Creación de un modelo Interfaz Cerebro-Computadora (BCI) basada en señales Potenciales Visualmente Evocados en Estado Estacionario (SSVEP) para aplicaciones de rápida respuesta.

Abstract

It has been proven that the characteristics of the stimulus can influence the strength of the SSVEP response in the user. However, a complete study of the influence of different characteristics was not considered in previous works. In this study, we have decided to test the most important characteristics of a visual stimulus that can affect the results of a BCI system based on SSVEP i.e., frequency, type of light source, color, illuminance, and size. For the present study, we have used a cyclic methodology called action research that allows us to perform different experiments to compare the best results generated by a certain characteristic of a visual stimulus and combine it with other characteristics, finding in this way, the optimal combination of stimulus’ characteristics that can allow to implement a more precise BCI system. For the experimentation, we have used the CCA algorithm and a new algorithm which we have called RCA for signal classification. The experiments’ results indicated that a white color circular visual stimulus of 7 to 10 Hz generated by an LED with a diffuser filter reaching 6 lux delivered the best SSVEP responses. With this combination of characteristics, it was possible to build different BCI prototypes having up to 4 simultaneous stimuli with a precision higher than 82% and a response time of 0.5s. In a similar way, it was also possible to simulate BCI systems of up to 7 simultaneous stimuli with a precision higher than 90% and a response time of 2.5s.