Fig. 3: Learning indices (LI) for conjunction learning with global and feature-specific feedback (Experiments 1 and 2).
From: Visual perceptual learning of feature conjunctions leverages non-linear mixed selectivity
a Conjunction LIs across training blocks. The initial 144 trials of the learning phase were taken as baseline (Base). Thick lines are average LIs across subjects, light thin lines are individual subjects’ LIs. Global feedback (CLGF) in Experiment 1 resulted in more learning than feature-specific feedback (CLFF) in Experiment 2 (Experiment×Block F(6,210) = 4.5752, p = 0.0002, η2 = 0.1156) b Orientation LIs. These values were computed based on the accuracy in the orientation dimension during conjunction learning. Global feedback (CLGF) in Experiment 1 resulted in more learning than feature-specific feedback (CLFF) in Experiment 2 (Experiment×Block F(6,210) = 8.7148, p < 10−7, η2 = 0.1994) c Color LIs. These values were computed based on the accuracy in the color dimension during conjunction learning. Global feedback (CLGF) in Experiment 1 resulted in more learning than feature-specific feedback (CLFF) in Experiment 2 (Experiment × Block F(6,210) = 3.9228, p = 0.001, η2 = 0.1008). Error bars represent the standard error of the mean.
