Português English

Dissertação de Jaime Andrés Riascos Salas

Detalhes do Evento

Aluno: Jaime Andrés Riascos Salas
Orientador: Prof. Dr. Dante Augusto Couto Barone

Título: Do I have a third arm? Towards a Supernumerary Motor Imagery Brain-Computer Interface in Virtual Reality

Linha de Pesquisa: Interação Humano-Computador, Realidade Virtual e Aumentada

Data: 15/03/2019
Hora: 09h
Local: AUD-1 (Auditório 1) do Prédio 43412 do Instituto de Informática da UFRGS

Banca Examinadora:
Prof. Dr. Marcelo Walter (UFRGS)
Prof. Dr. Edison Pignaton de Freitas (UFRGS)
Prof. Dr. Fabien Lotte (Inria – por videoconferência)

Presidente da Banca: Prof. Dr. Dante Augusto Couto Barone

Abstract: Brain-Computer Interface (BCI) opened the possibility of communicating the human with systems and devices using brain signals. This area has inspired researchers to develop several applications such as medical rehabilitation of disabled people, robotic prostheses, games, and assisted virtual reality (VR) scenarios. In effect, VR is currently employed to improve the BCI’s reliability through realistic and natural training and feedback. Motor imagery Brain-Computer Interface (MI-BCI) is a paradigm widely used for controlling external devices by imagining bodily movements. So far, nevertheless, MI-BCI has only used embodied limbs for the imaginary tasks, even though that psychology has conclusively demonstrated the existence of body transfer illusions (rubber hand illusion).Thus, this thesis studies and explores the inclusion of an imaginary third arm as a part of the control commands for MI-BCI while comparing the effectiveness of using the conventional arrows and fixation cross as training step (Graz condition) against a first-person view of a human-like avatar performing the corresponding task (Hands condition). Both conditions made in a VR scenario. Ten healthy subjects participated in a two-session experiment involving open-close hand tasks (including a third arm that comes out from the chest). The EEG analysis shows a strong power decrease in the sensory-motor areas for the third arm task in both conditions. Furthermore, the offline classification results show that a third arm can be effectively used as a control command (accuracy > 0.62%). Likewise, Hands condition (67%) outperforms Graz condition (63%) significantly, suggesting that the realistic scenario can reduce the abstractness of the third arm and improve the performance, however, this condition induces a cognitive load. Finally, with this thesis, a door is open towards the creation of a supernumerary MI-BCI system with the inclusion of non-embodied motor imagery tasks.

Keywords: Brain-Computer Interface, Virtual Reality, Rubber Hand Illusion, Cognitive Load, Electroencephalography.