Automated Detection of Aberrant Episodes in Epileptic Conditions: Leveraging EEG and Machine Learning Algorithms.

Area of Science:

• Neurology
• Biomedical Engineering
• Data Science

Background:

• Epilepsy is a neurological disorder requiring precise seizure detection for effective treatment.
• Electroencephalography (EEG) is the standard for monitoring brain electrical activity.
• Current detection methods can be improved with advanced computational techniques.

Purpose of the Study:

• To develop and evaluate machine learning models for precise epileptic seizure detection using EEG data.
• To investigate the utility of the Hurst exponent for identifying long-term EEG signal characteristics related to epilepsy.
• To assess the feasibility of a computationally efficient epilepsy detection framework for potential edge hardware implementation.

Main Methods:

• Feature extraction using the Hurst exponent and Daubechies 4 discrete wavelet transform.
• Feature selection employing ANOVA test and random forest regression.
• Classification using Support Vector Machine, Random Forest Classifier, and Long Short-Term Memory network models.
• Validation on the CHB-MIT scalp EEG database.

Main Results:

• The random forest classifier achieved the highest performance with 97% accuracy and 97.20% sensitivity.
• The proposed method demonstrated effective generalization on unobserved data.
• The approach utilizes single-channel EEG with minimal handcrafted features.

Conclusions:

• The developed machine learning framework offers a computationally efficient and accurate method for epilepsy seizure detection.
• This technique has the potential to enhance individualized epilepsy treatment and improve patient outcomes.
• The model's efficiency makes it suitable for deployment on edge devices for real-time analysis.