PPGINF PÓS-GRADUAÇÃO EM ENGENHARIA DA INFORMAÇÃO FUNDAÇÃO UNIVERSIDADE FEDERAL DO ABC Phone: Not available http://propg.ufabc.edu.br/ppginfo

Banca de DEFESA: GODOFREDO QUISPE MAMANI

Uma banca de DEFESA de DOUTORADO foi cadastrada pelo programa.
DISCENTE : GODOFREDO QUISPE MAMANI
DATA : 16/12/2019
HORA: 14:00
LOCAL: Auditório, 8º andar, Bloco B, Campus SA da Fundação Universidade Federal do ABC, localizada na Avenida dos Estados, 5001, Santa Terezinha, Santo André, SP
TÍTULO:

Topographic biomarkers to support early diagnosis of Alzheimer's disease based on computational analysis of electroencephalograms registered during working memory tasks


PÁGINAS: 145
GRANDE ÁREA: Engenharias
ÁREA: Engenharia Biomédica
SUBÁREA: Bioengenharia
ESPECIALIDADE: Processamento de Sinais Biológicos
RESUMO:

Alzheimer's disease (AD) is a neurodegenerative syndrome that affects millions of people worldwide. In addition, people with mild cognitive impairment (MCI) are in a risk group that should be closely followed, as there is a high probability of progression to AD. In this research, three studies were designed to improve and validate topographic biomarkers based on computational analysis of the electroencephalogram (EEG) collected during the execution of N-back working memory tasks, in order to support early diagnosis of MCI and AD. Participants were 15 patients diagnosed with AD, 21 individuals diagnosed with MCI and 27 healthy elderly (HE) as control group. Subjects underwent a three-level visual N-back task, with rising difficulty in working memory load, during which EEG signals (32 channels) were recorded.

The first study aimed to explore the use of Event-Related Potentials (ERP) and Event-Related Synchronization / Desynchronization (ERS / ERD). Significant differences were found between patient (MCI and AD) and control (HE) groups in the ERP during the execution of N-back tasks, predominantly in the frontocentral-parietal electrodes. Most of the differences were observed at 400-700 ms post-stimulus, where the control group showed larger amplitudes than the patients (MCI and AD), i.e., differences were observed in the amplitude of the P450 component, typically related to working memory update. Additionally, ERD / ERS measurements revealed that subjects in the HE group elicited consistently higher ERD-alpha responses than patients with MCI and AD while performing the 1-back task with match trials, with differences located in the frontal, central and occipital regions. In addition, in the non-match trials, it was possible to distinguish between patients with MCI and AD at the 0-back level, with MCI presenting more ERD-theta than AD in the temporal and frontotemporal areas.

The second study aimed to analyze source-space EEG,  focusing on the analysis of EEG segments before and after the same N-back working memory tasks, this time exclusively aiming to support the diagnosis of MCI. To this end, from sensor-space EEG data, the LORETA software provided the location of Brodmann regions (source-space) in which there were significant differences in the MCI vs. HE comparison. Significant differences were found in all working memory tasks, mainly located in the frontal and temporal lobe regions.

Finally, the third study aimed to extract features in the source space, identified as topographic biomarkers from the EEG analysis captured during N-back tasks, to automatically classify participants into MCI and HE categories. For this purpose, machine learning techniques with classifiers implemented through the Scikit-learn library in Python language were used. The prediction of the classifier reached accuracy, sensitivity and specificity around 95%. This shows that this source-space feature extraction method, focusing on the analysis of EEG segments before and after N-back tasks, can surpass the best performances reported in the literature, indicating that this technique may represent a clinically relevant tool to support early diagnosis of MCI and AD.


MEMBROS DA BANCA:
Presidente - Interno ao Programa - 1672975 - JOAO RICARDO SATO
Membro Titular - Examinador(a) Interno ao Programa - 1761107 - RICARDO SUYAMA
Membro Titular - Examinador(a) Externo ao Programa - 1955999 - ANDRE MASCIOLI CRAVO
Membro Titular - Examinador(a) Externo à Instituição - PAULO AFONSO MEDEIROS KANDA
Membro Titular - Examinador(a) Externo à Instituição - MARCIA REGINA COMINETTI - UFSCAR
Membro Suplente - Examinador(a) Externo ao Programa - 1722875 - DAVID CORREA MARTINS JUNIOR
Membro Suplente - Examinador(a) Externo à Instituição - CARLOS EDUARDO THOMAZ - FEI
Notícia cadastrada em: 11/11/2019 11:37
SIGAA | UFABC - Núcleo de Tecnologia da Informação - ||||| | Copyright © 2006-2020 - UFRN - sigaa-1.ufabc.int.br.sigaa-1