Particle swarm optimization algorithm based parameters estimation and control of epileptiform spikes in a neural mass...

Area of Science:

• Computational Neuroscience
• Biomedical Engineering
• Systems Biology

Background:

• Epileptic spikes are key biomarkers indicating transitions from normal to seizure activity.
• Accurate detection and control of these transitions are crucial for epilepsy management.
• Existing methods may have limitations in parameter estimation and real-time detection.

Purpose of the Study:

• To propose a novel epilepsy detection and closed-loop control strategy.
• To utilize the Particle Swarm Optimization (PSO) algorithm for enhanced performance.
• To demonstrate the effectiveness in suppressing epileptic spikes in neural mass models.

Main Methods:

• Employing the Particle Swarm Optimization (PSO) algorithm for parameter estimation and spike detection.
• Utilizing neural mass models to simulate brain activity and epileptic events.
• Implementing a proportion-integration controller for immediate inhibition of epileptiform spikes based on a threshold.
• Comparing PSO performance with the unscented Kalman filter for parameter estimation.

Main Results:

• The proposed PSO-based strategy effectively suppresses epileptic spikes in neural mass models.
• PSO accurately estimates the time evolution of key model parameters, outperforming the unscented Kalman filter.
• Epileptiform spikes are practically detected, and immediate inhibition is achieved when the excitatory-inhibitory ratio exceeds a threshold.
• Numerical simulations validate the method's effectiveness for early seizure detection and closed-loop control.

Conclusions:

• The PSO-based strategy offers a promising approach for model-based early epilepsy detection.
• This method demonstrates significant potential for developing effective closed-loop seizure control treatments.
• The enhanced parameter estimation capabilities of PSO are valuable for understanding and managing epilepsy.