UC Irvine

Decades of research have revealed the remarkable complexity of the midbrain dopamine (DA) system, which comprises cells principally located in the ventral tegmental area (VTA) and substantia nigra pars compacta (SNc). Neither homogenous nor serving a singular function, the midbrain DA system is instead composed of distinct cell populations that (1) receive different sets of inputs, (2) project to separate forebrain sites, and (3) are characterized by unique transcriptional and physiological signatures. To appreciate how these differences relate to circuit function, we first need to understand the anatomical connectivity of unique DA pathways and how this connectivity relates to DA-dependent motivated behavior. We and others have provided detailed maps of the input-output relationships of several subpopulations of midbrain DA cells and explored the roles of these different cell populations in directing behavioral output. In this study, we analyze VTA inputs and outputs as a high dimensional dataset (10 outputs, 22 inputs), deploying computational techniques well-suited to finding interpretable patterns in such data. In addition to reinforcing our previous conclusion that the connectivity in the VTA is dependent on spatial organization, our analysis also uncovered a set of inputs elevated onto each projection-defined VTA^DA cell type. For example, VTA^DA→NAcLat cells receive preferential innervation from inputs in the basal ganglia, while VTA^DA→Amygdala cells preferentially receive inputs from populations sending a distributed input across the VTA, which happen to be regions associated with the brain's stress circuitry. In addition, VTA^DA→NAcMed cells receive ventromedially biased inputs including from the preoptic area, ventral pallidum, and laterodorsal tegmentum, while VTA^DA→mPFC cells are defined by dominant inputs from the habenula and dorsal raphe. We also go on to show that the biased input logic to the VTA^DA cells can be recapitulated using projection architecture in the ventral midbrain, reinforcing our finding that most input differences identified using rabies-based (RABV) circuit mapping reflect projection archetypes within the VTA.