Table 2.

The summary of the time domain features used in EEG-based DDD.

Features		Mathematic Expression
Statistical measure	Maximum (Max) [80]	$\mathit{Max}=x_k\in \left\{x_1\dots x_N\right\}$
	Minimum (Min) [80]	$\mathit{Min}=x_k\in \left\{x_1\dots x_N\right\}$
	Standard deviation (STD) [76,81,82,83]	$\sigma =\sqrt{\frac{1}{N}{\displaystyle {\displaystyle \sum }_{k=1}^N}{(x_k-\mu )}^2}$
	Root mean square (RMS) [70]	$RMS=\sqrt{\frac{{{\displaystyle \sum }}_{k=1}^N{(x_k)}^2}{N}}$
	Integration [42,44,75,80]	$Integration={\displaystyle {\displaystyle \sum }_{k=1}^N}\left|x_k\right|$
The Number of Zero-Crossing (NZC) [42,46,70,80]		$NZC\left(x\right)={\displaystyle {\displaystyle \sum }_{k=1}^{N-1}}s\left(x_k,x_{k+1}\right)$ $s(x_k,x_{k+1})=\left\{\begin{array}{l}1\\ 0\end{array}\right.$ $\left.\begin{array}{c}if \left(x_k{x}_{k+1}\right)<0\\ if (x_k{x}_{k+1}\ge 0\end{array}\right\}$
Hjorth parameters [70]		$\mathrm{Activity}: Act=\frac{{{\displaystyle \sum }}_{k=1}^N{(x_k-\mu )}^2}{N}$ $\mathrm{Mobility}: Mob=\sqrt{\frac{\mathit{var}({\dot{x}}_k)}{\mathit{var}(x_k)}}$ $\mathrm{Complexity}: Com=\frac{Mob\left({\dot{x}}_k\right)}{Mob\left(x_k\right)}$
Barlow parameters [70]		$\mathrm{Absolute} \mathrm{Mean} \mathrm{Amplitude}: MA=\frac{{{\displaystyle \sum }}_{k=1}^N\left(\left|x_k\right|\right)}{N}$ $\mathrm{Mean} \mathrm{Frequency}: MF=\frac{\frac{1}{N}{{\displaystyle \sum }}_{k=1}^N\left|{\dot{x}}_k\right|}{MA}$ $\mathrm{Spectral} \mathrm{Purity} \mathrm{Index}: SPI=\frac{{{\displaystyle \sum }}_{k=1}^N\left|{\dot{x}}_k\right|}{{{\displaystyle \sum }}_{k=1}^N\left|{\ddot{x}}_k\right|}$
Auto regressive model coefficients(ARMC) [70]		$x_{k+1}=c+{\displaystyle {\displaystyle \sum }_{k=1}^N}{\varphi }_k{x}_k+{\epsilon }_{k+1}$ where c is the intercept and φ is ARMC which specifies how much the xk contributes to the xk+1. εk+1 is assumed to be the random zero mean noise.
Entropy	Shannon entropy (SE) [36,45,75]	$SE\left(x\right)=\frac{-{\displaystyle \sum }_{k=1}^{N}p\left(x_k\right)\times {\log }_{b}p\left(x_k\right)}{{\log }_{b}M}$ where p(xk) represents the probability that the xk occurs anywhere in the EEG samples x. The p(xk) is estimated by a histogram method where the x is linearly divided into M bins.
	R’enyi entropy (RE) [37,45,60]	$H_{\alpha }\left(x\right)=\frac{1}{1-\alpha }\log ({\displaystyle {\displaystyle \sum }_{k=1}^N}p{(x_k)}^{\alpha })$ where α is the order, α ≥ 0 and α ≠ 1
Mean comparison test (MCT) [61]		$M\left(i\right)=\frac{{\mu }_1-{\mu }_2\left(i\right)}{\sqrt{\frac{{\sigma }_1^2}{t_1}+\frac{{\sigma }_2^2\left(i\right)}{t_2}}}$ where μ1 indicates the fixed reference window, μ2(i) indicates the ith dynamic window.
Mahalanobis Distance (MD) [43]		$M\left(x\right)=\sqrt{{(x-\mu )}^T{S}^{-1}\left(x-\mu \right)}$