ABSTRACT
This chapter shows the results of our researches conducted with Brain–Computer Interfaces (BCIs) based on Electroencephalography (EEG). Regarding the performance, relevant issue in BCIs is the ability of efficiently converting user intentions into correct actions, and how to properly measure this efficiency. Active BCIs have outputs derived from brain activity, which are directly and consciously controlled by user and thus independent of external events. Reactive BCIs have outputs derived from brain activity, which arise due to reaction to external stimulation, which is indirectly modulated by the user. Passive BCIs have outputs derived from implicit information on actual user mental state, which arises arbitrarily without the purpose of voluntary control. The traditional metrics to measure the efficiency of a BCI are classification accuracy, kappa, false-positive rate, true-positive rate, and information transfer rate. BCI applications are employed in four major areas: assistance for patients with severe motor disabilities, diagnosis of disorders of consciousness, entertainment applications, and recognition of affective or cognitive states.
