Using Constrained Square-Root Cubature Kalman Filter for Quantifying the Severity of Epileptic Activities in Mice

Area of Science:

• Neuroscience
• Biomedical Engineering
• Signal Processing

Background:

• Quantifying seizure severity during electrical stimulation is crucial for epilepsy treatment and evaluation.
• Existing methods for analyzing electroencephalogram (EEG) data may not accurately reflect neural network excitability.

Purpose of the Study:

• To develop and evaluate a novel parameter ratio derived from constrained square-root cubature Kalman filter (CSCKF) for quantifying neural network excitability.
• To compare the effectiveness of the CSCKF parameter ratio against band power, Teager energy operator, and sample entropy in assessing epileptiform discharge severity.

Main Methods:

• Generated simulated EEG data (normal and epileptic) using a mathematical model.
• Recorded experimental EEG data from mice undergoing electrical stimulation.
• Collected EEG segments before, during, and after stimulation for analysis.

Main Results:

• Experimental data showed inconsistent results among the four indicators, though trends in seizure severity were observed.
• Simulated data analysis revealed that only the CSCKF parameter ratio accurately reflected the model network's excitability.
• CSCKF parameter ratio demonstrated superior performance in capturing neural network excitability, particularly when the DC component was preserved.

Conclusions:

• CSCKF did not outperform other indicators for standard EEG analysis.
• CSCKF shows promise for quantifying neural network excitability and seizure severity in EEG data with a preserved DC component.