A system theoretic state description for temporal transitions in the electroencephalogram data of severe epileptic patients

This paper describes a new system theoretic analysis of electroencephalogram data obtained in the temporal neighborhood of a seizure episode from intracranial electrodes. The analysis provides distinct state descriptions (patient invariant characterization) of the seizure states. This characterization enables seizure detection before onset with sufficient time to warn the individual or execute actions to abort the seizure formation. Clinically classified segments of intracranial EEG data obtained from epileptic patients are separated into training and testing sets. The powers in various frequency bands are calculated as the state-discriminating features for each EEG segment. The mean power in the various frequency bands is calculated based on the training set data and a new system theoretic model, the fuzzy measure theoretic quantum approximation of an abstract system (FMQAS), is then used to develop system measures between the various frequency bands and the states. These system measures provide state descriptions of the transition into a seizure state (preictal), the seizure state (ictal), and the state immediately following a seizure (postictal). Finally, an FMQAS-based classifier is developed to individually analyze the features from short time segments of EEG. The FMQAS-based classifier shows evidence of seizure activity in advance of visual EEG manifestation and also functions well as a seizure detector for both the training and the testing datasets with low indication of seizure activity during the postictal periods where the time to a seizure onset is greater than ten minutes.