Epileptic seizure prediction using machine learning techniques

• Autor(a) principal: Salvador, Carolina Duarte
• Data de Publicação: 2023
• Tipo de documento: Dissertação
• Idioma: eng
• Título da fonte: Repositório Científico de Acesso Aberto de Portugal (Repositórios Cientìficos)
• Texto Completo: http://hdl.handle.net/10400.6/13643
• topic: Eeg; Epilepsia; Machine Learning; Previsão de Ataques; Domínio/Área Científica::Engenharia e Tecnologia::Bioengenharia

Epileptic seizures affect about 1% of the world’s population, thus making it the fourth most common neurological disease, this disease is considered a neurological disorder characterized by the abnormal activity of the brain. Part of the population suffering from this disease is unable to avail themselves of any treatment, as this treatment has no beneficial effect on the patient. One of the main concerns associated with this disease is the damage caused by uncontrollable seizures. This damage affects not only the patient himself but also the people around him. With this situation in mind, the goal of this thesis is, through methods of Machine Learning, to create an algorithm that can predict epileptic seizures before they occur. To predict these seizures, the electroencephalogram (EEG) will be employed, since it is the most commonly used method for diagnosing epilepsy. Of the total 23 channels available, only 8 will be used, due to their location. When a seizure occurs, besides the visible changes in the EEG signal, at the moment of the seizure, the alterations before and after the epileptic seizure are also noticeable. These stages have been named in the literature: • Preictal: the moment before the epileptic seizure; • Ictal: the moment of the seizure; • Postictal: the moment after the seizure; • Interictal: space of time between seizures. The goal of the predictive algorithm will be to classify the different classes and study different classification problems by using supervised learning techniques, more precisely a classifier. By performing this classification when indications are detected that a possible epileptic seizure will occur, the patient will then be warned so that he can prepare for the seizure.