A deep learning approach for epilepsy seizure detection using EEG signals
Electroencephalogram (EEG) is an effective non-invasive way to detect sudden changes in neural brain activity, which generally occurs due to excessive electric discharge in the brain cells. EEG signals could be helpful in imminent seizure prediction if the machine could detect changes in EEG pattern...
| Autores Principales: | Kaushik, Manoj, Singh, Divyanshu, Kishore-Dutta, Malay, Travieso, Carlos M. |
|---|---|
| Formato: | Artículo |
| Idioma: | Inglés |
| Publicado: |
Editorial Tecnológica de Costa Rica (entidad editora)
2022
|
| Materias: | |
| Acceso en línea: |
https://revistas.tec.ac.cr/index.php/tec_marcha/article/view/6461 https://hdl.handle.net/2238/14154 |
| Sumario: |
Electroencephalogram (EEG) is an effective non-invasive way to detect sudden changes in neural brain activity, which generally occurs due to excessive electric discharge in the brain cells. EEG signals could be helpful in imminent seizure prediction if the machine could detect changes in EEG patterns. In this study, we have proposed a one-dimensional Convolutional Neural network (CNN) for the automatic detection of epilepsy seizures. The automated process might be convenient in the situations where a neurologist is unavailable and also help the neurologists in proper analysis of EEG signals and case diagnosis. We have used two publicly available EEG datasets, which were collected from the two African countries, Guinea-Bissau and Nigeria. The datasets contain EEG signals of 318 subjects. We have trained and verify the performance of our model by testing it on both the datasets and obtained the highest accuracy of 82.818%. |
|---|