Figure 5.

Categorization accuracy levels of all subjects on the posttraining explicit categorization task. Overall accuracy values reflect performance on both trained and novel stimuli. Novel stimuli accuracy values reflect performance on the novel hard category sounds that were heard during ERP recording. The two individuals scoring below the 50% chance level for overall accuracy are a subset of the five individuals scoring below chance level for novel stimuli accuracy; these five individuals were identified as “bad learners” for their failure to learn the categories robustly enough to generalize them to the novel, untrained MMN stimuli.