CNN-LSTM to classify EEG signals based on motor imagery. Morlet Wavelet Transform was used for preprocessing as outlined in Construction of a Morlet Wavelet Power Spectrum for feature extraction of EEG Signals. Emotion Recognition based on EEG using LSTM was referenced for the construction of the model.
The EEG data represents what letter of the alphabet the subject was thinking of drawing. The data has 801 time points and 24 channels.
Dataset of EEG data can be downloaded here