Real-Time Seizure Detection Using EEG - Hyewon Jeong

Explore real-time seizure detection using EEG in this comprehensive 53-minute lecture by Hyewon Jeong from Stanford University. Compare state-of-the-art models and signal feature extractors in a practical framework suitable for real-world applications. Learn about the general pipeline of seizure detection methods, deep neural networks for EEG seizure detection, and evaluation metrics including a newly proposed one for assessing practical aspects of detection models. Discover insights on optimal sliding window lengths, feature transformers, and signal feature extractors. Examine the performance of various approaches on raw data and for different seizure types. Gain valuable knowledge on EEG lead information, the TUH EEG dataset, and the challenges of implementing seizure detection in real-time scenarios.

Intro
Diagnosis of Epileptic Seizure
General Pipeline of Seizure Detection Methods
Signal Feature Extractors for Seizure Detector
EEG Lead Information of TUH EEG
Deep Neural Networks for EEG Seizure Detection
Evaluation Metrics for Seizure Detection: EPOCH
Exploring the optimal Length of Sliding Window
Feature Transformer, Guided Feature Transformer
Exploring the optimal Signal Feature Extractors
Comprehensive Comparison on Seizure Detection: Raw Data
Binary Seizure Detection Performance for each seizure type
Real-Time Seizure Detection with EEG
Dataset