Individual Differences in Dopaminergic Modulation of Exploration-Exploitation Behaviour

Abstract

Dopamine (DA) has been implicated in exploration-exploitation behaviour, i.e., exploring novel, potentially better options vs. exploiting known, previously rewarding options. Impairments in this trade-off occur in psychiatric disorders involving DAergic dysfunction, including addiction and schizophrenia. Pharmacological studies revealed a contribution of DA to exploration, but inconsistent findings suggest that interindividual variability in baseline DA may modulate effects. To address this, we investigated the effects of the DA precursor L-DOPA on exploration-exploitation during reinforcement learning in a sample of N = 75 healthy participants (n = 32 women), following a randomised, double-blind, placebo-controlled, pre-registered design (https://osf.io/p2r7u). We assessed whether putative baseline DA markers, including spontaneous eye blink rate, working memory (WM) capacity, and impulsivity, modulated drug effects and probed visual fixation patterns and pupil dilation as markers of exploration. L-DOPA had no overall effect on computational model parameters of random exploration, directed exploration or choice perseveration. WM capacity moderated drug effects on random exploration, with stronger effects at higher WM capacity. Remaining DA proxies showed no credible effects. Pooling the data from male participants with that from an earlier male-only study (Chakroun et al., 2020; total N = 74), L-DOPA increased uncertainty-dependent value weighting and perseveration strength, while decreasing habit updating, indicating a stronger tendency to repeat previous choices and slower decay of their influence over time. No credible drug effects were observed in female participants. Pupil dilation was tonically increased under L-DOPA and scaled with exploration behaviour and prediction error, confirming that pupillometry can index exploration-exploitation dynamics. Visual exploration patterns reflected uncertainty-driven sampling, but were unaffected by L-DOPA. Taken together, results suggest that DAergic modulation of exploration and perseveration behaviour may be contingent on cognitive capacity and sex, rather than exerting uniform effects across individuals.