ABSTRACT
This chapter presents the brain–computer interfaces (BCIs) developed in our laboratory at UFES/Brazil along 20 years of investigation. A limitation of both Steady-State Visual Evoked Potential(SSVEP)-based BCIs is that they are dependent on eye gaze, which are classified as “dependent SSVEP-based BCIs”. This BCI is based on depth-of-field using SSVEP with stimuli by LEDs, which is used to command the telepresence robot movements (with bidirectional communication of video/audio) and its onboard alternative communication system. Currently, new methods are being applied in our laboratory to improve the success rate of our BCIs for motor imagery detection. For example, short-time Fourier transform, sparseness constraints, and total power in time-frequency representation (to locate the subject-specific bands with the highest power), in addition to Riemannian geometry (to extract spatial features) and a fast version of neighborhood component analysis (to increase the class separability) are used.
