Cognitive and neural bases of salience-driven incidental learning

Abstract Humans incidentally adjust their behavioral strategies along feedback dimensions even without explicit reasons to do so. How it occurs may depend on stable individual preferences and contextual factors, such as visual salience. We here examined how task-irrelevant visual salience exerts influence on attention and valuation systems that further drives incidental learning. We first established the baseline behavior with no salience emphasis in Exp.1. We then highlighted either the utility or performance dimension of the chosen outcome using colors in Exp.2. We demonstrated that the difference in switching frequency increased along the salient dimension, confirming a strong salience effect. Moreover, the salience effect was abolished when directional information of feedback was removed in Exp.3, suggesting that the observed salience effect is specific to directional feedback. We then generalized our findings using text emphasis in Exp.4 and replicated the non-specific salience effects in Exp.5 with simultaneous eye-tracking. The fixation difference between chosen and unchosen values was enhanced along the feedback-specific salient dimension (Exp.4) but kept unchanged when removing directional information (Exp.5). Moreover, behavioral switching correlates with fixation difference, confirming that salience guides attention and further drives incidental learning. Lastly, our neuroimaging study (Exp.6) showed that the striatum subregions encoded salience-based outcome evaluation, while the vmPFC encoded salience-based behavioral adjustments. The connectivity of the vmPFC-ventral striatum accounted for individual differences in utility-driven, whereas the vmPFC-dmPFC for performance-driven behavioral adjustments. Our results provide a neurocognitive account of how task-irrelevant visual salience drives incidental learning by involving attention and the frontal-striatal valuation systems.