Abstract
Previous studies have shown that attention is drawn to the location of manipulable objects and is distributed across pairs of objects that are positioned for action. Here, we investigate whether central, action-related objects can cue attention to peripheral targets. Experiment 1 compared the effect of uninformative arrow and object cues on a letter discrimination task. Arrow cues led to spatial-cueing benefits across a range of stimulus onset asynchronies (SOAs: 0 ms, 120 ms, 400 ms), but object-cueing benefits were slow to build and were only significant at the 400-ms SOA. Similar results were found in Experiment 2, in which the targets were objects that could be either congruent or incongruent with the cue (e.g., screwdriver and screw versus screwdriver and glass). Cueing benefits were not influenced by the congruence between the cue and target, suggesting that the cueing effects reflected the action implied by the central object, not the interaction between the objects. For Experiment 3 participants decided whether the cue and target objects were related. Here, the interaction between congruent (but not incongruent) targets led to significant cueing/positioning benefits at all three SOAs. Reduced cueing benefits were obtained in all three experiments when the object cue did not portray a legitimate action (e.g., a bottle pointing towards an upper location, since a bottle cannot pour upwards), suggesting that it is the perceived action that is critical, rather than the structural properties of individual objects. The data suggest that affordance for action modulates the allocation of visual attention.
Keywords:
Notes
1 A reviewer expressed concerns that, when oriented towards the upper target locations, the jug not only portrays an implausible action but is also in a noncanonical orientation. To address this concern we repeated this analysis, excluding jug trials. We found the same pattern of results. Objects that portrayed a legitimate action in all directions produced cueing benefits of 20 ms (lower targets) and 26 ms (upper targets), while objects that only portrayed a legitimate action towards lower targets produced cueing benefits of 17 ms (lower targets), but only 4 ms (upper targets). However, with fewer trials the interaction was not statistically reliable, F(1, 51) = 1.76, p = .19.