Automated analysis and detection of epileptic seizures in video recordings using artificial intelligence
We are excited to share an impressive phase-2 study recently published in Frontiers in Neuroinformatics, an independent collaboration with the Danish Epilepsy Centre in Dianalund and Professor Sándor Beniczky. The study evaluates the algorithmic performance of Nelli in video-based detection of several types of motor seizures.
“Automated analysis and detection of epileptic seizures in video recordings using artificial intelligence” by Pragya Rai, Andrew Knight, Matias Hiillos, Csaba Kertész, Norma Elizabeth Morales Cruz, Daniella Terney, Sidsel Armand Larsen, Tim Østerkjerhuus, Jukka Peltola & Sándor Beniczky.
This study involved day and night-time video recordings from 230 patients (adults and pediatrics), with a total of 334 seizures featuring motor components included for analysis. Nelli’s performance was verified against video-EEG findings.
The wide range of seizures included in the study allowed for a sensitivity analysis of six different seizure types: Tonic-Clonic, Hyperkinetic, Tonic, Automatisms, Other motor seizures, and Psychogenic Non-Epileptic Seizures.
The results highlight Nelli as a clinically applicable solution for seizure screening within diagnostic workflows. Two key use-case scenarios for clinical application include:
- Patient Safety: Automated, real-time video monitoring in healthcare institutions
- Seizure Detection & Classification: Data reduction in diagnostic home-video monitoring
For further details, please refer to the full clinical paper here:
https://www.frontiersin.org/journals/neuroinformatics/articles/10.3389/fninf.2024.1324981/full