Utilize este identificador para referenciar este registo:
https://hdl.handle.net/1822/61996
Título: | Epileptic event detection algorithm for ambulatory monitoring platforms |
Autor(es): | Pinho, F. Ferreira, João Reis, Joana Sousa, Nuno Cerqueira, João José Correia, J. H. Dias, N. S. |
Palavras-chave: | Electroencephalography Algorithm design and analysis Event detection Epilepsy Monitoring Real-time systems Classification algorithms root mean square ambulatory |
Data: | Jul-2014 |
Editora: | IEEE |
Citação: | Pinho, F., Ferreira, J., Reis, J., Sousa, N. J., Cerqueira, J. J., Correia, J. H., & Dias, N. S. (2014, June). Epileptic event detection algorithm for ambulatory monitoring platforms. In 2014 IEEE International Symposium on Medical Measurements and Applications (MeMeA) (pp. 1-6). IEEE. |
Resumo(s): | Detecting epileptic electroencephalography (EEG) signals, both automatically and accurately, is significant in ambulatory long-term monitoring patients with epilepsy. In this study, it is presented a novel epileptic-like event detection algorithm based on a mixture of amplitude, frequency and spatial analysis with rule-based decision. In this work, EEG signals from 6 different subjects were searched for epileptic-like and normal data segments. The herein proposed algorithm detects putative epileptic EEG channels by comparing the RMS values of EEG activity with a hysteresis threshold, on a channel basis. The raw EEG signals are filtered with an artefact attenuation technique. The threshold is calculated on a reviewer-visually-selected baseline epoch, free of artefacts. Generalized epileptic activity detection is based on a spatial decision rule. Experimental results have shown detection rates as high as 95% with a false-negative rate as low as 1%. The algorithm seems to show a promising detection performance, even on artifact contaminated datasets. The proposed algorithm is intended to be used in real-time ambulatory monitoring of epileptic patients, with subject personalization, small size window analysis, good artefact immunity and no need for classifier training |
Tipo: | Artigo em ata de conferência |
URI: | https://hdl.handle.net/1822/61996 |
ISBN: | 978-1-4799-2920-7 |
e-ISBN: | 978-1-4799-2921-4 |
DOI: | 10.1109/MeMeA.2014.6860104 |
Versão da editora: | https://ieeexplore.ieee.org/abstract/document/6860104 |
Arbitragem científica: | yes |
Acesso: | Acesso restrito autor |
Aparece nas coleções: |
Ficheiros deste registo:
Ficheiro | Descrição | Tamanho | Formato | |
---|---|---|---|---|
pinho2014.pdf Acesso restrito! | 1,66 MB | Adobe PDF | Ver/Abrir |