Notably, this was the case both for intact rats and for rats in which cholinergic signaling in the pDMS, an area previously shown to be necessary for goal-directed behavior (Yin et al., 2005), was disrupted via RG7204 molecular weight several different manipulations (Figure 1, left). This intact initial learning, and sensitivity to devaluation, demonstrated that behavior was indeed goal directed (i.e., guided by an expectation of the specific outcome and its unique attributes [Dickinson and Balleine, 1994]) and that intact cholinergic activity in the pDMS is not necessary for this fundamental

learning process. But what happens if the world changes? This was tested in a second phase of training, in which rats faced three challenges (Figure 1, top), each designed to test how changes in the associative structure of the environment would be incorporated into the earlier learning. The first involved contingency degradation—the outcome associated with one of the levers was presented for free, meaning that rats no longer needed to work to receive that reward. The second, reversal learning, involved switching the outcomes associated with each lever, followed by another devaluation test of the effects of satiety on responding. The third, extinction, involved removal of all outcomes for actions, followed by a “reinstatement

test” in which one of the outcomes was delivered to test whether it could reinstate pressing on the lever most recently associated with that JQ1 mw outcome. In each challenge, the critical question was whether rats would appropriately create new states in which to represent the new environmental contingencies. If so, each challenge should selectively affect responding on the lever most recently associated with the degraded, devalued or reinstated outcome. Any nonspecific effects on both levers would suggest that something had gone awry.

Endonuclease The results were amazingly clear-cut: in each case, intact rats exhibited selective effects on subsequent testing (Figure 1, middle), whereas rats in which cholinergic signaling in the pDMS had been disrupted showed intermediate or nonspecific effects on testing (Figure 1, bottom). Importantly, this same behavioral pattern was induced by bilateral lesions of the parafascicular nucleus, crossed lesions of the parafascicular nucleus and the pDMS, or a pharmacological manipulation that disrupted cholinergic signaling in the pDMS only during the learning phases. This exhaustive characterization of the phenomenon shows both that it is reliable and that it depends on cholinergic signaling at the time of learning, with the latter explicitly confirmed using an immunohistochemistry tool specific to cholinergic interneurons that was recently developed by the Balleine lab ( Bertran-Gonzalez et al., 2012). In addition, these results were shown to be specific to cholinergic disruption in the pDMS as they was not reproduced by manipulations of cholinergic function in the anterior portion of the dorsomedial striatum.