Specialized coding of sensory, motor and cognitive variables in VTA dopamine neurons

Abstract

There is increased appreciation that dopamine (DA) neurons in the midbrain respond not only to reward 1 and reward-predicting cues 1,2 , but also to other variables such as distance to reward 3 , movements 4-9 , and behavioral choices 10,11. Based on these findings, a major open question is how the responses to these diverse variables are organized across the population of DA neurons. In other words, do individual DA neurons multiplex multiple variables, or are subsets of neurons specialized in encoding specific behavioral variables? The reason that this fundamental question has been difficult to resolve is that recordings from large populations of individual DA neurons have not been performed in a behavioral task with sufficient complexity to examine these diverse variables simultaneously. To address this gap, we used 2-photon calcium imaging through an implanted lens to record activity of >300 midbrain DA neurons in the ventral tegmental area (VTA) during a complex decision-making task. As mice navigated in a virtual reality (VR) environment, DA neurons encoded an array of sensory, motor, and cognitive variables. These responses were functionally clustered, such that subpopulations of neurons transmitted information about a subset of behavioral variables, in addition to encoding reward. These functional clusters were spatially organized, such that neighboring neurons were more likely to be part of the same cluster. Taken together with the topography between DA neurons and their projections, this specialization and anatomical organization may aid downstream circuits in correctly interpreting the wide range of signals transmitted by DA neurons. Users may view, print, copy, and download text and data-mine the content in such documents, for the purposes of academic research, subject always to the full Conditions of use:

A solution to the learning dilemma for recurrent networks of spiking neurons

“…We wondered whether e-prop can fill this void. As an example we consider the task that was studied in the experiments of Morcos and Harvey (2016) and Engelhard et al (2019). There a rodent learnt to run along a linear track in a virtual environment, where it encountered several visual cues on the left and right, see Fig.…”

“…they predict upcoming rewards in the case of dopamine or movement errors in the case of the error-related negativity (ERN), see MacLean et al (2015). Furthermore both dopamine signals (Engelhard et al, 2019;Roeper, 2013) and ERN-related neural firing (Sajad et al, 2019) are reported to be specific for a target population of neurons, rather than global. We refer to such top-down signals as learning signals in our learning model.…”

“…Solving a task with difficult temporal credit assignment by e-prop. A) Setup of corresponding rodent experiments ofMorcos and Harvey (2016) andEngelhard et al (2019), see Movie S2. B) Input spikes, internal spiking activity of 10 out of 50 sample LIF neurons and 10 out of 50 sample ALIF neurons, softmax output, sample learning signals and samples of slow components of eligibility traces in the bottom row.…”

Brain-wide electrical spatiotemporal dynamics encode reward anticipation

“…S2a-b). Thus, in contrast to prior studies that probed the role of individual brain regions or circuits within a putative anticipation network [5,15,[17][18][19][20][21], this approach allowed us to interrogate many circuit elements concurrently and to discover how they are integrated across space and multiple timescales from hundreds of milliseconds to seconds. Electrophysiological signals were recorded from regions comprising multiple cell types and projections distributions; thus, we chose ICA because this machine learning model allowed each region or circuit to contribute to multiple networks, as quantified by their learned feature weight within each network/component.…”

“…Pre-reward neural activation is potentiated in anticipation of greater potential rewards across several of these regions [5,7,14], a finding recapitulated by in vivo studies in animal models using electrical and neurochemical recordings of VTA and NAc [15][16][17]. These latter experiments, which allow for faster measurements than those possible using human imaging technology, have also revealed a ramping temporal profile of the anticipatory signals (i.e., neural signals were positively correlated with the progress of animals towards obtaining the reward) [15,[17][18][19][20][21][22][23].…”