NeuroGNN is a state-of-the-art framework for precise seizure detection and classification from EEG data. It employs dynamic Graph Neural Networks (GNNs) to capture intricate spatial, temporal, semantic, and taxonomic correlations between EEG electrode locations and brain regions, resulting in improved accuracy. Presented at PAKDD '24.

The official code for the paper 'Mixture of Experts for EEG-Based Seizure Subtype Classification'.

Epilepsy Ictal Phase Detection Dataset. A multimodal, non-EEG seizure detection dataset featuring acceleration and heart rate data from the Open Seizure Database (OSDB) and annotated by clinical experts to label the phases of the ictal state