Epileptic spike recognition in electroencephalogram using deterministic finite automata.

Area of Science:

• Neuroscience
• Biomedical Engineering
• Signal Processing

Background:

• Epileptic seizures are characterized by abnormal brain activity detected via EEG.
• Accurate and automated detection of epileptic spikes is crucial for diagnosis and management.
• Existing methods may require significant training data or human interpretation.

Purpose of the Study:

• To develop and evaluate an automated method for epileptic spike detection in EEG signals.
• To utilize Deterministic Finite Automata (DFA) for modeling the detection system.
• To assess the system's performance without requiring human input or prior training.

Main Methods:

• EEG signals were recorded at 256 Hz and preprocessed to remove baseline shifts.
• An Infinite Impulse Response (IIR) Butterworth band-pass filter was applied.
• A system modeled with Deterministic Finite Automata (DFA) was designed for spike detection.

Main Results:

• The automated system achieved an average recognition rate of 95.68% for epileptic spikes.
• The system processed single-channel EEG data files automatically.
• No human intrusion or training data was necessary for the system's operation.

Conclusions:

• Deterministic Finite Automata (DFA) offer a viable approach for automated epileptic spike detection in EEG.
• The developed system demonstrates high accuracy and efficiency.
• This automated method shows potential for extension to continuous EEG data processing.