Active fault tolerant deep brain stimulator for epilepsy using deep neural network.

Area of Science:

• Neuroscience
• Biomedical Engineering
• Computational Neuroscience

Background:

• Epileptic seizures affect millions globally, with severe cases often managed by deep brain stimulation.
• Controlling severe epileptic seizures remains a challenge, particularly when ion channel conductance varies.

Purpose of the Study:

• To propose a Hodgkin-Huxley (HH) model-based Active Fault Tolerant Deep Brain Stimulator (AFTDBS) for controlling epileptic seizures.
• To suppress epileptic seizures by addressing ion channel conductance variations using a Deep Neural Network (DNN).

Main Methods:

• Utilized Support Vector Machine (SVM) and K-Nearest Neighbor (KNN) for seizure detection.
• Employed Long Short-Term Memory (LSTM) to predict ion channel conductance variations.
• Developed a Reconfigurable Deep Brain Stimulator (RDBS) with Proportional Integral (PI) and Model Predictive Controllers (MPC).

Main Results:

• Classified seizures into normal and epileptic states based on sodium and potassium ion channel conductance variations.
• Designed current-controlled deep brain stimulators for effective epileptic suppression.
• Demonstrated the efficacy and stability of the proposed DNN-based AFTDBS through closed-loop performance analysis.

Conclusions:

• The proposed DNN-based AFTDBS effectively suppresses epileptic seizures by adapting to ion channel conductance changes.
• The AFTDBS offers a promising approach for controlling severe epileptic seizures, enhancing patient outcomes.
• The study validates the robustness and stability of the developed control schemes in managing neurological disorders.