Seeing Seizures: A Large-Scale Empirical Study of Video and EEG Deep Learning for Automated Rodent Seizure Detection and Severity Grading

Key points

• Manual annotation of synchronized video and EEG recordings is a bottleneck in preclinical seizure research.1
• Ten video architectures and seven single-channel EEG models are evaluated independently on binary detection and three- and five-class severity tasks.1
• A separate video-only binary-detection evaluation on RodEpil achieves a mean seizure-class F1 of 0.979 after uniform re-encoding to reduce identified class-correlated encoding cues.1
• The study characterizes existing models for behavior-defined severity labels under the evaluated protocol, motivating further investigation of expert-assisted annotation workflows.1

Sources (1)

• [1]Seeing Seizures: A Large-Scale Empirical Study of Video and EEG Deep Learning for Automated Rodent Seizure Detection and Severity Grading
  arXiv (AI, ML, NLP, CV, robotics, multi-agent) · Oct 15, 07:11 PM

  We present an empirical study of seizure detection and behavioral severity grading using 24,131 paired clips from 601 sessions across 20 rats in a soman-induced epilepsy model.

  Manual annotation of synchronized video and EEG recordings is a bottleneck in preclinical seizure research.

Extractive summary: sentences quoted from the sources.