Proefschrift_Holstein

Reward modulation of cognitive function: the nucleus accumbens

Introduction When an animal is foraging for berries and suddenly comes across a nut tree (a cue) it may want to update its current goal and switch behaviour appropriately (from searching for berries to gathering nuts). Now consider that an animal is foraging in fall- or springtime, the expected reward will be higher in fall, and thus an animal may exert more control (e.g. be more flexible) to find blackberries and nuts in this season. Many everyday actions require flexible adaptation when environmental conditions change, and the use of cues can facilitate such adaptation. Cues can be differentially informative, with certain cues indicating which action will be rewarded, and others signifying the amount of reward to be received. It is well known that cognitive control processes are under the influence of reward motivation, allowing agents to select the most appropriate and beneficial course of action (Balleine and Dickinson, 1998; Locke and Braver, 2008; Padmala and Pessoa, 2010, 2011; Aarts et al., 2014b), (for reviews see: Pessoa and Engelmann, 2010; Aarts et al., 2011; Braver et al., 2014). Despite a well-established role for motivation in influencing cognitive control, surprisingly little is known about which neural mechanisms are crucial for such integration. Several functional neuroimaging studies in humans report increased neural activity in response to reward in regions typically involved in cognitive control, such as the prefrontal cortex, e.g. the inferior frontal gyrus (Locke and Braver, 2008) and the dorsal striatum (Aarts et al., 2010). Furthermore, several theories suggest a role for the human ventral striatum (VS) or rodent nucleus accumbens core (AcbC) in mediating a link between motivation and cognitive/ action control (Mogenson et al., 1980; Pessoa, 2009; Mannella et al., 2013; Floresco, 2015), suggesting that the VS/AcbC modulates the efficient pursuit of rewards or other goals in a constantly changing environment. However, direct evidence for a causal role of the VS/AcbC has thus far not been provided. Assessing whether there is a direct, causal role of the VS in integrating reward and cognitive control in humans is prevented by ethical andmethodological issues.The difficulty in assessing this in rodents, on the other hand, is that existing paradigms for measuring (rewarded) flexible control are conceptually different from the paradigms used in human studies, preventing direct cross-species comparison. Specifically, task-switching paradigms employed in humans typically require a trial-by-trial adaptation to task-sets in response to external cues (Meiran, 1996; Monsell, 2003). In contrast, rodent paradigms assessing flexible cognitive control include reversal learning, set-shifting, and extradimensional shift (EDS) paradigms, none of which involve the use of cues to initiate behavioural changes nor manipulate reward size, but instead assess an animals’ capacity to learn that the rule has changed (Birrell and Brown, 2000; Ragozzino et al., 2002; Floresco et al., 2008a; for a review see: Bizon et al., 2012). Although informative about aspects of flexible control, these paradigms ignore the more efficient process of using cues to switch actions (e.g. switching to foraging for nuts is more efficient in the presence of a nut tree than having to encounter several nuts before switching) and ignore the integration of motivational processes (which are generally held constant for such tasks).

119