As for the PCA , the MCCA can be done on the epoched data (averaged ERP or single-trial epochs) or the average variance-covariance for each subject.
Note that due to the format of EEG data cutted to the reaction time. The method discards some data to have square matrices (a warning informs user how much data is kept)
Minor:
Cosmetic and few fixes on the pca computation and argument choice
Added option to compute components based on MCCA instead of PCA.
The implementation is based on https://github.com/ANCPLabOldenburg/MCCA/blob/main/MCCA.py (see https://www.sciencedirect.com/science/article/pii/S1053811923002252).
As for the PCA , the MCCA can be done on the epoched data (averaged ERP or single-trial epochs) or the average variance-covariance for each subject. Note that due to the format of EEG data cutted to the reaction time. The method discards some data to have square matrices (a warning informs user how much data is kept)
Minor: