Automatic detector of high frequency oscillations for human recordings with macroelectrodes
1Department of Biomedical Engineering at McGill University and the Montreal Neurological Institute, Montreal, QC, H3A 2B4, Canada. rina.zelmann@mail.mcgill.ca
Summary
A new automatic detector for High Frequency Oscillations (HFOs) in EEG has been developed. This tool accurately identifies HFOs, a key biomarker for epileptogenic tissue, improving diagnostic efficiency.
Area of Science:
- Neuroscience
- Biomedical Engineering
- Signal Processing
Background:
- High Frequency Oscillations (HFOs) detected in electroencephalography (EEG) are significant biomarkers for identifying brain regions affected by epilepsy.
- Manual detection of HFOs is laborious and prone to inter-rater variability, necessitating automated solutions.
Purpose of the Study:
- To develop and validate a novel automated detector for High Frequency Oscillations (HFOs) in EEG signals.
- To improve the efficiency and objectivity of HFO detection for potential clinical use.
Main Methods:
- A new automatic HFO detector was designed, integrating previously identified baseline information.
- The detector was trained on data from 72 EEG channels and validated on 278 channels.
Main Results:
- The automated detector achieved a high mean sensitivity of 96.8%.
- A low mean false positive rate of 4.86% was recorded, using visually confirmed negative segments for validation.
Conclusions:
- The developed automatic HFO detector demonstrates high accuracy and efficiency.
- This tool holds promise for systematic HFO research and future clinical applications in epilepsy diagnosis.
