Codificação eficiente para caracterização de eletroencefalograma de pacientes epiléticos

Abstract

Epilepsy is a neurological condition characterized by recurrent seizures, caused by brief disturbances in the electrical functions of the brain. Currently seizures are treated via anticonvulsant drugs that improve the quality of life of epileptic patients, controlling seizures with as few side effects as possible. It is possible to distinguish four stages in the electroencephalogram signal of a patient with the disease: interictal, period without seizure activity; ictal, corresponds to the seizure itself; pre-ictal period before the seizure; and post-ictal period, which precedes the seizure. Studies suggest that in some types of epilepsy it is possible to predict a seizure by analyzing the signs of the pre-ictal period, because at this stage the so-called "aura" appears as a warning of the crisis. This work proposes a tool for the characterization of convulsive seizures in epileptic patients through the Electroencephalogram, using efficient coding. Signal processing was performed using the Efficient Coding technique to obtain the efficient codes related to two EEG signal periods of epileptic patients: interictal and pre-ictal. We used 21 hours of EEG from two patients in order to find out if the efficient coding would be able to find differences between the signals of the two stages studied. The results show that efficient coding is able to characterize and differentiate the interictal and pre-ictal stages of the Electroencephalogram of an epileptic patient from a given frequency range.