Table 5. An overview of the major works concerning the mind-based control of humanoid robots.

The abbreviations for the robot walking behaviors are defined as follows: “MI” is a motor imagery model, “ITR” is the information transfer rate, and the P300 models indicated with an asterisk (*) use 5 repetitions per trial.

Publication	Brainwave models	Environment	Task	Evaluation criteria
				Accuracy (%)	Response time (s)	ITR (bits/min)
Bell et al., 2008[ 12 ]	4-class P300*	Real	Object selection	95	5	24
Li et al., 2011[ 13 ]	3-class MI	Real	Humanoid walking	N/A	N/A	N/A
Chung et al., 2011[ 40 ]	3-class SSVEP	Virtual	Navigation	77.5	N/A	N/A
M Bryan et al., 2011[39], 2012[41]	5-class SSVEP	Virtual & Real	Control of a humanoid arm	N/A	N/A	N/A
Thobbi et al., 2010[22]	2-class MI	Real	Navigation	78.35	N/A	N/A
Finke et al., 2013[38]	P300 and2-class MI	Real	Tasks related to assistance and telepresence	P300: 80 MI: 78	N/A	N/A
Gergondet et al., 2011[42], 2012[43]	4-class SSVEP	Real	Navigation and object selection	N/A	N/A	N/A
Chae et al., 2012[37]	3-class MI	Real	Navigation	80.9	1.84	14.02
Choi et al., 2013[10]	2-class SSVEP, 2-class MI and 4-class P300*	Real	Navigation and object selection	SSVEP: 84.4 MI: 84.6 P300: 91	N/A N/A 5.35	11.6 11.8 15.2
Tidoni et al., 2014[11]	6-class SSVEP	Real	Navigation and pick-and-place tasks	N/A	7.52	N/A
Bouyarmane et al., 2014[9]	2-class MI	Virtual	Moving up stairs	N/A	N/A	N/A
Present work	4-class SSVEP and 6-class P300*	Real	Humanoid walking and navigation	SSVEP: 90.3 P300: 91.3	3.65 6.6	24.7 18.8