Proefschrift_Holstein

Chapter 6

have established a role for the AcbC in helping animals to choose the best action based on prospective rewards. Disrupted functioning of the AcbC in these studies generally reduces the facilitation of appropriate responses, for example assessed with general Pavlovian to instrumental transfer (g-PIT) or cue-induced reinstatement (Floresco et al., 2008b; Corbit and Balleine, 2011). In a PIT procedure, animals receive instrumental- (learning to press a lever to obtain a food reward) and Pavlovian conditioning (learning that a cue predicts the delivery of an outcome), followed by a test on which the Pavlovian stimulus (CS) is delivered with the levers present. The presence of a CS increases lever pressing, i.e. the association between a cue and a reward facilitates instrumental performance, but only in animals with an intact AcbC (Corbit and Balleine, 2011). In addition to this general enhancement of reward- related cues, Pavlovian cues can also selectively increase performance on the basis of the specific outcome predicted by the cue. Animals with lesions of the AcbC are still able to show this specific PIT effect. In the current study, optimal behaviour would entail the facilitation of flexible control in the high reward context over the low reward context. This facilitation of flexible control under high reward was impaired after lesions of the AcbC. Also, lesions of the AcbC are known to reduce the inhibition of inappropriate actions, during outcome devaluation (Shiflett and Balleine, 2010) (Corbit et al., 2001), (but see de Borchgrave et al., 2002). After instrumental training on two levers which deliver two distinct outcomes, the value of one of the outcomes is reduced (e.g. by allowing animals to consume one of the outcomes freely). During a subsequent test, animals with an intact AcbC reduce responding to the lever which previously delivered the devalued outcome. After lesions of the AcbC however, animals no longer show this inhibition. In the current study, the suboptimal strategy would be to maximize cognitive control in the low reward context over the high reward context. In line with a role for the AcbC in the inhibition of inappropriate behaviour, we observed that animals with lesions of the AcbC exhibited more cognitive control in the low reward context than did animals with sham lesions. Combined, these results suggest that the AcbC is necessary to facilitate appropriate responses (e.g. when a CS predicts a reward) and to inhibit irrelevant responses (e.g. when the outcome is undesirable). Another field of research suggesting a role for the AcbC in facilitating and suppressing goals shows that animals without an intact AcbC are impaired when facing a situation that requires a shift in strategy (Floresco et al., 2006a). However, this deficit was clearly distinct from set-shifting deficits typically observed after lesions of cognitive control areas, such as the dorsomedial striatum (DMS) or the prelimbic cortex (PL). Animals with lesions in the DS and PL generally fail to make the initial shift (i.e. they persevere on the old rule (Ragozzino et al., 2002; Ragozzino, 2007). After lesions of the AcbC on the other hand, animals show no deficit on the initial switch. Instead, their deficit is characterized by an inability to eliminate irrelevant responses after they initially switch, i.e. animals occasionally go back to the previously correct (i.e. now incorrect) response. These studies fit well with a role for the AcbC in orienting behaviour to optimize cognitive control in order to obtain rewards (Floresco et al., 2006a). A failure to optimize cognitive control may result in inefficient facilitation of control in a low

132