Redução do Eletroencefalograma durante monitorização contínua de paciente crítico

• Ano de defesa: 2017
• Autor(a) principal: Santos, Talita Endriely Batista dos
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Federal do Rio de Janeiro; Brasil; Instituto Alberto Luiz Coimbra de Pós-Graduação e Pesquisa de Engenharia; Programa de Pós-Graduação em Engenharia Biomédica; UFRJ
• Palavras-chave em Português: Engenharia biomédica; HaEEG; aEEG; Monitorização contínua; Transformada de Hilbert; Redução de dados; CNPQ::ENGENHARIAS::ENGENHARIA BIOMEDICA
• Link de acesso: http://hdl.handle.net/11422/8330

"[ENG] This work aimed to apply the Hilbert Transform to the Electroencephalogram (EEG) data reduction. The scope of this work was to analyze differences between two EEG reduction methods, amplitude-integrated EEG (aEEG) and Hilbert aEEG (HaEEG). The main difference is in the signal envelope obtention: the aEEG uses a 5th - order Butterworth lowpass filter, and the HaEEG uses the analytical signal module of Discrete Hilbert Transform (DHT). One aimed at investigating the use of DHT to obtain the envelope in EEG reduction; to compare different parameters to obtaining the envelope using the Butterworth filter; to analyze and to compare the Butterworth and Hilbert envelopes; and to compare the filtering of HaEEG signal within the band 2- 15 Hz and within 1 and 70 Hz. The methodology consisted of evaluating the Butterworth filter order and the scaling parameter for the aEEG by means of a known signal, to visually compare aEEG and HaEEG in the same frequency range (2 to 15 Hz) and to compare HaEEG in frequency ranges 2 to 15 Hz and 1 to 70 Hz. One fulfilled the comparison between segments of 15 s of raw signal, 2-15 Hz and 1-70 Hz filtered signal, and between the respective envelopes, of the aEEG and HaEEG. In conclusion the Hilbert Transform is an effective method to reduce EEG signals. The aEEG envelope is better obtained by means of 2nd order Butterworth using Square Law and No Gain. Hilbert envelope resembles the original signal and highlights better the seizure segments in relation to background activity compared to Butterworth. The frequency range within 2-15 Hz can omit epileptic characteristics."