Number of RFA selected features for procedures 3, 4 and 5.

The optimal subset of variables is represented for each subject by its average size (second and fourth columns) and standard deviation (third and fifth columns) across blocks for the exploration (second and third column) and mental imagery (fourth and fifth) sessions. The last line gives the mean and standard deviation across subjects. Results are presented in terms of mean and standard deviation across the number of features obtained after each cross-validation step.