Abstract
Stimuli which signal upcoming reward have extensive influence on visual selective attention and the processing of a visual scene. Reward expectation can draw visual attention in a scene even when this is irrelevant or counterproductive for the observer's actual task. The present experimental investigation examined the influence of reward-predicting stimuli on learning of repeated context regularities when these stimuli were properties of a visual scene. In a contextual cueing task reward-signaling colors were assigned to three different reward magnitudes, with one color present in every search display, which consistently predicted the subsequent reward outcome. We assumed that reward magnitude would affect subsequent learning of context configurations and would favor learning of repeated relative to novel context configurations. Participants performed a contextual cueing task in two sessions on separate days which allowed examining long-term effects. Results revealed an acceleration of contextual cueing, i.e., faster responses to targets presented in repeated compared to novel contextual configurations, an effect that was most pronounced when contexts were associated with high reward relative to medium or low reward magnitudes. In addition to target response times also the analysis of eye movements revealed that the first fixation was closer to the respective target location in repeated contexts associated with high reward. Differences in response times between novel and repeated context configurations were found for both experimental sessions, although the contextual cueing effect was less pronounced in the second session. As the colors signaling reward magnitudes were neither response relevant nor predictive of the target location, we concluded that reward expectation facilitated context learning due to prioritized processing of contexts associated with high reward. This influence manifested also in early modulations of eye movements emphasizing the crucial role of attention guidance for the modulation of emerging context learning.
Meeting abstract presented at VSS 2018