EEGdenoiseNet

[scott-huberty]

This is the model I breifly mentioned in our last meeting: https://de.mathworks.com/help/signal/ug/ocular-artifact-removal-from-electroencephalogram-signals-using-deep-learning-regression.html

As compared to our WIP, There are some similarities (they use an LSTM architecture), and some differences (IIRC, they simulated EEG data with and without EOG artifact, and trained a regression model to take the the EEG signal (with EOG artifact) as input, and minimize the loss between the predicted clean EEG signal and the actual "clean" version of the EEG signal (i.e. the label).

Just leaving this here in case it is relevant to us down the road.