Semantic Meaning Enhances Feature-Binding but not Quantity or Precision of Locations in Visual Working Memory

Recent studies showed that real-world items are better remembered in Visual Working Memory (VWM) than visually similar stimuli that are stripped of their semantic meaning. However, the exact nature of this advantage remains unclear. We used meaningful and meaningless stimuli in a location-reproduction VWM task. Employing a Mixture-modeling analysis, we examined whether semantic meaning enables more item locations to be remembered, whether it improves the precision of the locations stored in memory, or whether it improves binding between the specific items and their locations. Participants were presented with streams of 4 (Experiments 1 2) or 6 (Experiment 3) real-world items, or their scrambled, meaningless counterparts. Each item was presented at a unique location and the task was to reproduce one item’s location. Overall, location memory was consistently better for real-world items compared to their scrambled counterparts. Furthermore, the results revealed that participants were less likely to make swap errors for the meaningful items but there was no effect of conceptual meaning on the guess rate or the precision of the report. In line with previous findings, these results indicate that conceptual meaning enhances VWM for arbitrary stimulus properties such as item location, and this improvement is primarily due to a more efficient identity-location binding rather than an increase in the quantity or quality (precision) of the locations held in memory.