Seizure detection by recurrent backpropagation neural network analysis

Author

Bates, Robyn R. ; Sun, Mingui ; Scheuer, Mark L. ; Sclabassi, Robert J.

Author_Institution

Dept. of Neurological Surg., Pittsburgh Univ., PA

fYear

2003

fDate

24-24 Sept. 2003

Firstpage

312

Lastpage

317

Abstract

Detection of the onset of seizure-activity in the EEG record of an epileptic patient is an important step in the localization of seizure foci. We employ the use of three multilayer recurrent neural-network architectures to detect seizure-activity onset in multichannel subdural EEG (SEEG) data. Representing each architecture as (input layer-hidden layer-output layer), the three neural networks examined were 5-10-5, 5-5-1, and 5-10-1

Keywords

backpropagation; electroencephalography; medical signal processing; neurophysiology; patient diagnosis; recurrent neural nets; EEG record; epileptic patient; multichannel subdural EEG data; multilayer recurrent neural-network architectures; recurrent backpropagation neural network analysis; seizure foci localization; seizure-activity detection; Artificial neural networks; Backpropagation; Biological neural networks; Data analysis; Electroencephalography; Epilepsy; Neural networks; Recurrent neural networks; Signal analysis; Surgery;

fLanguage

English

Publisher

ieee

Conference_Titel

Uncertainty Modeling and Analysis, 2003. ISUMA 2003. Fourth International Symposium on

Conference_Location

College Park, MD

Print_ISBN

0-7695-1997-0

Type

conf

DOI

10.1109/ISUMA.2003.1236179

Filename

1236179