Abstract
In this study we explored whether the presentation of an irrelevant object with previously acquired motivational value causes conflict for task-directed visual orienting. To assess visual orienting, we used a conventional spatial cuing paradigm in which a non-predictive cue (face stimulus) preceded target (simple circle) presentation by a short (100 ms) or long (1500 ms) interval; the task was to locate the target as quickly as possible (left or right). On half the trials (cued trials), cue location matched that of the target; on remaining trials (uncued trials), cue and target appeared at opposite locations. Prior to engaging in the spatial cuing task, participants learned to associate monetary reward or loss with different faces images by making repeated choices in an instrumental learning task. Both valence (win, loss) and probability (high, low) of outcome was manipulated in the learning task yielding different expected values for each stimulus (including zero, i.e., never yielding any outcome). These learned stimuli were then presented as cues in the spatial cuing task. For long cue-target intervals, the expected slowing of responding for cued versus uncued locations (inhibition of return) was found to be unmodulated by the expected value of the cue. However, for short cue-target intervals, we observed longer RTs when cues had non-zero expected value versus zero value (matched in familiarity), but only when targets appeared at the same location as the cue. Uncued RTs were unaffected. This location-dependent effect of cue expected value worked to abolish the typical speed advantage for targets appearing at cued locations with short cue-target lags. These results were interpreted within a framework of spatial interference, created by converging and overlapping neural representations of cue value and target location present in cued trials but not uncued trials.