An EEG dataset for interictal epileptiform discharge with spatial distribution information

Area of Science:

• Neuroscience
• Medical Informatics
• Biomedical Engineering

Background:

• Interictal epileptiform discharges (IEDs) and their spatial distribution are crucial for epilepsy management.
• Existing datasets lack sufficient data and spatial information, hindering accurate diagnosis and treatment.
• There is a need for comprehensive, annotated EEG datasets to advance epilepsy research.

Purpose of the Study:

• To present a novel, large-scale, and meticulously annotated EEG dataset for epilepsy research.
• To provide detailed spatial distribution and state of consciousness information for IEDs.
• To facilitate the development and validation of improved IED detection models.

Main Methods:

• Collected 28 hours of continuous raw EEG recordings from 84 epilepsy patients.
• Annotated 2,516 interictal epileptiform discharge (IED) epochs and 22,933 non-IED epochs (4 seconds each) by multiple EEG experts.
• Categorized IEDs based on occurrence regions (generalized, frontal, temporal, occipital, centro-parietal) and annotated states of consciousness (wake/sleep).

Main Results:

• Developed and validated a VGG-based model for IED detection, demonstrating improved performance with spatial and consciousness data.
• The dataset contains 2,516 annotated IED epochs and 22,933 non-IED epochs.
• The inclusion of consciousness and spatial distribution information enhanced IED detection model performance.

Conclusions:

• The presented comprehensive EEG dataset addresses limitations of existing resources for epilepsy research.
• This dataset enables more accurate IED detection and classification, aiding epilepsy diagnosis and treatment.
• It serves as a valuable benchmark for evaluating and comparing IED detection algorithms.