Skip to content. | Skip to navigation

Personal tools
Document Actions

J.G. Bednark, E.A. Franz, Tom Stafford, Peter Redgrave, and N.J. Reynolds (2009)

Tracking the learning of actions: An evaluation of the frontal P3a component

In: Neuroscience Meeting Planner, Society for Neuroscience.

The ability to correctly identify and evaluate actions that result in an unexpected and/or behaviorally significant sensory event is crucial to an organism’s learning and survival. Theories regarding the functioning of the striatum propose that a phasic dopamine signal strengthens inputs that were active just prior to a sensory event. However, at the time of the dopamine signal the striatum is receiving a large number of inputs, not only those directly associated with the sensory event. Evaluating these various inputs is the anterior cingulate cortex (ACC). The ACC has been implicated in a variety of evaluative processes (e.g. performance monitoring, conflicted detection, reward evaluation). Typically, event-related potential components such as N2 or error-related and feedback-related negativities (ERN/FRN) are used to assess ACC activity during evaluation. Our novel hypothesis is that the P3a, a neural signal typically associated with novelty, also captures increased ACC activity during the learning of actions that have caused a sensory event. Specifically, we hypothesized that the P3a component will be present in early stages of action learning, but will decrease in amplitude as the cause of a sensory event is learned. To test this hypothesis, we used tasks that track the learning of actions leading to a sensory event. Specifically a ‘Where Task’ was employed in which participants had to discover, by exploring the screen with a cursor, the location that elicited a visual stimulus. After acquiring the visual stimulus, the cursor was repositioned and the participant had to move back to that location to re-elicit the visual stimulus. Data from the Where Task were compared to control tasks that did not require action discovery, but controlled for other potential confounds. EEG recordings were made during each of the tasks, along with cursor movements and search times as behavioral measures. Event-related potential techniques were used to analyze EEG data. The P3a component (maximal at FCz) was only elicited in the Where Task, confirmed by highly significant differences found between the Where Task and control tasks (p < 0.0001). No differences were found across control tasks (p > 0.1). Significant decreases in P3a amplitude were found when comparing block 1 to subsequent blocks of the Where Task (p < 0.05). Effects of learning were confirmed by behavioral analysis. These results support our novel proposal that the P3a is a measure of ACC activity probably modulated by dopamine, as an organism learns to correctly identify and evaluate the actions that result in an unexpected and/or behaviorally significant sensory event.
Humans; Learning; Mental Processes; Models; Psychological; Reinforcement (Psychology); 12020450