A contribution to the dynamics of epilepsy: identification and control for seizure attenuation

• Ano de defesa: 2023
• Autor(a) principal: Brogin, João Angelo Ferres
• Tipo de documento: Tese
• Tipo de acesso: Acesso aberto
• Idioma: eng
• Instituição de defesa: Universidade Estadual Paulista (Unesp)
• Palavras-chave em Português: Epilepsy; System identification; Seizure attenuation; Epilepsia; Identificação de sistemas; Atenuação de crises
• Link de acesso: https://hdl.handle.net/11449/253611

Epilepsy is one of the most common neurological disorders worldwide, affecting millions of people every year and leading to physical, psychological and social implications that may significantly compromise their lives. There are treatments today that help to mitigate the frequency and intensity of seizures, but none is fully successful, and the precise mechanisms that trigger them are not yet completely understood either. This has motivated researchers to address epilepsy from a theoretical point of view, resulting in the development of mathematical computational models that provide insights into seizure generation, synchronization and propagation. The Epileptor model is a breakthrough in this field due to its highly representative features, relatively easy implementation and adherence with experimental observations. Similarly, data-driven models have also been presented in the literature to describe the temporal signatures of electrophysiological recordings. Relevant questions that arise in this context are how to obtain dynamic models from real data for which a control system can be designed and how to define the nature of the control inputs that achieve seizure attenuation in both theoretical and experimental models. In this sense, the main contributions of this thesis are strategies to define the control inputs able to suppress or mitigate the epileptiform activity described by the Epileptor and real signals, based on the design of controllers, state observers and system identification techniques. Both are developed computationally but overcoming a series of numerical, theoretical and practical restrictions to obtain a final control approach with a strong practical appeal that can be potentially used by physicians as an auxiliary tool to better define the shape, intensity and frequency of the input stimulus that must be applied to attenuate seizures in electrical stimulation therapies. The results show that it is possible to reconstruct and attenuate the epileptiform patterns of both models from control signals that are specific to each level of brain activity and consistent with previous studies in the literature.