The role of dopaminergic and cholinergic modulation on the striatal network : a computational investigation

Abstract

The famous words from the French philosopher René Descartes (1596-1650), “I think therefore I am”, proclaims that since we are thinking we must also exist. At the time when this was stated, very little was known about the main organ involved in thinking, the nervous system. Today we know that the nervous system consists of interconnected cells, so called neurons that communicate with each other through electro-chemical signals. This has been known for little over a century and during this time we have gathered an impressive amount of detailed data on neurons and the circuits they make up. Despite this, we still don’t have a detailed description of the overall computing mechanism of the central nervous system, the brain, or even single nuclei within the brain. One reason for this is the transient nature of the brain, continuously going in and out of operational modes, or so called brain states. The state of the brain is heavily influenced by neuromodulators – molecules changing the properties of neurons and the connections between them. One area strongly affected by neuromodulators is the striatum, the main input structure of the basal ganglia.

The basal ganglia are an evolutionary conserved set of interconnected nuclei tightly connected to the cerebral cortex and thalamus, with which they form a loop. From pathological states like Parkinson’s disease we know that the basal ganglia are involved in motor control. More specifically they have been proposed to drive formation and control of automatic motor response sequences (including habits), but like in the rest of the brain, the modus operandi of the basal ganglia is not known. To bridge the gap between data and function we therefore need models and testable theories.

In this thesis I have studied the role of neuromodulation in the striatal microcircuit, with the aim of understanding how subcellular changes affect cellular behavior. The technique used is biophysically detailed computational modelling. The essence of these models tries to mimic the electro-chemical signals within and between neurons using as detailed a description of individual neurons as possible. From this standpoint a good model minimizes the number of assumptions used in construction, by restricting the model to experimentally measured entities. Simulations of the striatal projection neurons in such models show that complex spikes – a particular type of neuronal signal associated with learning in other brain regions – may be triggered following manipulation of certain conductances in the cell membrane. In our simulations, the complex spikes were associated with large calcium signals in the dendrites, indicating a more robust form of crosstalk in the soma-to-dendrites direction than following regular action potentials. Together these simulations extend the theory of striatal function and learning.