The Impact of Striatal Neuropeptides and Topography on Action Sequence Selection

Many common behaviours are a sequence of several actions. As action sequences are learned their activation often becomes habitual, allowing smooth, rapid, and semi-automatic execution; learning and performing action sequences is central to normal motor function. The striatum is the primary input nucleus for the basal ganglia and receives glutamatergic cortical afferents. These afferents innervate localised populations of medium spiny neurons (MSNs) and may encode 'action requests'. Striatal interactions ensure that only non-conflicting, high salience requests are selected, but the mechanisms enabling clean, rapid switching between sequential actions are poorly understood. Substance P (SP) and enkephalin are neuropeptides co-released with GABA by MSNs preferentially expressing D1 or D2 dopamine receptors respectively. SP facilitates subsequent glutamatergic inputs to target MSNs while enkephalin has an inhibitory effect. We construct models of these glutamatergic effects and integrate them into a basal ganglia model to demonstrate that diffuse neuropeptide connectivity enhances action selection. For action sequences with an ordinal structure, patterning SP connectivity to reflect this ordering enhances the selection of correctly–ordered actions and suppresses disordered selection. We also show that selectively pruning SP connections allows context–sensitive inhibition of specific undesirable requests that otherwise interfere with action group selection. We then construct a striatal microcircuit model with physical topography and show that inputs to this model generate oscillations in MSN spiking. Input salience and active neuronal density have differentiable impacts on oscillation amplitude and frequency, but the presence of oscillations has little effect on the mean MSN firing rate or action selection. Our model suggests that neuropeptide interactions enhance the contrast between selected and rejected action requests, and that patterned SP connectivity enhances the selection of ordered sequences. Our model further suggests that striatal topography does not directly impact action selection, but that evoked oscillations may represent an additional form of population coding that could bind together semantically related MSN groups.