Seizure tracking of epileptic EEGs using a model-driven approach

Area of Science:

• Neuroscience
• Computational Biology
• Medical Signal Processing

Background:

• Epilepsy is a chronic neurological disorder causing recurrent seizures that significantly impact patients' neuro-biologic, cognitive, and psychological well-being.
• Understanding the evolution of epileptic electroencephalography (EEG) during seizures is crucial for developing effective clinical control strategies.

Purpose of the Study:

• To develop a model-driven approach for tracking seizure development using epileptic EEG data.
• To propose an improved time-delay Wendling model with sub-populations (TD-W-SP model).
• To introduce a method for tracking seizure progression based on trained model parameters.

Main Methods:

• A new time-delay Wendling model with sub-populations (TD-W-SP model) was developed.
• A model-driven seizure tracking approach was introduced, utilizing EEG features for model training.
• A tracking index was defined based on the parameters of the trained model.

Main Results:

• The proposed method successfully simulated epileptic-like EEGs.
• The approach effectively tracked the evolution of seizures through pre-ictal, ictal, and post-ictal stages.
• Numerical results were validated on eight patients from the CHB-MIT database.

Conclusions:

• The study presents a valuable method for tracking epileptic seizures by integrating neural mass modeling (NMM) with data analysis.
• The developed TD-W-SP model and tracking approach show promise for enhancing the understanding and management of epilepsy.