Seizure detection in intracranial EEG using a fuzzy inference system

Area of Science:

• Neurology
• Biomedical Engineering
• Signal Processing

Background:

• Epilepsy monitoring requires accurate seizure detection from intracranial EEG (iEEG) data.
• Manual analysis of long-term iEEG is time-consuming and prone to error.
• Automated seizure detection systems can improve patient care and research.

Purpose of the Study:

• To develop and evaluate a fuzzy rule-based system for automatic seizure detection in iEEG.
• To assess the system's sensitivity, false detection rate, and detection latency.
• To compare the automated system's performance with expert visual analysis.

Main Methods:

• A fuzzy rule-based system was developed for iEEG seizure detection.
• Temporal, spectral, and complexity features were extracted from preprocessed iEEG data.
• Spatio-temporal integration and thresholding were employed for decision making.

Main Results:

• The system demonstrated a sensitivity of 98.7%.
• A low false detection rate of 0.27/h was achieved.
• The average detection latency was 11 seconds.
• Results showed strong correlation with expert visual analysis.

Conclusions:

• The developed fuzzy rule-based system is effective for automatic seizure detection in iEEG.
• The system offers high sensitivity and a low false detection rate for long-term monitoring.
• This tool can assist clinicians and researchers in epilepsy management and studies.