Patient non-specific algorithm for seizures detection in scalp EEG

Abstract

Epilepsy is a brain disorder that affects about 1% of the population in the world. Seizure detection is an important component in both the diagnosis of epilepsy and seizure control. In this work a patient non-specific strategy for seizure detection based on Stationary Wavelet Transform of EEG signals is developed. A new set of features is proposed based on an average process. The seizure detection consisted in finding the EEG segments with seizures and their onset and offset points. The proposed offline method was tested in scalp EEG records of 24–48h of duration of 18 epileptic patients. The method reached mean values of specificity of 99.9%, sensitivity of 87.5% and a false positive rate per hour of 0.9.