Pavlovian reward learning elicits attentional capture by reward-associated stimuli

Abstract

Feature-reward association elicits value-driven attentional capture (VDAC) regardless of the task relevance of associated features. What are the necessary conditions for feature-reward associations in VDAC? Recent studies claim that VDAC is based on Pavlovian conditioning. In this study, we manipulated the temporal relationships among feature, response, and reward in reward learning to elucidate the necessary components of VDAC. We presented reward-associated features in a variety of locations in a flanker task to form a color-reward association (training phase) and then tested VDAC in a subsequent visual search task (test phase). In Experiment 1, we showed reward-associated features in a task display requiring response selection and observed VDAC, consistent with most previous studies. In Experiment 2, features presented at a fixation display before a task display also induced VDAC. Moreover, in Experiment 3, we reduced the time interval between features and rewards so that features appeared after a task display and we obtained marginally significant VDAC. However, no VDAC was observed when features and rewards were simultaneously presented in a feedback display in Experiments 4 and 5, suggesting that a direct association between feature and reward is not sufficient for VDAC. These results are in favor of the idea that response selection does not mediate feature-reward association in VDAC. Moreover, the evidence suggests that the time interval of feature and reward is flexible with some restriction in the learning of feature-reward association. The present study supports the hypothesis that theories of Pavlovian conditioning can account for feature-reward association in VDAC.

Cross-Modal Integration of Reward Value during Oculomotor Planning

“…This pattern of result is in line with a series of recent studies showing that associative reward learning can enhance the salience of rewarded stimuli and lead to valuedriven attentional capture, even in the following nonreward phase (Anderson et al, 2011;Theeuwes and Belopolsky, 2012;Yantis et al, 2012;Chelazzi et al, 2013;Hickey and van Zoest, 2013;Bucker and Theeuwes, 2018;Mine and Saiki, 2018). The impact of associative value during the nonreward phase depends on whether the previously rewarded stimuli serve as targets or distractors.…”

Dissociable Components of Experience-Driven Attention

“…Our findings are consistent with this distinction and further indicate that their influence is fundamentally dissociable. Our findings are also consistent with the idea that the effects of statistical learning on attention, which is related to history as a former target, are distinctly habit based [31], which can be contrasted with the mechanisms of Pavlovian reward learning hypothesized to underlie value-driven attention [26][27][28][29]. However, a direct comparison of these hypothesized mechanisms has been lacking.…”

“…When a stimulus predicts reward, the learning of this predictive relationship results in the stimulus gaining competitive priority in the visual system, biasing selection in its favor. This is in spite of the fact that participants were only ever rewarded for orienting away from the stimulus, providing powerful evidence for the role of associative learning in the orienting of attention [26][27][28][29]. On the other hand, repeatedly performing an orienting response in the presence of a particular stimulus facilitates the future In each experiment, three colors were used (previously highvalue/frequent, previously unrewarded/infrequent, and never presented during training).…”

Cross-modal integration of reward value during oculomotor planning

“…This pattern of result is in line with a series of recent studies showing that associative reward learning can enhance the salience of rewarded stimuli and lead to value-driven attentional capture, even in the following non-reward phase (Theeuwes and Belopolsky, 2012;Yantis et al, 2012;Chelazzi et al, 2013b;Hickey and van Zoest, 2013;Bucker and Theeuwes, 2018;Mine and Saiki, 2018). The impact of associative value during the non-reward phase depends on whether the previously rewarded stimuli serve as targets or distractors.…”