ABSTRACT

Designing brain–computer interface (BCI) experiments requires knowledge in many different disciplines: from neurosciences to signal processing and machine learning, through psychology. However, very few people have skills in all these disciplines. Yet, a lack of knowledge in a single aspect of BCIs is likely to result in flaws in the experimental design, statistical analyses, or the interpretation of the results. Moreover, because the BCI field is relatively 614young, no widely accepted guidelines are available yet, while at the same time an exponentially increasing number of research teams contribute to this field and would benefit from such guidelines. Thus, the objective of this chapter is to propose, in a pedagogical way, step-by-step guidance to design a rigorous BCI experiment. We name potential pitfalls and explain how to avoid them using concrete examples. This chapter could be seen as a checklist of points that should be addressed when the aim is to design rigorous and scientifically valid BCI studies and experiments. It is structured into three categories: (1) the acquisition of brain signals, (2) data processing issues, and (3) the experimental design and consideration of the BCI user.