Local field potential-based brain-machine interface to inhibit epileptic seizures by spinal cord electrical...

Area of Science:

• Neuroscience
• Biomedical Engineering
• Machine Learning

Background:

• Epileptic seizures pose a significant challenge, necessitating novel therapeutic strategies.
• Current treatments for epilepsy have limitations, driving research into alternative interventions.
• Brain-machine interfaces (BMIs) offer a promising avenue for real-time seizure detection and management.

Purpose of the Study:

• To propose and evaluate a closed-loop brain-machine interface (BMI) for inhibiting epileptic seizures.
• To utilize spinal cord stimulation (SCS) as a therapeutic intervention triggered by a seizure detection system.
• To apply a semi-supervised machine learning approach for early seizure detection based on Local Field Potential (LFP) patterns.

Main Methods:

• Acquisition of LFP signals from the hippocampus and motor cortex in Wistar rats.
• Band-pass filtering of LFP signals (1-13 Hz) and time-frequency analysis using Continuous Wavelet Transform.
• Novel Z-score-based Phase Lock Value (PLV) normalization with modified k-means and Davies-Bouldin clustering for seizure detection.
• Closed-loop system triggering 30-second SCS upon seizure event detection.

Main Results:

• The proposed BMI system significantly reduced seizure intensity and duration compared to control conditions.
• The system accurately detected synchronized seizure activity in the hippocampus and motor cortex.
• Spinal cord stimulation effectively inhibited seizure symptoms, demonstrating the system's therapeutic potential.

Conclusions:

• Low-frequency LFP signals from the hippocampus and motor cortex are valuable for seizure detection.
• Closed-loop BMIs integrating SCS show promise as an alternative treatment for epilepsy.
• The developed semi-supervised machine learning approach enables accurate and timely seizure inhibition.