Pavlovian Conditioning is Responsible for Value-Modulated Attentional Capture Unless This Learning Process is Disrupted and/or Information-Seeking is Encouraged.

Value-modulated attentional capture (VMAC) is when a stimulus attracts more attention beyond the attention it attracts from its physical salience as a result of its association with a valuable outcome. With some exceptions, most previous research into VMAC argues that it arises from Pavlovian conditioning between the stimulus and valuable outcome and is not based on information-seeking whereby the individual instrumentally attends to the stimulus to acquire information about the upcoming valuable outcome. The current experiments tested the roles of Pavlovian conditioning and information-seeking in VMAC using the paradigm developed by Le Pelley et al. (2015). In Experiment 1, participants were told explicitly about associations between colours and rewards before and throughout training to provide an opportunity for information-seeking and provided or not provided with feedback about reward outcomes on each trial to provide and exclude any opportunity for Pavlovian conditioning, respectively. We found equal evidence for VMAC in both groups, consistent with information-seeking rather than Pavlovian conditioning. Experiment 2 adopted a method by Failing and Theeuwes (2017) where participants were sometimes non-contingently satiated with information about the reward available on the upcoming trial to devalue this information. This manipulation reduced the evidence for VMAC, again consistent with an information-seeking explanation for the effect. However, Experiments 3 and 4, that replicated Experiment 2 but included a preliminary training phase before any information devaluation, established that VMAC responses can be insensitive to this devaluation, consistent with Pavlovian conditioning. Our findings suggest that either Pavlovian conditioning or information-seeking will give rise to VMAC depending on which learning process is provided an opportunity to take place first.