No task is accomplished by the use of only a single brain region. Areas of the brain communicate with each other in a flexible manner that allows for complex processing to occur. Over the last decade and a half functional brain imaging has proven to be an immensely popular technique in cognitive neuroscience. While this has led to enormous progress in understanding brain function, it has also led to an increasingly phrenological view of the brain, where a single brain region is proposed to be necessary for the experience of cognitive states, or the completion of tasks. Rather, processing is accomplished by a distributed network, much of which involves connectivity between cortical and subcortical regions of the brain. The purpose of this dissertation is to explore the interaction of cortical and subcortical regions and resulting effects on performance in three task domains.
In chapter 1 a novel association learning task is used to isolate learning from positive and negative reinforcement. When a low dose (1.25 mg) of bromocriptine, a dopamine (D2) receptor agonist was administered performance was impaired in learning from positive reinforcement, while there was a boost to performance in learning from negative reinforcement. It is suggested that this pattern of results on the two feedback valences is due to a reduction in phasic dopamine release due to presynaptic drug action. The effect of drug administration was further modulated by gender where males were much more affected by the drug. Individual differences in cortical and subcortical processes were examined using genetic data for two dopamine related genetic polymorphisms. Under placebo conditions, participants with the better functioning version (Val/Val) of the polymorphism associated with cortical dopamine function, COMT Val158Met, out performed the two other allele groups (Val/Met and Met/Met). Lastly, the main effect of drug administration was best predicted by the polymorphism associated with subcortical dopamine functioning, DRD2/ANKK1-TaqIa, where participants with a higher DA receptor density (A1-) had less of a drop in performance when learning from negative feedback after drug administration, than did (A1+) participants. Thus, performance in learning from positive and negative reinforcement, is at least to some extent, reliant on the interaction of cortical and subcortical dopaminergic systems.
In chapter 2 the same group of participants from experiment 1, also completed a working memory (WM) task. The WM task was completed on the same sessions as the tasks for experiment 1. Of particular interest in the WM task were three trial types: low load, where a minimal amount of information was to be held in WM; high load, where the amount of items to be remembered was increased; and filter, where some items were to be ignored. After administration of bromocriptine there was an unexpected slowing in reaction time (RT) for the easiest low load trials. Participants' RTs on the low load trials slowed to such an extent from drug administration that they were then responding more quickly to the more difficult high load and filter trials. When examining the polymorphism associated with cortical DA function, COMT Val158Met, under placebo conditions participants with the better functioning Val/Val allele, surprisingly, had superior performance when the number of items to be remembered was increased, than Val/Met or Met/Met participants. Bromocriptine administration resulted in improved load performance for participants with the lower functioning COMT allele, Met/Met. Finally, when examining the polymorphism associated with subcortical DA function, DRD2/ANKK1-Taq-IA, participants with the allele associated with lower receptor density, A1+, were the most susceptible to changes in RT from drug administration. In summary, ACC in working memory performance under increased load is best predicted by the cortical dopamine polymorphism, COMT Val158Met, while the polymorphism associated with subcortical dopamine function, DRD2/ANKK1-Taq-IA, predicts the extent to which dopamine modulation affects RT.
Lastly, in chapter 3 a visuomotor adaptation task was used to explore the role of primary motor cortex (M1) in the retention of a new sensorimotor transformation. Specifically the focus was on the role that success or error in a reaching movement plays on retention of the sensorimotor transformation in M1. On each trial participants made a reaching movement in a virtual environment while a perturbation was applied around the cursor, in this case in the form of a visuomotor rotation. To investigate M1s role in consolidation single pulses of transcranial magnetic stimulation (TMS) were delivered on specific trials. In experiment 3A feedback was only given about success in reaching for the target, not about the size of any error. Participants who received a TMS pulse on target hits failed to adjust their reach angles to compensate for the perturbation, regardless of pulse timing. Due to participants in the hit and delayed hit TMS groups not adjusting for the perturbation, the effect on consolidation was not able to be investigated. In experiment 3B participants were additionally shown the endpoint feedback of their reach. Under these conditions all participants adjusted to the perturbation, however, there were no differences in consolidation of the learning. This was true whether the TMS pulse was delivered on all trials, only on hit trials, or only on miss trials. Lastly, a direct replication was attempted by the addition of endpoint feedback during the retention test. Only a no TMS and an all trials TMS group were included. While weak evidence for impaired retention in the all trials TMS group was found, if present, the effect is much smaller than previously reported. The results in chapter 3 emphasize the importance of properly controlling for the mildly aversive nature of the TMS itself when using TMS to target specific trials. This leaves open the question of the role that success and error play in the consolidation of sensorimotor learning in M1.