Humans automatically detect events that, in deviating from their expectations, may signal prediction failure and a need to reorient behaviour. The pupil dilation response (PDR) to violations has been associated with subcortical signals of arousal and prediction resetting. However, it is unclear how the context in which a deviant occurs affects the size of the PDR. Using ecological musical stimuli that we characterised using a computational model, we showed that the PDR to pitch deviants is sensitive to contextual uncertainty (quantified as entropy), whereby the PDR was greater in low than high entropy contexts. The PDR was also positively correlated with unexpectedness of notes. No effects of music expertise were found, suggesting a ceiling effect due to enculturation. These results show that the same sudden environmental change can lead to differing arousal levels depending on contextual factors, providing evidence for a sensitivity of the PDR to long-term context.

Reinforcement learning algorithms have a long-standing success story in explaining the dynamics of instrumental conditioning in humans and other species. While normative reinforcement learning models are critically dependent on external feedback, recent findings in the field of perceptual learning point to a crucial role of internally-generated reinforcement signals based on subjective confidence, when external feedback is not available. Here, we investigated the existence of such confidence-based learning signals in a key domain of reinforcement-based learning: instrumental conditioning. We conducted a value-based decision making experiment which included phases with and without external feedback and in which participants reported their confidence in addition to choices. Behaviorally, we found signatures of self-reinforcement in phases without feedback, reflected in an increase of subjective confidence and choice consistency. To clarify the mechanistic role of confidence in value-based learning, we compared a family of confidence-based learning models with more standard models predicting either no change in value estimates or a devaluation over time when no external reward is provided. We found that confidence-based models indeed outperformed these reference models, whereby the learning signal of the winning model was based on the prediction error between current confidence and a stimulus-unspecific average of previous confidence levels. Interestingly, individuals with more volatile reward-based value updates in the presence of feedback also showed more volatile confidence-based value updates when feedback was not available. Together, our results provide evidence that confidence-based learning signals affect instrumentally learned subjective values in the absence of external feedback.

Reinforcement learning algorithms have a long-standing success story in explaining the dynamics of instrumental conditioning in humans and other species. While normative reinforcement learning models are critically dependent on external feedback, recent findings in the field of perceptual learning point to a crucial role of internally-generated reinforcement signals based on subjective confidence, when external feedback is not available. Here, we investigated the existence of such confidence-based learning signals in a key domain of reinforcement-based learning: instrumental conditioning. We conducted a value-based decision making experiment which included phases with and without external feedback and in which participants reported their confidence in addition to choices. Behaviorally, we found signatures of self-reinforcement in phases without feedback, reflected in an increase of subjective confidence and choice consistency. To clarify the mechanistic role of confidence in value-based learning, we compared a family of confidence-based learning models with more standard models predicting either no change in value estimates or a devaluation over time when no external reward is provided. We found that confidence-based models indeed outperformed these reference models, whereby the learning signal of the winning model was based on the prediction error between current confidence and a stimulus-unspecific average of previous confidence levels. Interestingly, individuals with more volatile reward-based value updates in the presence of feedback also showed more volatile confidence-based value updates when feedback was not available. Together, our results provide evidence that confidence-based learning signals affect instrumentally learned subjective values in the absence of external feedback.