Abstract
Stimuli matching visual working memory (VWM) representations are often attended. Effects consistent with this proposal have been observed in many tasks, such as RSVP, visual search, and probe detection. However, methodological limitations make it unclear whether the effect is automatic or the result of task-specific strategies. For example, some experiments required observers to hold an item in VWM and presented a matching item over the retention interval, but did not eliminate the demand to intentionally search for VWM matches. The current experiments took several precautions to address these demand characteristics and reduced the opportunity for verbal rehearsal while retaining the VWM task. Subjects performed articulatory suppression and held an object in VWM. Over the retention interval two objects were simultaneously flashed on the screen followed by a response probe. On critical trials, a VWM match and a novel item were flashed on the screen. If VWM matches attract attention, then probe responses should be faster when the probe appears at the location where the VWM match appeared. In an attempt to eliminate demand characteristics, only one-sixth of the trials were critical trials. On the remaining trials two task irrelevant objects were flashed. In Experiment 1, subjects who did not intentionally search for VWM matches over the retention interval showed a match-trial disadvantage. In Experiment 2, subjects were instructed to “let the display determine… [their] response”, in order to further reduce demand characteristics. We again found a match-trial disadvantage. These experiments suggest that VWM matches do not always attract attention.