Talk: A Bayesian map of the striatal microciruitry reveals different wiring processes for D1 and D2…
Speaker: Francois Cinotti, University of Nottingham (grid.4563.4)
Title: A Bayesian map of the striatal microciruitry reveals different wiring processes for D1 and D2 spiny projection neurons
Emcee: Elenor Morgenroth
Backend host: Ali Rigby
Details: https://neuromatch.io/abstract?submis...
Presented during Neuromatch Conference 3.0, Oct 26-30, 2020.
Summary: Serving as input region of the basal ganglia, a set of interconnected midbrain structures thought to determine action selection, the striatum is mainly made up of spiny projection neurons (SPNs) which fall into two main categories based on the expression of either D1-type dopamine receptors or D2-type dopamine receptors. These distinct neuronal populations project into two different pathways towards the globus pallidus, but also make lateral connections which could constitute feedback inhibition of the circuit. In addition, interneurons in the striatum also connect to these SPNs. This complex microcircuitry is often overlooked in models of the basal ganglia which typically adopt a simple feedforward view of striatal function, when in fact, the connections between SPNs in the striatum could be crucial in determining their output to the globus pallidus. Fortunately for modellers, experimental studies to characterize these connections have been conducted and it seems that D1 neurons make fewer connections than D2 neurons and prefer to connect to other D2 neurons. However, the robustness of these results is undermined by the fact that when reporting these findings, experimenters failed to provide an estimate of the measurement error. We take the opportunity of correcting this oversight to introduce a novel Bayesian approach for estimating and comparing probabilities of connection. This method allows us to easily combine different data sets from different studies, to plug in the results of previous experiments in the form of a Bayesian prior, and to easily compare connection probabilities visually. By applying this method to two previous experimental studies, we find that D1 neurons do indeed form fewer connections than D2 neurons, but that evidence for preferential targeting by D1 neurons of other D1 neurons is currently lacking. We then go on to propose a method for extracting the probability of connection given distance between neurons.