Multimodal Spectral Metrics with Discrete Finite Automata for Predicting Epileptic Seizures

Abstract

This Paper presents a multimodal spectral metrics and Deterministic Finite Automata (DFA) in a manner to enhance the detection of spikes and seizures in epileptiform activity from Electroencephalograms (EEG). To develop robust classification rules for identifying epileptiform activity in the human brain the authors present a new methodology to connect their previous work where successful prediction of epileptiform activity was achieved through the use of DFA. The link between DFA seizure detection operating in the time-domain and frequency domain seizure detection has been a non-trivial task and this paper presents a means to link four power spectra metrics of rat EEG experiencing epilepsy seizures and previous work by the authors where the same rats experience seizures that the authors DFA algorithm identified the seizures. We propose a system that links 1) four power spectra metrics capable of detecting seizure activity with 2) Deterministic Finite Automata (DFA). It is a common goal for those skilled in the art of epilepsy prediction to create classifiers that are used to make rules and isolate characteristic events leading to an epileptic seizure. Herein, we present a means to link time and frequency domains using various spectral metrics and DFA to identify the electrographic onset of a seizure.