FIGURE 5.

(A) Individual decoding accuracy achieved by the classifier (mean across all folds per participant) using all the channels (“All,” 505 channels), and using the channels within (“Motor,” 372 channels) and outside (“No-Motor,” 133 channels) the motor-learning network, respectively. (B) Example of gamma-band activity (Participant #1) used as predictors to which SVM classifier was trained. Each row corresponds to one 0.3 s-long trial (i.e. observation), and each column corresponds to one feature (i.e. channel). The squares point-out features/channels outside the motor-learning network. The horizontal line separates the motor-trials (first-half) from the rest-trials (second-half). (C) Linear predictor coefficient β across brain areas. Per participant and brain area, the absolute β (rescaled [0–1]) value was extracted. It allowed exploring the contribution of the different brain areas in the decision function. The analysis was performed using the two set of channels (all the channels vs. the channels within the motor-learning network), separately.