State Optimism as a State-Dependent Prior About Rewarding Outcomes

Abstract Optimism has recently been shown to have a state-like aspect that fluctuates within individuals. Given that optimism represents expectations for one’s own future, its intraindividual changes may interact with learning and decision-making. However, how optimism can be placed in a formal learning theory remains elusive. Here, we tested potential dynamic relationships between state optimism and learning by conducting behavioral tasks and questionnaire surveys pre- and post-intervention to increase optimism ( n = 95 for the intervention group and n = 41 for the control group), which has been shown to be effective for trait optimism in previous meta-analyses. After demonstrating that this intervention was also effective for state optimism, the following preregistered hypotheses were tested: intraindividual changes in state optimism (1) relate to positive belief updating bias and (2) relate to prior expectation about reward likelihood. While the results did not support hypothesis (1), hypothesis (2) was partially supported. Specifically, intraindividual changes in prior about reward, estimated through a reward learning task, showed a positive association with changes in state optimism, measured by questionnaire, in the intervention group. This result suggests a functional role of state optimism in value-based learning and decision-making: it defines a state-dependent general prior of rewarding outcomes, which can optimize the degree of exploration under that state.