posted on 2021-06-23, 20:18authored bySatoshi Nishida, Antoine Blanc, Naoya Maeda, Masataka Kado, Shinji Nishimoto
The distribution of the maximum values of voxelwise prediction accuracy for individual participants was shown separately for each model (A and B: fastText; C and D: GloVe; E and F: word2vec; G and H: binary-labeling) and each dataset (A, C, E, and G: movie set 1; B, D, F, and H: movie set 2). The vertical dashed line in each panel indicates the mean value of the maximum prediction accuracy averaged across participants.