135 / 218
133
Reward modulation of cognitive function: the nucleus accumbens
reward context, and the failure to allocate more control in a high than low reward context. We
demonstrate, for the first time, that reward plays a crucial role in this process, by showing that
manipulating the size of a reward can alter cognitive control in rodents, and that the AcbC is
crucial for maintaining the balance between suppressing irrelevant and facilitating relevant
goals.
Although not directly tested in the current study, we would like to elaborate on a potential
neural circuitry by which information about rewards can modulate cognitive and action
goals. Such a mechanism should allow information from the AcbC about which goals to
pursuit, to be conveyed to cognitive-control regions involved in flexible control i.e. the DMS
(Ragozzino, 2007). Previous work in human subjects has revealed that reward motivation can
enhance task-related signalling in the human homologue of the rodent DMS (i.e. the caudate
nucleus: (Aarts et al., 2010; Aarts et al., 2015). One mechanism perfectly suited to subserve
this interaction is via spiralling striato-nigro-striatal connections (Haber et al., 2000). These
dopaminergic midbrain connections allow information in the AcbC to be conveyed to more
dorsal regions of the striatum involved in goal-directed control (i.e. DMS) and actions and
habits (i.e. the dorsolateral striatum; DLS) (Balleine and O’Doherty, 2010). A functional role
for these connections has been shown previously by exploiting the knowledge that the AcbC is
involved in the acquisition of drug seeking, but that drug seeking is mediated by the DLS after
prolonged training (Belin and Everitt, 2008). Using an elegant design, these authors showed
that disconnecting the AcbC from the DLS impairs the transition to habits. Importantly, in
one hemisphere the AcbC remained intact (but its connection with the DLS was disrupted),
while in the other hemisphere the DLS was intact (but its dopaminergic input from the AcbC
was disrupted). Future work will have to reveal whether input from the AcbC to the DMS, or
signalling in the AcbC itself, is crucial for optimal motivation-cognition integration.
Using the rewarded task-switching paradigm in humans, we repeatedly observed dopamine-
dependent effects on behaviour and striatal responses (e.g. Aarts et al., 2010; Aarts et al.,
2015). In addition, we showed that a dopamine D2 receptor agonist did not alter behavioral
integration of reward and task-switching, but that it didmodulate task-switching performance
(irrespective of reward) (van Holstein et al., 2011). It is possible that dopamine D1 receptor
stimulation is modulating the effect of reward on task switching, also given the role for
dopamine D1 receptor stimulation in reward processing (Ikemoto et al., 1997; Meririnne et
al., 2001). Future work in rodents should test this, as D1 specific agents are not available for
research in human subjects.
Our previous observation that dopamine D2 receptor stimulationmodulates flexible cognitive
control is well in line with set-shifting work in rodents (Floresco et al., 2006b). However, when
comparing these task- switching and set- shifting studies, it is important to keep in mind that
they are conceptually different and that different neural systems may underlie these processes.
One important distinction to keep in mind is that the task-switching paradigm requires the
alternation between well-established task-sets, in which the striatum is involved (Cools et al.,