TY - GEN
T1 - Epileptic Seizure Prediction from Scalp EEG Using Ratios of Spectral Power
AU - Salvatierra, Noe
AU - Sakanishi, Renato
AU - Flores, Christian
N1 - Publisher Copyright:
© 2020 IEEE.
PY - 2020/10/21
Y1 - 2020/10/21
N2 - Epilepsy is a neurological disorder that affects a wide range of people from different gender and ages. Some known Epilepsy causes are biological conditions, brain tumors, brain infections, and seizures' occurrence is its impact on a patient. Living daily with this condition is not an easy task, so it is necessary to innovate in treatments. Many outstanding articles reported that epileptic seizure prediction provides quality of life to patients. In this work, we present the use of Absolute Spectral Power and Spectral Power Ratio on electroencephalography (EEG) signals to predict seizures in epileptic patients. We propose threshold-definition criteria by using statistical measures of central tendency and spread from the analysis of segments of data before a seizure. The brain signals of epileptic subjects, who suffered several seizures, were obtained from the CHB-MIT Scalp EEG Database. The results reported that the best prediction time was 60 minutes, and the success rate was 100%, using three segments of data before a seizure for testing. These results are compared to a seizure prediction algorithm so that our results matched it with a low computational cost.
AB - Epilepsy is a neurological disorder that affects a wide range of people from different gender and ages. Some known Epilepsy causes are biological conditions, brain tumors, brain infections, and seizures' occurrence is its impact on a patient. Living daily with this condition is not an easy task, so it is necessary to innovate in treatments. Many outstanding articles reported that epileptic seizure prediction provides quality of life to patients. In this work, we present the use of Absolute Spectral Power and Spectral Power Ratio on electroencephalography (EEG) signals to predict seizures in epileptic patients. We propose threshold-definition criteria by using statistical measures of central tendency and spread from the analysis of segments of data before a seizure. The brain signals of epileptic subjects, who suffered several seizures, were obtained from the CHB-MIT Scalp EEG Database. The results reported that the best prediction time was 60 minutes, and the success rate was 100%, using three segments of data before a seizure for testing. These results are compared to a seizure prediction algorithm so that our results matched it with a low computational cost.
UR - https://www.scopus.com/pages/publications/85097809951
U2 - 10.1109/EIRCON51178.2020.9254056
DO - 10.1109/EIRCON51178.2020.9254056
M3 - Conference contribution
AN - SCOPUS:85097809951
T3 - Proceedings of the 2020 IEEE Engineering International Research Conference, EIRCON 2020
BT - Proceedings of the 2020 IEEE Engineering International Research Conference, EIRCON 2020
PB - Institute of Electrical and Electronics Engineers Inc.
T2 - 2020 IEEE Engineering International Research Conference, EIRCON 2020
Y2 - 21 October 2020 through 23 October 2020
ER -