Proefschrift_Holstein

Chapter 8

aspects of flexible control by analysis of the types of errors: Animals may fail to shift (i.e. they persevere on the old rule) or they may exhibit deficits in rule maintenance (i.e. they initially shift, but then fail to maintain this new strategy). One draw-back, when trying to compare the results obtained with this paradigm to those observed with the human cued task-switching paradigm is that switches are not signalled by a cue, but need to be learned. In addition, set- shifting paradigms only allow one switch per test, rather than testing the fast trial-by-trial adaptation generally measured in humans. Extensive studies on strategy set-shifting have revealed a prominent role for the rodent medial PFC, in particular the prelimbic cortex (Ragozzino et al., 1999; Floresco et al., 2006b; Floresco et al., 2008a). In addition, dopamine signalling plays an important role in successful shifting behaviour (Floresco et al., 2006b; Haluk and Floresco, 2009), as does the dorsal striatum (Ragozzino et al., 2002). Human cued task switching work generally reports neural responses in the dorsal striatum and dorsal prefrontal cortex (inferior frontal gyrus) and a role for dopamine (Crone et al., 2006; Stelzel et al., 2010; Aarts et al., 2012; Aarts et al., 2014a), (see Derrfuss et al., 2005) for a meta-analysis. An interesting open question relates to if and how strategy set-shifting differs from cued task switching. A particularly important distinction resides in the learning component. Whereas the strategy set-shifting paradigm assesses how well animals learn to change their behaviour; cued switching requires the ability to quickly change task-sets. Given the role for the nucleus accumbens in learning (Schultz et al., 1997), it is perhaps not surprising that the rodent nucleus accumbens is crucial for certain aspects of successful strategy shifting (Floresco et al., 2006a). Interestingly, this nucleus is not involved in the initial shift, a role reserved for the dorsal striatum, but it is crucial to maintain a recently introduced novel rule. The first study that tested cued task switching in rodents (Baker and Ragozzino, 2014b, a) revealed that flexible behaviour was disrupted by lesions of the medial PFC and the basal ganglia, including the dorsal striatum. However, one major confound in this paradigm is caused by the occurrence of ~4 times more repeat than switch trials. As a consequence, switching between task-sets is no longer the only difference between switch and repeat trials. Instead, potential confounds are introduced (e.g. increased novelty or saliency of switch trials, and the overall expectation that a task will repeat). This confound is not present in the novel paradigm presented in chapter 6 where trials switch and repeat equally often. Together, the work in human subjects and in rodents both show an important role for dopamine, the dorsal striatum and corticostriatal circuit in task-switching, in particular the inferior frontal gyrus in humans and the prelimbic cortex in rodents (Robbins, 2007; Klanker et al., 2013), suggesting that these processes translate quite well across species. So far it is unclear how the role of the nucleus accumbens in strategy set-shifting (Floresco et al., 2006a) can be reconciled with its role in motivated cognitive control ( chapter 6 ). One possibility may be related to the role of this region in the approach of reward-related stimuli and its role in facilitating appropriate actions. The nucleus accumbens may play a role in how much updating takes place in prefrontal cortex, thereby mediating a balance between flexible

172