Author(s): Marques, Joao M. C. ; Cerdeira, Hilda A. [UNESP] ; Tanaka, Edgar ; Vitor, Conrado de ; Gomez, Paula ; Zheng, H. ; Callejas, Z. ; Griol, D. ; Wang, H. ; Hu, X ; Schmidt, H. ; Baumbach, J. ; Dickerson, J. ; Zhang, L.
Date: 2019
Persistent ID: http://hdl.handle.net/11449/185428
Origin: Oasisbr
Made available in DSpace on 2019-10-04T12:35:19Z (GMT). No. of bitstreams: 0 Previous issue date: 2018-01-01
Predicting epileptic seizure occurrence has long been a goal of the community surrounding it. Accurate prediction, however, is still elusive. This work presents a modified pipeline for the training of seizure prediction systems which aims to attenuate the effects of current data labeling strategies - and consequent data mislabeling of samples that heavily affect classifiers that are trained on it. This paper also presents a seizure prediction system trained following the proposed pipeline, which improved our system's performance by reducing its time-in-warning (TiW) by over 14%, while improving its prediction sensitivity to 72.4%, bringing its performance closer to the state-of-the-art performance (83.1% prediction sensitivity) for systems with similar TiW (41%) [1], while only requiring input from two scalp EEG electrodes - without making use of any variables external to the single EEG channels.
Epistem Gomez & Gomez Ltda ME, Cietec, Ave Prof Lineu Prestes 2242,Sala 244, BR-05508000 Sao Paulo, Brazil
Univ Sao Paulo, Escola Politecn, Rua Prof Luciano Gualberto Travessa 3,380, BR-05508010 Sao Paulo, Brazil
Univ Estadual Paulista, Inst Fis Teor UNESP, Rua Dr Bento Teobaldo Ferraz 271,Bloco 2, BR-01140070 Sao Paulo, Brazil
Univ Estadual Paulista, Inst Fis Teor UNESP, Rua Dr Bento Teobaldo Ferraz 271,Bloco 2, BR-01140070 Sao Paulo, Brazil
