Dissociation between Striatal Regions while Learning to Categorize via Feedback and via Observation

Abstract

Abstract Convergent evidence from functional imaging and from neuropsychological studies of basal ganglia disorders indicates that the striatum is involved in learning to categorize visual stimuli with feedback. However, it is unclear which cognitive process or processes involved in categorization is or are responsible for striatal recruitment; different regions of the striatum have been linked to feedback processing and to acquisition of stimulus-category associations. We examined the effect of the presence of feedback during learning on striatal recruitment by comparing feedback learning with observational learning of an information integration task. In the feedback task, participants were shown a stimulus, made a button press response, and then received feedback as to whether they had made the correct response. In the observational task, participants were given the category label before the stimulus appeared and then made a button press indicating the correct category membership. A region-of-interest analysis was used to examine activity in three regions of the striatum: the head of the caudate, body and tail of the caudate, and the putamen. Activity in the left head of the caudate was modulated by the presence of feedback: The magnitude of activation change was greater during feedback learning than during observational learning. In contrast, the bilateral body and tail of the caudate and the putamen were active to a similar degree in both feedback and observational learning. This pattern of results supports a functional dissociation between regions of the striatum, such that the head of the caudate is involved in feedback processing, whereas the body and tail of the caudate and the putamen are involved in learning stimulus-category associations. The hippocampus was active bilaterally during both feedback and observational learning, indicating potential parallel involvement with the striatum in information integration category learning.