Neurocircuitry controlling reward-directed behaviour in rats: Contribution of striatal sub-regions and prelimbic cortex

Rodent striatum is involved in sensory-motor transformations and reward-related learning, with lesion studies suggesting functional differences between striatal subregions. Dorsomedial striatum (DMS) is associated with goal-directed behaviour; dorsolateral striatum (DLS) mediates automated stimulus-response and nucleus accumbens (NAc) is involved in reward expectation. Corticostriatal communication from prelimbic cortex (PrL) to DMS and NAc likely modulates appetitive behaviour. The studies reported here investigated how specific elements of reward-related behaviour are maintained by striatum and cortico-striatal interaction. To better understand the functional significance of DLS sensory responses we developed a novel tactile discrimination task in head-fixed rats. Initial results using this task linked DLS sensory responses to either reward-expectation or motor-initiation but could not distinguish between the two. Next, to separate reward and motor components of striatal neural responses and to examine the role of cortico-striatal interaction, we developed a novel discrimination task requiring rats to either respond or suppress responding to reward-predicting cues. Neuronal responses in DLS, DMS, NAc and PrL were recorded during the discrimination task in overtrained rats. In both striatum and PrL, neuronal responses to cue-onset did not appear to be influenced by differences in reward expectation. However, responses in NAc and DMS showed a possible contribution from motor preparatory processes. Overall, striatal and PrL responses as well as synchronisation between striatal sub-regions and between PrL and striatal sub-regions were greater in error trials (false alarms and misses) than correct response trials (hits and correct rejections). Error responses during performance of an overtrained task may signal trials in which the animal tests the consistency of the learned stimulus response contingencies and thus engage striatal networks associated with goal-directed rather than habitual behaviour. The trial type-dependent differences in synchronisation between PrL and all three striatal subregions may indicate modulation from other brain areas or interactions between different cortico-striatal-thalamic circuits.