On striatum in silico

Abstract

The basal ganglia are a collection of subcortical nuclei involved in movement and action selection. The striatum is the main input nucleus with extensive projections from the cortex and thalamus, and dopaminergic projections from SNc and VTA. The two main cell types are the striatal projection neurons (SPNs), which are divided into the direct (dSPN) and indirect (iSPN) pathways, based on the downstream projections and the expression of dopamine D1 and D2 receptors, respectively. The remaining 5% consists mainly of GABAergic interneurons, such as parvalbumin-expressing fastspiking interneurons (FS) and low threshold spiking interneurons (LTS). The cholinergic interneuron (ChIN) is spontaneously active and unlike the other interneurons releases acetylcholine. This thesis is focused on investigating the function of the striatum and the role of SPNs and the striatal interneurons. This is achieved by building a platform, tools, and a database of multi-compartmental models of SPN, FS, ChIN, and LTS; and through simulations systematically uncovering the roles of these striatal neuron types and external input and, more specifically, the role of neuromodulation and intrastriatal inhibition.

In Paper I, Snudda, a platform for simulating large-scale networks, is developed and includes multicompartmental models of dSPN, iSPN, FS, LTS, and ChIN. The tools include methods to generate external input from the cortex and thalamus; and dopaminergic modulation from SNc. Paper II investigates the relationship between ChIN and LTS. The ChIN releases ACh, which activates both nicotinic and muscarinic receptors within the striatum. The dominating effect on LTS is inhibition caused by muscarinic M4 receptors. LTS, on the other hand, releases NO which excites ChINs. Paper II showed that the interaction between these neuromodulators could control the activity of ChIN and LTS, which are generally spontaneously active. In the subsequent Paper III, Snudda was complemented with the neuromodulation package called Neuromodcell, a Python Package, for creating models of neuromodulation, which can be included in large-scale network simulations in Snudda. The method of simulating neuromodulators in Snudda was expanded to include multiple simultaneously active modulators. This resulted in several simulations with simultaneous ACh pause with DA burst as well as an ACh burst with a DA burst. In Paper IV, the effect of intrastriatal surround inhibition on striatal activity was investigated by utilizing ablations, clustered input, dopaminergic modulation, and other features in Snudda. These simulations demonstrated that shunting inhibition could reduce the amplitude of corticostriatal input onto SPNs. The surround inhibition can further modulate the plateau potentials in SPNs, which is dependent on the GABA reversal. Lastly, the competition between populations of SPNs can be modified by varying the strength, size, and positions of populations. Furthermore, dopaminergic modulation can enhance the effect of dSPNs, while increasing the inhibition onto iSPNs.

Overall, this thesis provides an analysis of the striatal microcircuit and a tool for further investigations of the striatum in silico; and demonstrates the importance to consider the different components of the striatal microcircuit and how neuromodulators can reshape microcircuits on both single neuron and network levels.