Epileptic MEG Spike Detection Using Statistical Features and Genetic Programming with KNN

Area of Science:

• Neurology
• Biomedical Engineering
• Machine Learning

Background:

• Epilepsy affects millions globally, necessitating accurate brain activity monitoring.
• Identifying seizure onset through spike detection is crucial for effective epilepsy treatment.
• Magnetoencephalography (MEG) offers high-density brain activity data, but manual analysis is challenging due to data volume.

Purpose of the Study:

• To develop and evaluate an automated method for detecting interictal spikes in MEG data.
• To explore the efficacy of genetic programming (GP) and K-nearest neighbor (KNN) for epilepsy diagnosis.
• To improve the efficiency and accuracy of analyzing large MEG datasets for neurological disorders.

Main Methods:

• A three-stage approach involving preprocessing, genetic programming-based feature generation, and classification.
• Utilized eight statistical features combined with GP for feature extraction.
• Employed the K-nearest neighbor (KNN) algorithm for classifying interictal spikes.

Main Results:

• The proposed method achieved an average sensitivity of 91.75%.
• The system demonstrated an average specificity of 92.99% in detecting epileptic spikes.
• Successful evaluation on real MEG data from 28 epilepsy patients.

Conclusions:

• The combined GP and KNN approach is effective for automated interictal spike detection in MEG data.
• This method offers a promising solution for the challenges of analyzing large MEG datasets in epilepsy diagnosis.
• The findings support the use of advanced machine learning techniques in neurological disorder monitoring.