Investigating striatal dopamine contributions to risk-taking during a mouse probabilistic decision-making task

Abstract

Midbrain projections from the ventral tegmental area and the substantia nigra pars compacta to the striatal regions are known to play distinct roles in behaviour. This mouse study used a calcium indicator (AAV1-CAG-flex-GCaMP6f) to measure dopaminergic axon activity in the nucleus accumbens core (NAcC) and the dorsolateral striatum (DLS) during a probabilistic decision-making task to compare their contributions to decision-making and reward processes. Specifically, the encoding of reward prediction error (RPE) was investigated in both regions through comparison of reward and omission responses across risky and reliable decision trials. Risky decisions offered a 65% chance of omission, a 17.5% chance of a jackpot reward (4x single), and 17.5% chance of a single reward. Reliable choices offered a 30% chance of omission but a 70% chance of a single reward. This probabilistic structure was designed to probe individual risk preferences, as although the risky choice was optimal for rewards, they were less consistently delivered. In the DLS and the NAcC, risky decisions elicited a significantly higher response to a single reward, suggesting these regions may encode positive RPE signals. Reward omission in the NAcC following a reliable decision elicited a significantly more negative response, providing further support for the encoding of negative RPE signals in this region. The findings of this study highlight distinct roles of the NAcC and DLS in a probabilistic risky decision-making task, with the NAcC reflecting encoding of reward value, uncertainty and RPE at outcome. In contrast, the DLS only reflects RPE partially, aligning with the idea that this area is more involved in movement and habitual behaviour. This study provides further insight into how dopamine signalling in the NAcC and DLS contribute to processes which may be involved in disorders of DA dysfunction, such as gambling disorder. As a pilot study implementing an adapted open-source behavioural device with a small sample size, findings should be interpreted cautiously. Replication in a larger sample with further evaluation of task contingency learning is needed to further confirm these observations.

Date of Award	17 Mar 2026
Original language	English
Awarding Institution	University of Bristol
Supervisor	Paul Dodson (Supervisor) & Johan Alsio (Supervisor)