“…Traditionally, BCI applications rely on dense, high-dimensional feature vectors produced by multi-electrode scanning caps with high temporal resolution (Lotte et al, 2007), which threatens the responsiveness of BCI from a user experience standpoint and places high requirements on end-user hardware. (Crowley et al, 2010;Grierson and Kiefer, 2011;Chuang et al, 2013;Johnson et al, 2014). However, the use of consumer EEGs for the direct, real-time control of software interfaces has proven more difficult, as the number of electrodes on these headsets limit the spatial resolution required to discriminate between mental gestures (Carrino et al, 2012;Larsen and Hokl, 2011).…”