The epileptic seizures result due to the nonlinear dynamic asynchronization of the brain waves. The temporal dynamics of the brain waves are usually depicted from long-term EEG signals. Prediction of seizures craves f...
详细信息
The epileptic seizures result due to the nonlinear dynamic asynchronization of the brain waves. The temporal dynamics of the brain waves are usually depicted from long-term EEG signals. Prediction of seizures craves for the capacity to reliably recognize the distinct preictal state comprehended from the remaining seizure states. The differentiation between the preictal and interictal states is the key challenge in such a prediction problem, which intently requires the least number of false alarms. computational models trace the seizure dynamics and descend the segment changeover of the seizure states from interictal toward preictal, preictal toward ictal and ictal toward postictal. In this work, a model driven-computational approach for interpreting the nonlinear temporal dynamics of the four different seizure states is developed. An Augmented Wendling Model for generating type 1 to 10 seizure-like EEG waveforms is established. Tracking and prediction of artificially generated seizure-like EEG waveforms are done via an Optimized Model Driven approach using Grey Wolf optimizer. Then, the seizure-like EEG waveforms are controlled, using an fuzzy-based neural computational modelling, comprising of Optimized Forward Back Propagation-based Closed Loop Controller (OFBP-CLC) is implemented. Ordered fuzzy Number System (OFNS) is applied on to the neurons of the OFBP-CLC approach to reduce the complexity of the learning approach. The pattern driven method in concurrence with a computational model is established. Epileptic seizure states are modelled, tracked, predicted and controlled using an arithmetic ordered fuzzy number system.
暂无评论