Epileptic seizure detection based on expected activity measurement and Neural Network classification

Area of Science:

• Neurology
• Biomedical Engineering
• Signal Processing

Background:

• Epilepsy is a common neurological disorder, frequently requiring specialist consultation.
• Electroencephalograms (EEGs) are crucial for diagnosing and monitoring epilepsy.
• Accurate and automated seizure detection in EEGs remains a significant challenge.

Purpose of the Study:

• To develop and validate a new automated algorithm for detecting epileptic seizures in EEG signals.
• To improve the efficiency and accuracy of epilepsy diagnosis through advanced signal processing techniques.

Main Methods:

• The proposed method utilizes image transformation and Discrete Wavelet Decomposition (DWT) for EEG analysis.
• Feature extraction is performed using Expected Activity Measurement (EAM) on wavelet coefficients.
• A backpropagation Artificial Neural Network (ANN) classifies EEG segments into seizure and non-seizure periods.

Main Results:

• The algorithm demonstrated high performance on a public EEG dataset.
• The automated detection system achieved an accuracy of up to 99.44% in identifying epileptic seizures.
• The correlation between wavelet coefficients across sub-bands proved effective for feature extraction.

Conclusions:

• The developed algorithm offers a promising, highly accurate, and automated approach for epileptic seizure detection in EEGs.
• This method has the potential to significantly aid neurophysiologists in epilepsy diagnosis and management.
• The integration of DWT, EAM, and ANN provides a robust framework for EEG-based seizure detection.