Algorithm 1 Pipeline of EEG signal analysis

Input: EEG signal eeg(t).   procedure Acquisition (eeg(t)):      Sampling equipment selection.      Choose invasive or noninvasive acquisition.      Choose the best reference electrode.      Return: EEG data eeg(t).   procedure Denoising (eeg(t)):      Using different denoising methods to process EEG signals.      Return: Denoised EEG signal eeg${\left(\mathrm{t}\right)}_{new}$.   procedure Feature_Engineering (eeg(t)):      Perform time–frequency, high-order spectrum analysis or nonlinear analysis.      Return: More expressive features, eeg(t).   procedure Classification (eeg(t), choice):      Perform classification tasks based on choice:      If choice equal traditional approach do           Using KNN, SVM, ... to classify.      If choose the same deep learning approach do           Using CNN, GAN, ... to classify.      Return: Accuracy, precision, etc.