Research for suppression epilepsy using Cubature Kalman Filter and photic stimulation

Abstract

Temporal lobe epilepsy is a common neurological disease by which the common symptoms are stopping of movement and amnesia of a single memory or set of memories and tonic seizure is the leading cause of sudden unexpected death of these patients It is necessary to develop a warning system for the patient in time and the model describing average dynamic of neuron papulation called the neural mass model may assist to develop the warning system The neural mass model a non-linear model based on the electrophysiology of the neuron populations can describe the EEG of interictal and ictal In previous studies a constrained square root Cubature Kalman filter (CSC Kalman filter) was developed based on the neural mass model for predicting the occurrence of epilepsy However due to the large amount of calculations only off-line detection of seizure was achieved on a single subject The goals of this study are real-time suppress seizure by using photic stimulation and to develop new indicators based on the state of the nervous system for detection of seizure The kainic acid-treated seizure mouse was used to verify the accuracy of the new indicators Four indicators namely parameter B of the neural mass model the sample entropy the state-based inhibitory index and the excitorary index were computed for their capability to detect seizure The results show that thresholds determined based on probability distribution can improve the accuracy of seizure detection but the sensitivity In addition the average accuracy of identifying seizure using parameter B is lowest among all indicators The inhibitory state indicator with accuracy up to 98% can detect the depth EEG waveform of ictal that can not be detected by sample entropy and warn patients earlier The inhibitory index outperforms the other three indicators and it might be applied for real-time detection of seizure if the algorithm could be implemented on hardware with higher computing speed and larger memory The integration of the CSC Kalman filter and photic stimulation may have contribution to development of new method for treating temporal lobe epilepsy

Date of Award	2021
Original language	English
Supervisor	Ming-Shaung Ju (Supervisor)