Fig. 6: Pupillary signal of subjective decision-values computed under the best-fitting model.

a Coefficient estimates from the pupillary OLS regression model correlating with self-reward values (i.e. the orange line and shading) (b) perceived emotions of the opponent which dynamically modulate the inequality term (c) the magnitude of self–other inequality. These panels illustrate the components of the subjective decision values. Computed under the best-fitting model as shown above the figure in a black frame (also in Eq. (9)). Pupil response to inequality peaks prior to decision onset (the window marked with dashed lines designating the decision reaction time (RT)) and indicates that pupil dilates more to offers associated with greater inequality (**p < 0.01, *p < 0.05). Error shading denotes ±1 SEM across all panels. All regressors were normalised ahead of model fitting to allow direct comparisons.