Abstract: According to the biased competition model of Desimone and Duncan (1995), the contents of working memory (WM) may be crucial to resolve the competition for selection amongst different stimuli in the visual scene. Feedback from the item maintained in WM can strengthen the matching representation in early visual cortex, allowing it to win the competition for selection against the other non-matching representations. Previous studies have shown clear evidence that visual WM can automatically guide attention in some conditions, which can be established on both feature-based (e.g., Soto, Heinke, Humphreys, & Blanco, 2005) and dimension-based (e.g., Pan, Xu, & Soto, 2009) matching between memory and attention tasks. Moreover, Soto and Humphreys (2007) demonstrated that when the memory item was verbally represented, the feature-based effect could occur only when there were explicit memory requirements for observers. In other words, implicit verbal memory was not sufficient to automatically bias attention in favor of feature-matching stimuli in the field. In the current study, the author assessed the effects of explicit / implicit verbal memory on dimension-based visual selection. Here, the questions to be tested were that (1) whether the dimensional information held in verbal WM could guide attention to select the matching dimension of objects in the visual scene? and (2) if the answer was yes, then whether explicit memory of verbal stimuli was necessary for this dimension-based effect?
The present study included two experiments. Experiment 1 explored the effect of explicit verbal memory on visual selection. Twelve naive students participated for cash compensation, and all of them reported having normal or corrected-to-normal visual acuity and normal color vision. All participants were right-handed. The experiment was programmed using Presentation (Version 0.71), and was run on a Pentium IV computer with a 17-inch color monitor. Participants were asked to discriminate whether the colors (or the shapes) of the two objects simultaneously presented to the left and right sides of the screen (separated by approximately 6° from center to center) were the same or different during the retention interval of the verbal WM task. Experiment 2 was designed to explore the effect of implicit verbal memory on visual selection, and its method was very similar to that used in Experiment 1, except that a new group of nine volunteers participated. Here, a go/no-go procedure was used to make participants encode verbal stimuli into implicit memory system. Participants were instructed to withhold the response in the attention task on 25% of trials where a Chinese word referring to either “white” or “square” appeared.
The results showed a significant dimensional congruency effect in both Experiments 1 and 2. That is, regardless of whether verbal memory was explicit or implicit, response times in the color attention task on incongruent trials (i.e., the relevant dimensions of verbal memory and attention tasks were different) were reliably slower than on congruent trials. Moreover, though the size of congruency effect in Experiment 1 was a bit larger than in Experiment 2, this difference did not approach significance. However, there was no dimensional congruency effect when the relevant dimension of the attention task was shape, suggesting that the effect of dimension-based guidance of attention from verbal memory was not very strong and thus could be modulated.
The present findings extend previous work that showed the automatic guidance of visual attention to the objects’ features that matched the particular feature value maintained in verbal WM (Soto & Humphreys, 2007). Here, the author demonstrated that visual selection was involuntarily biased to the objects’ features that matched the specific dimension held in explicit / implicit verbal memory. Moreover, the current results also suggest that feature dimensions may be more difficult to be ignored than feature values (Pan, Xu, & Soto, 2009). Once dimensional information is processed, regardless of whether it is maintained in explicit or implicit memory, it will automatically guide attention in favor of the matching visual dimension in the scene.

Key words: verbal memory, feature dimension, visual selective attention, congruency effect