A novel genetic programming approach for epileptic seizure detection.

Area of Science:

• Neuroscience and Biomedical Engineering
• Signal Processing and Machine Learning

Background:

• Epilepsy is characterized by seizures resulting from disruptions in brain activity.
• Electroencephalograph (EEG) signals reflect neuronal activity and are crucial for epilepsy diagnosis.
• Automatic detection of unpredictable epileptic seizures in EEG is essential.

Purpose of the Study:

• To develop an automated system for detecting epileptic seizures in EEG signals.
• To enhance classification accuracy using a novel genetic programming approach.

Main Methods:

• Empirical Mode Decomposition (EMD) for feature extraction from EEG signals.
• A Constructive Genetic Programming (CGP) model with integrated hill climbing and dynamic fitness computation.
• Novel operators for crossover and mutation, and a method for selecting Globally Prime offspring.

Main Results:

• The proposed CGP model achieved outstandingly high accuracies in classifying normal, interictal, and ictal EEG signals.
• Experimental results demonstrated superior performance compared to existing methods on benchmark datasets.
• The method shows significant potential for accurate and efficient epileptic seizure detection.

Conclusions:

• The developed automated system effectively detects epileptic seizures in EEG signals.
• The novel CGP approach offers a robust and computationally efficient solution for epilepsy diagnosis.
• This method holds promise for improving clinical diagnosis and patient management.