# Contribution of non-canonical dopamine pathways to model-based learning

> **NIH NIH R01** · UNIVERSITY OF CALIFORNIA LOS ANGELES · 2023 · $582,282

## Abstract

PROJECT SUMMARY
Model-based learning affords individuals the ability to contemplate the specific outcomes of actions or events. This
facilitates flexible decision making. While we know of brain regions that contribute to model-based learning, the
wider pathways and circuits that facilitate development of these flexible representations in these regions are less
explored. Given that substance use disorders are characterized by deficits in model-based decision making, a gap
in the knowledge of the neural circuits contributing to model-based learning prevents us from making clinical
advances in the treatment of these deficits. The overarching goal of this proposal is, thus, to expose the neural
circuits that mediate model-based decision making.
 Recent evidence from our team and others has implicated ventral tegmental area dopamine neurons (VTADA)
as critical to driving model-based learning. This was surprising because phasic VTADA activity was typically
restricted to assigning general value to cues, which prevents this signal from contributing to more flexible
associative relationships characterizing model-based learning. This work acts as our catalyst to investigate how
this dopamine signal is used in the circuits necessary for model-based learning. We are particularly interested in
the dopamine pathways to the basolateral amygdala (VTADABLA) and lateral hypothalamus (VTADALH). We
have shown that BLA and LH are important for the development of model-based associations. However, while the
BLA and LH both contribute to model-based learning about cues proximal to rewards, the function of these regions
diverge when it comes to more distal predictors. Specifically, the BLA remains important for using distal predictors
to predict rewards, while the LH opposes learning about distal predictors. It is unknown how VTADA projections to
BLA or LH facilitate reinforcement learning generally, or model-based learning specifically. Thus, we hypothesize
that midbrain dopamine projections to the BLA and LH mediate the encoding of detailed model-based associative
memories that allow prioritization of information most relevant to rewards.
 Capitalizing on the overlapping and complementary expertise and perspectives from two labs, we will uncover
the function of these two non-canonical dopamine circuits in model-based learning. We will use a symmetrical and
multifaceted approach using modern cell-type and projection-specific manipulation and recording techniques in the
context of sophistical behavioral tasks to reveal the function VTADA projections to BLA and LH in proximal and distal
learning. We will use cell-type and projection-specific optogenetic inhibition, stimulation, and recording of the
VTADABLA and VTADALH pathways to expose the role of these pathways. We will use next-generation
dopamine sensors to provide novel measurements of dopamine release in BLA and LH. Finally, we
chemogenetically inhibit VTADA projections to BLA or LH while optically imaging BLA...

## Key facts

- **NIH application ID:** 10607923
- **Project number:** 1R01DA057084-01A1
- **Recipient organization:** UNIVERSITY OF CALIFORNIA LOS ANGELES
- **Principal Investigator:** Melissa Sharpe
- **Activity code:** R01 (R01, R21, SBIR, etc.)
- **Funding institute:** NIH
- **Fiscal year:** 2023
- **Award amount:** $582,282
- **Award type:** 1
- **Project period:** 2023-04-15 → 2028-01-31

## Primary source

NIH RePORTER: https://reporter.nih.gov/project-details/10607923

## Citation

> US National Institutes of Health, RePORTER application 10607923, Contribution of non-canonical dopamine pathways to model-based learning (1R01DA057084-01A1). Retrieved via AI Analytics 2026-05-24 from https://api.ai-analytics.org/grant/nih/10607923. Licensed CC0.

---

*[NIH grants dataset](/datasets/nih-grants) · CC0 1.0*