Table 3.

Feature classification methods in different EEG paradigms.

EEG paradigms	Authors	Classification methods
SSVEP	Chu et al. [125]	LDA, 3 classes (20, 15, 12 Hz)
	Bi et al. [126]	LDA, 2 classes (12, 13 Hz)
	Oikonomou et al. [92]	LDA, 5 classes (6.66, 7.5, 8.57, 10, 12 Hz)
	Maggi et al. [127]	RLDA, 5 classes (6, 7, 8, 10 Hz, idle)
	Singa and Haseena [128]	SVM, 4 classes (7, 9, 11, 13 Hz)
	Bi et al. [129]	SVM, 3 classes (12, 13 Hz, idle)
	Sakurada et al. [130]	SVM, 4 classes (6, 7, 8, nonfixation)
	Jian and Tang [35]	OVO RBF SVM, 5 classes (8, 10, 12, 14, 15 Hz)
	Cecotti and Gräser [131]	TDNN, 5 classes (13, 14, 15, 16, 17 Hz)
	Cecotti [132]	CNN, 5 classes (6.66, 7.5, 8.57, 10, 12 Hz)
	Hartmann and Kluge [133]	HMM, 3 classes (10, 12, 15 Hz)
	Ko et al. [134]	kNN, 2 classes (15, 20 Hz)
	Oikonomou et al. [92]	kNN, 5 classes (6.66, 7.5, 8.57, 10, 12 Hz)
P300	Gareis et al. [135]	LDA, P300 Speller
	Onishi and Natsume [136]	Ensemble Stepwise LDA, P300 Speller
	Elwardy et al. [137]	Disjunctive Normal Unsupervised LDA, P300 Speller
	Li et al. [31]	SVM, P300 speller
	Raju et al. [138]	Least Square SVM (LS-SVM), Competition III, Dataset II (P300 Speller)
	Li et al. [139]	Self-Training Semisupervised SVM, P300 Speller
	Yang et al. [140]	LVQNN, 7 classes (oddball paradigm)
	Turnip et al. [141]	MNN, raw data in Hoffmann et al.
	Cecotti and Gräser [142]	CNN, P300 Speller
	Helmy et al. [143]	HMM, raw data in Hoffmann et al.
	Speier et al. [144]	HMM, P300 Speller
	Syan and Harnarinesingh [80]	kNN, P300 Speller, BCI Competition II
	Chikara and Ko [145]	kNN, 2 classes
MI	Chen et al. [32]	LDA, 2 classes (left or right hand)
	Steyrl et al. [146]	Shrinkage RLDA, 2 classes (right hand and feet)
	Vidaurre et al. [147]	KALDA, 2 classes (left or right hand)
	Rathipriya et al. [148]	SVM, 2 classes, Dataset IVa (right hand, foot) and IVb (left hand, foot), BCI Competition III
	Oskoei et al. [149]	supervised and unsupervised SVM, 3 classes, Dataset V, BCI Competition III (left or right hand, word association)
	Siuly and Li [150]	LS-SVM, 2 classes, Dataset IVa and IVb, BCI Competition III
	Hamedi et al. [151]	BP, 3 classes (left or right hand, tongue)
	Wei et al. [152]	LVQNN, 2 classes (left or right hand)
	Hazrati and Erfanian [153]	APNN, 2 classes (left or right hand), BCI competition 2003, data set III
	Haselsteiner and Pfurtscheller [154]	TDNN, 2 classes (left or right hand)
	Siuly et al. [155]	Naïve Bayes, 2 classes, Dataset IVa and IVb, BCI Competition III
	Obermaier et al. [156]	HMM, 2 classes (left or right hand)
	Suk and Lee [157]	HMM, Dataset IIa, BCI Competition IV (2008), 4 classes (left or right hand, feet, tongue)
		kNN, 2 classes (left or right hand),
	Bashar et al. [89]	BCI Competition 2003 data set (motor imagery III)
	Bashar and Bhuiyan [158]	BCI Competition II data set (GRAZ motor imagery III)
	Diana Eva and Tarniceriu [159]	kNN, 2 classes (left or right hand), BCI Competition 2002