Analysis of the Influence of the MVDR Filter Parameters on the Performance of SSVEP-Based BCI

Brain-Computer Interface (BCI) is a communication method based on brain signals analysis. The interface enables controlling applications such as a wheelchair with minimal muscle effort, making BCI systems attractive in assistive technology development. Currently, Steady-State Visually Evoked Potential (SSVEP) represents one of the most promising BCI paradigms, since a specific physiological brain response is evoked when a subject is exposed to continuously flickering visual stimuli. In this study, we evaluated how the parameters of the Minimum Variance Distortionless Response (MVDR) filter impact the performance of the SSVEP-based BCI. Three parameters were analyzed: filter order, number of EEG signals combined at the filter input, and number of electrodes employed for filtering. Our results show that it is convenient to employ fewer electrodes, as they are closer to the visual cortex region, and to combine them spatially, using low filter orders. The best performance, among the tested configurations, was 80.20 +/- 6.65%, obtained with filter order nine, employing nine EEG signals and spatially combining the inputs with eight signals at a time.