from raw Neonatal EEG data
Kondaveeti Tejaswi1, Madala Vikas1, Himala Praharsha1
1Indian Institute of Science Education and Research, Thiruvananthapuram, India.
Insights
This study introduces an AI model for early seizure prediction in newborns using EEG data. The self-attention network achieves 91% accuracy, offering a promising tool for neonatal epilepsy management.
Area of Science:
- * Neuroscience and Biomedical Engineering
- * Artificial Intelligence in Healthcare
Background:
- * Neonatal seizures are subtle, difficult to detect, and require timely intervention.
- * Automated early prediction systems are crucial for improving neonatal outcomes.
- * Machine learning for seizure detection is established, but preemptive warning systems are less explored.
Purpose of the Study:
- * To develop and validate a self-attention-based neural network for early seizure prediction in neonates.
- * To address computational challenges in analyzing long-duration EEG data using a novel embedding technique.
- * To demonstrate the potential of AI for real-world clinical application in neonatal epilepsy.
Main Methods:
- * Development of a self-attention neural network processing raw electroencephalogram (EEG) data.
- * Introduction of a second-wise summary statistics-based embedding to reduce input sequence length and retain features.
- * Validation on a public dataset of 79 neonatal patients with physician-annotated EEG recordings.
Main Results:
- * The model achieved a maximum accuracy of 91% in distinguishing pre-ictal and ictal events from non-ictal signals.
- * Evaluation on unseen patients demonstrated strong potential for real-world applicability.
- * The embedding method effectively reduced computational burden while preserving essential EEG features.
Conclusions:
- * The study presents a proof-of-concept for a preemptive seizure warning system using AI.
- * The developed model shows significant promise for enhancing neonatal seizure detection and management.
- * This work paves the way for AI-driven epilepsy management and broader clinical applications in seizure prediction.
Abstract:
Epileptic seizures can occur unpredictably, making real-time monitoring and early warning systems critical, especially in neonatal patients, where timely intervention can significantly improve outcomes. Neonatal seizures are often subtle and difficult to detect, increasing the need for automated, early prediction methods to aid clinical decision-making. While machine learning models have been widely used for seizure detection, their application in preemptive seizure warning remains underexplored. In this study, we propose a self-attention-based neural network that processes raw EEG data to detect pre-ictal signals, enabling early seizure prediction. A key challenge in using attention mechanisms for EEG analysis is the computational burden of handling high-frequency, long-duration signals. To address this, we introduce a second-wise summary statistics-based embedding that significantly reduces the input sequence length while retaining essential features. We validate our model using a publicly available dataset of 79 neonatal patients with physician-annotated EEG recordings. Our classifier achieves a maximum accuracy of 91 percent in distinguishing pre-ictal and ictal events from non-ictal signals. Notably, we evaluate our model on completely unseen patients, demonstrating its potential for real-world applicability in neonatal seizure prediction. This study provides a proof-of-concept for a preemptive seizure warning system, paving the way for AI-driven neonatal epilepsy management and broader clinical applications in seizure detection.
