Immersive Virtual Reality Feedback in a Brain Computer Interface for Upper Limb Rehabilitation : A Pilot Study

• Autores: Pacheco Barrios, Kevin, Carranza-Urquizo, Erick, Achanccaray, David, Aguilar Mendoza, Luis Ángel
• Formato: artículo
• Fecha de Publicación: 2018
• Institución: Universidad San Ignacio de Loyola
• Repositorio: USIL-Institucional
• Lenguaje: inglés
• OAI Identifier: oai:repositorio.usil.edu.pe:20.500.14005/3920
• Enlace del recurso: https://www.scopus.com/record/display.uri?eid=2-s2.0-85034220410&origin=resultslist; https://hdl.handle.net/20.500.14005/3920; https://onlinelibrary.wiley.com/doi/full/10.1002/ana.25331; https://dx.doi.org/10.1002/ana.25331
• Nivel de acceso: acceso embargado
• Materia: Visualización de datos; Procesamiento de datos; Investigación sobre el cerebro

The brain‐computer interface (BCI) uses electrical signals from the brain and uses them as information to control an external device and has the potential to stimulate neuroplasticity in motor impairment. Objective. Implement an immersive virtual reality and BCI system for upper limb rehabilitation and test its safety. Methods: 34 healthy subjects participated. We used an EEG of 16 channels placed on the scalp corresponding to the motor and pre‐motor area, an amplifier (g.USBAmp) and a signal classification method (ANFIS classifier), to detect mu and beta signals related to the intention of movement and used to control an immersive virtual reality representation of an upper limb; and perform 2 movements: flexion and extension, according to the brain signals detected. In addition, the classification percentages and their association with spatial abilities of participants were measured. Results: In this study, the system showed to acquire, amplify and adequately classify the brain signals to control the virtual upper limb, with a success rate of 79.6% and 76.1%, for extension and flexion, respectively. No adverse effect was evidenced. Classifications success show a correlation with spatial abilities, especially mental rotation test (R = 0.56, p < 0.05), block design (R = 0.49, p < 0.05) and matrix reasoning (R = 0, 54, p < 0.05). Conclusions: The implemented system adequately controls the virtual reality software by intention of movement, showing an adequate level of safety.