g., win-stay) in choice tasks of this sort relate more to an explicit, working memory-based mechanism that may mask underlying 5-Fluoracil supplier incremental reinforcement learning ( Collins and Frank, 2012). Another possibility about how local adjustment and reversal might relate is that a deficit in learning from punishments (relative to reward) might exhibit itself as an apparently selective difficulty at reversal time,

when a cluster of negative feedback occurs. However, although this mechanism might predict dissociation between errors in reversal versus overall errors in initial acquisition, it does not seem to provide a good explanation of the observed pattern of DAT1 effects. This is because such a mechanism would couple the reversal deficit with global lose-shifting, and these are doubly dissociated by our DAT1 and SERT effects.

Accordingly, the EWA model also provided a better overall fit to choices than an alternative model involving differential learning from reward and punishments. An important interpretational caveat with the present study is that the task has only two, mutually exclusive response options, which makes it difficult to distinguish to what extent choice of either option relates to its own perceived strength versus the weakness HCS assay of the other. For instance, it may not be definitively possible to disentangle truly perseverative responding (in the sense of a sustained affirmative tendency most to seek the previously reinforced option) from impairment in acquiring or sustaining a response to the newly highly reinforced option. Nevertheless, the best-fitting model here suggests that DAT1-related perseveration occurs due to large, sustained value

on the previously favored option. Future studies should test this model using a task with a third option. Notwithstanding these finer distinctions, our finding relating DAT1 to reinforcement is in line with the conditioning literature suggesting that dopamine potentiates responding to cues previously associated with reward. Specifically, studies in rodents ( Goto and Grace, 2005 and Parkinson et al., 1999) have shown that enhanced levels of dopamine potentiate responding to previously rewarded stimuli. Furthermore, dopaminergic medication in patients with Parkinson’s disease has been shown to impair reversal learning ( Cools et al., 2001), possibly due to abnormal reward-related processing in the ventral striatum ( Cools et al., 2007). Interestingly, administration of the DAT blocker methylphenidate resulted in similar impairment in healthy volunteers depending on the degree to which the drug increased dopamine release ( Clatworthy et al., 2009). Thus, several lines of functional evidence have associated higher levels of dopamine with increased reward sensitivity and decreased behavioral flexibility.