Banca de QUALIFICAÇÃO: MARCELO BEZERRA GRILO JÚNIOR

Uma banca de QUALIFICAÇÃO de DOUTORADO foi cadastrada pelo programa.
DISCENTE : MARCELO BEZERRA GRILO JÚNIOR
DATA : 21/06/2021
HORA: 13:30
LOCAL: Remoto
TÍTULO:

Joint Blind Source Separation Applied to EEG Signals in the Search for Empathic Reactions

PÁGINAS: 52
GRANDE ÁREA: Engenharias
ÁREA: Engenharia Elétrica
RESUMO:

The ability to express empathy for the suffering of others is a fundamental skill of human social behavior.  The brain electrical activity has characteristics that reflects an emotional condition and EEG signals can be used to study empathic reactions. EEG signals suffer interferences of the brain, the body, and the environment. To overcome this problem, it is common to apply Blind Source Separation (BSS) techniques. By recording EEG signals from several subjects, it is assumed that they present correlated brain electrical activity behavior, thus, it is possible to use Joint Blind Source Separation (JBSS) instead. The goal of JBSS is to jointly identify the multiset source components and mixing matrices by exploiting the similarities between distinct datasets and the diversity of each dataset. Initially, we proposed a JBSS method that exploit the Joint Diagonalization (JDIAG) of covariance matrices, a technique based on the sole use of second order statistics. This method was compared with two other approaches based on Independent Vector Analysis (IVA), implemented through a gradient-descent algorithm (IVA-VG) and a Newton-based algorithm (IVA-NW). The algorithms based on IVA performed better than JDIAG in recovering the sources, considering a simulation scenario with 3 datasets and 3 sources. IVA-NW also presented a faster convergence. In addition, preliminary analysis of EEG signals recorded for this study has shown that a greater brain activity was observed when the subjects watched the video of an actor with which they identified themselves the most (a person from the same region and same sex). Since JBSS explores the correlation between different datasets, in the next steps, we will apply the JBSS methods studied to identify patterns related to  empathy recognition.

MEMBROS DA BANCA:
Presidente - Interno ao Programa - 1544392 - ALINE DE OLIVEIRA NEVES PANAZIO
Membro Titular - Examinador(a) Interno ao Programa - 1946319 - DIOGO COUTINHO SORIANO
Membro Titular - Examinador(a) Interno ao Programa - 1761107 - RICARDO SUYAMA
Membro Suplente - Examinador(a) Externo ao Programa - 3008017 - DENIS GUSTAVO FANTINATO