Interface cérebro-computador híbrida e colaborativa no processo de tomada de decisão

Detalhes bibliográficos
Ano de defesa: 2017
Autor(a) principal: Schuh, Ânderson Rodrigo lattes
Orientador(a): Campos, Márcia de Borba lattes
Banca de defesa: Não Informado pela instituição
Tipo de documento: Dissertação
Tipo de acesso: Acesso aberto
Idioma: por
Instituição de defesa: Pontifícia Universidade Católica do Rio Grande do Sul
Programa de Pós-Graduação: Programa de Pós-Graduação em Ciência da Computação
Departamento: Faculdade de Informática
País: Brasil
Palavras-chave em Português:
EEG
Área do conhecimento CNPq:
Link de acesso: http://tede2.pucrs.br/tede2/handle/tede/7711
Resumo: Brain-Machine Interface (BCI) or Brain-Computer Interface (BCI) is a computer system capable of establishing communication between human neurophysiological activity and a computer. A hybrid BCI (hBCI) consists of a combination of two or more types of BCIs, two or more signal acquisition techniques, or a combination of BCI with other non-BCI based interaction techniques. A Collaborative BCI (cBCI) integrates the brain activity of a group of individuals, mainly acting in the increase of the human capacity. Low-cost electroencephalogram (EEG) equipment is currently available in the market, one of which is the Emotiv EEG, which is portable, has 14 electrodes, and in addition to registering the neurophysiological signals, it processes and makes available them in the form of neural measurements, such as levels of attention and excitement. In addition to neural measures, other measures may reveal an individual's behavior, such as the speed with which he responds to a challenge, which may suggest how confident he is about this decision-making. This work has as main objective "To verify if neural and behavioral measures have relation with the right and wrong decision making". Initially, a systematic review of the literature was carried out. Afterwards, a data collection system and a decision-making task based on Rapid Serial Visual Presentation (RSVP) were developed. The experiment consisted of 10 participants, in which each one performed 112 tests, recording the neural measurements taken by Emotiv EEG, besides the Reaction Time (RT) as a behavioral measure and the response given by the user, both collected by a conventional keyboard. Statistical techniques, such as descriptive analysis, including data summarization and boxplot charts, and multivariate analysis were used for the data analysis, using logistic regression to estimate the relationship between neural and behavioral measures with the decisions made. The proposed task proved to be efficient, revealing in the results that the difficulty was effective. The developed database proved to be efficient in synchronizing the task data and the recorded measurements. After different approaches of statistical analysis of the data, a regression model that could explain with high explanatory power the data sampled was not found. Thus, based on the experiment performed and statistical analyzes, no relationship was found between neural and behavioral measures and the correct or wrong decision-making.