Table 3. Comparison with other studies.

Author	Electrodes	Task	Classified feature(s)	Classifier	Accuracy (%)
Ong et al. [50]	32	VEP	Power spectral density (PSD)	kNN	83
Alyasseri et al. [51]	6	Mental task (counting)	Multi-objective flower pollination algorithm with wavelet transform (MOFPA-WT)	ANN	85.12
Falzon et al. [42]	32	VEP	Frequency components up to 5th harmonic	kNN	91.7
Koutras et al. [52]	56	Sleep	Time-Domain Descriptors (10), Frequency-Domain Descriptors (17), Wavelet Domain Descriptors (4)	kNN	95
Fukami et al. [43]	4	VEP	5 frequency components	Mahalanobis distance	95
Yang et al. [53]	4	Motor movement, imagery	Wavelet log-DCT (WLD)	Fisher’s Linear Discriminant	98.5
Kaewwit et al. [54]	4	Resting state	Combined ICA and AR	kNN	98.51
Arnau et al. [55]	32	Video stimulation	PSD	ANN	99
Thomas et al. [56]	19	Resting state	PSD	Mean correlation coefficients	99
Schetinin et al. [57]	64	Motor movement, imagery	Group Method of Data Handling (GMDH)	SVM	100
Proposed approach	4	Relaxation before stimulation	IHAR	QDA, SVM, kNN	100
	4	Visual stimulation	IHAR	kNN	98.8 ± 0.9
	4	Mental recall	IHAR	QDA, SVM, kNN	100