When receiving a reward after a sequence of multiple events, how do we determine which event caused the reward? This problem, known as temporal credit assignment, can be difficult for humans to solve given the temporal uncertainty in the environment. Research to date has attempted to isolate dimensions of delay and reward during decision-making, but algorithmic solutions to temporal learning problems and the effect of uncertainty on learning remain underexplored. To further our understanding, we adapted a reward learning task that creates a temporal credit assignment problem by combining sequentially delayed rewards, intervening events, and varying uncertainty via the amount of information presented during feedback. Using computational modeling, two learning strategies were developed: an eligibility trace, whereby previously selected actions are updated as a function of the temporal sequence, and a tabular update, whereby only systematically related past actions (rather than unrelated intervening events) are updated. We hypothesized that reduced information uncertainty would correlate with increased use of the tabular strategy, given the model’s capacity to incorporate additional feedback information. Both models effectively learned the task, and predicted choices made by participants (N = 142) as well as specific behavioral signatures of credit assignment. Consistent with our hypothesis, the tabular model outperformed the eligibility model under low information uncertainty, as evidenced by more accurate predictions of participants’ behavior and an increase in tabular weight. These findings provide new insights into the mechanisms implemented by humans to solve temporal credit assignment and adapt their strategy in varying environments. <<<

According to the Bayesian framework, both our perceptual decisions and confidence about those decisions are based on the precision-weighted integration of prior expectations and incoming sensory information. While it is generally assumed that priors influence both decisions and confidence in the same way, previous work has found priors to have a stronger impact at the confidence level, challenging this assumption. However, these patterns were found for high-level probabilistic expectations that are flexibly induced in the task context. It remains unclear whether this generalizes to low-level perceptual priors that are naturally formed through long term exposure. Here we investigated human participants’ confidence in decisions made under the influence of a long-term perceptual prior: the slow-motion prior. Participants viewed tilted moving-line stimuli for which the slow-motion prior biases the perceived motion direction. On each trial, they made two consecutive motion direction decisions followed by a confidence decision. We contrasted two conditions – one in which the prior impacted discrimination performance, and one in which it did not. We found a confidence bias favoring the condition in which the prior influenced discrimination decisions, even after accounting for performance differences. Computational modeling revealed this effect to be best explained by confidence using the prior-congruent evidence as an additional cue, beyond the posterior evidence used in the perceptual decision. This is in agreement with a confirmatory confidence bias favoring evidence congruent with low-level perceptual priors, revealing that, in line with high-level expectations, even long-term priors have a greater influence on the metacognitive level than on perceptual decisions. <<<