Table 1.

This table provides a summary of the feature extraction methods.

A Summary of Feature Extraction Methods
Temporal methods	Statistical Features [19,20]	$Mean\left(\overline{x}\right)=\frac{1}{T}{\sum }_{t=1}^T\left|x_t\right|$,
		$Std.Dev\left(\sigma \right)=\sqrt{\frac{1}{T}{\sum }_{t=1}^T{(x_t-\overline{x})}^2}$
		$Variance\left(x\right)=\frac{1}{T}{\sum }_{t=1}^T{(x_t-\overline{x})}^2$
		$skewness=\frac{\frac{1}{T}{\sum }_{t=1}^T{(x_t-\overline{x})}^3}{{\sigma }^3}$
		$kurtosis=\frac{\frac{1}{T}{\sum }_{t=1}^T{(x_t-\overline{x})}^3}{{\sigma }^4}$
	Hijorth features [21]	$Activity=\frac{{\sum }_{t=1}^T{(x_t-\overline{x})}^2}{T}$
		$Mobility=\frac{Var\left(\widehat{x_t}\right)}{Var\left(x_t\right)}$
		$Complexity=\frac{Mobility\left(\widehat{x_t}\right)}{Mobility\left(x_t\right)}$
	RMS [20]	$RM{S}_t=\sqrt{\frac{1}{N}{\sum }_{i-1}^N{x}_i^2}$
	IEEG [20]	$IEE{G}_t={\sum }_{i=1}^N|x_I$
	Fractal Dimension [22]	$D=\frac{log(L/a)}{log(d/a)}$
	Autoregressive modeling [21]	$x_t={\sum }_{i=1}^p{a}_i{x}_{t-i}+{ϵ}_t$ where {a for i = 1,…, p} are AR model coefficients and p is the model order
	Peak-Valley modeling [23,24]	Cosine angles, Euclidean distance between neighbouring peak and valley points
	Entropy [25,26]	$S=-{\sum }_{i=1}^N{p}_{i}ln{p}_i$
	Quaternion modeling [27]	$Mean\left(\mu \right)=\frac{\sum \left(q_{mod}\right)}{N}$ $Variance\left(\sigma \right)=\frac{{(\sum {\left(q_{mod}\right)}^2-\mu )}^2+\sum {\left(q_{mod}\right)}^2}{2N}$ $Contrast\left(con\right)=\frac{\sum {\left(q_{mod}\right)}^2}{N}$ $Homogeneity\left(H\right)=\sum \frac{\left(1\right)}{1+{\left(q_{mod}\right)}^2}$ $ClusterShade\left(cs\right)=\sum {\left(q_{mod}-\mu \right)}^3$ $Clusterprominence\left(cp\right)=\sum {\left(q_{mod}-\mu \right)}^4$
Spectral methods	Band power [19]	$F\left(s\right)={\sum }_{n=0}^{N-1}{x}_n{e}^{\frac{-2\pi }{N}sn},s=0,1,\dots ,N-1$ $Pow\left(s\right)={\int }_{f_{l}ow}^{f_{h}igh}F{\left(s\right)}^{2}ds$
	Spectral Entropy [26]	$SH=-{\sum }_{f1}^{f2}\widehat{P}\left(f\right)log\left(\widehat{P}\left(f\right)\right)$ $\widehat{P}\left(f\right)=P\left(f\right)/{\sum }_{f1}^{f2}P\left(f\right)$, P(f) is PSD of signal
	Spectral statistical Features [19]	Mean Peak Frequency, Mean Power, Variance of Central Frequency etc.
Time-frequency Methods	STFT [28]	$S(m,k)={\sum }_{n=0}^{N-1}s(n+mN)w\left(n\right)e^{-j\frac{2\pi }{N}nk}$
	Wavelet transform [29]	${\psi }_{s,\tau }\left(t\right)=\frac{1}{\sqrt{s}}\psi \left(\frac{t-\tau }{s}\right)$
	EMD [30]	$x\left(t\right)={\sum }_{i=1}^n{c}_i\left(t\right)+r_n\left(t\right)$
Spatial Methods	CSP [31]	$J\left(w\right)=\frac{w^T{C}_{1}w}{w^T{C}_{2}w}$
	BSS [32,33]	$x\left(t\right)=As\left(t\right)$ $s^{\prime }\left(t\right)=Bx\left(t\right)$ Approaches like ICA, CCD estimate $s^{\prime }\left(t\right)$
Spatio-temporal methods	Sample covariance matrices [34]	$C_i=\frac{X_i{X}_i^T}{tr\left(X_i{X}_i^T\right)}$ Where $C_i$ is covariance matrix of single trial