Comparison of epileptic seizure prediction performance for different EEG derivation schemes

Authors

  • O. Yu. Panichev National Technical University of Ukraine "Igor Sikorsky Kyiv Polytechnic Institute", Ukraine
  • A. O. Popov National Technical University of Ukraine "Igor Sikorsky Kyiv Polytechnic Institute", Ukraine
  • V. I. Kharytonov TMO "Psychiatry", Kyiv, Ukraine

DOI:

https://doi.org/10.20535/RADAP.2017.68.54-58

Keywords:

epilepsy, epileptic seizure, correlation, electroencephalography, EEG, seizure prediction, derivation scheme

Abstract

Introduction. The influence of EEG derivation scheme on performance of epileptic seizure prediction is considered in the paper. Comparison of epileptic seizure prediction performance for different EEG derivation schemes is conducted. Database of EEG recordings from 20 patients (between 1 and 25 years old) suffering epilepsy was used. Correlation coefficients between channels of signal extracted with different combination of window and preictal lengths were used as features. Support vector machine was used to classify EEG data into interictal and preictal classes. Epileptic seizures prediction performance was evaluated using area under the ROC-curve for two types of derivation schemes: monopolar scheme with reference electrode on ipsilateral ear and scheme with averaged reference. Conclusions. It was found that there is a difference in epileptic seizures prediction performance for different schemes and patients. For subgroup of patients (9 of 20) usage of scheme with reference electrode in ipsilateral ear shows higher performance for window length between 2 and 10 seconds and in entire range of preictal lengths. Scheme with averaged reference shows higher performance when window length is in the range between 90 and 300 seconds.

Author Biographies

O. Yu. Panichev, National Technical University of Ukraine "Igor Sikorsky Kyiv Polytechnic Institute"

Panichev Oleg

A. O. Popov, National Technical University of Ukraine "Igor Sikorsky Kyiv Polytechnic Institute"

Popov Anton

V. I. Kharytonov, TMO "Psychiatry", Kyiv

Kharytonov Volodymyr

Published

2017-03-30

Issue

Section

Radioelectronics Medical Technologies