Epileptic seizure detection in EEG recordings using phase congruency

Area of Science:

• Biomedical Engineering
• Signal Processing
• Neurology

Background:

• Epileptic seizures are neurological disorders characterized by abnormal brain activity.
• Accurate and robust detection of epileptic seizures from electroencephalogram (EEG) signals is crucial for diagnosis and treatment.
• Existing methods often rely on magnitude-based features, which can be sensitive to noise and variations.

Purpose of the Study:

• To present a novel method for robust epileptic seizure detection using phase congruency.
• To evaluate the effectiveness of phase congruency as a feature for classifying EEG data into epileptic and seizure-free classes.
• To compare the performance of the proposed phase congruency method with existing magnitude-based detection techniques.

Main Methods:

• Phase congruency was calculated using log Gabor wavelets applied to EEG data.
• The number of spikes detected from the phase congruency was extracted as a one-dimensional feature.
• Two classes of EEG data, epilepsy and seizure-free, were analyzed.

Main Results:

• The proposed phase congruency method demonstrated robust seizure detection capabilities.
• The one-dimensional features derived from phase congruency yielded high detection accuracies.
• The performance was comparable to established magnitude-based EEG seizure detection methods.

Conclusions:

• Phase congruency is a viable and effective technique for robust epileptic seizure detection.
• The use of one-dimensional features from phase congruency offers a computationally efficient approach.
• This method shows promise for improving the accuracy and reliability of automated seizure detection systems.