Abstract
In complex visual scenes, one often needs to process multiple objects. Dividing attention over even two simultaneous objects can be challenging if processing capacity is limited. The most extreme situation is all-or-none serial processing: observers process only one stimulus and guess about the other. This situation is rare, but has recently been demonstrated for the categorization of simultaneously presented and masked words (White et al., 2018). To understand whether this extreme effect generalizes to other kinds of stimuli, we asked how much divided attention limits performance in judgments of visual objects. In Experiment 1, we used a view-invariant 3D object recognition task; in Experiment 2, we used a semantic categorization task similar to that previously used for words. Stimuli appeared in a rapid serial visual presentation task presented above and below fixation, and observers were cued to attend to one (“single-task”) or both locations (“dual-task”). In Experiment 1, the stimuli were greyscale photographs of isolated 3D objects (Scharff et al., 2013); observers judged whether the cued object matched an object shown from a different viewpoint. In Experiment 2, the stimuli were grayscale photographs of isolated nameable objects; observers judged whether the cued object belonged to a target category (e.g. “animal”). To measure the effect of divided attention, we compared performance in the single- and dual-task conditions. Results from Experiment 1 showed a divided attention effect for the 3D object judgments. However, its magnitude was less than the all-or-none serial prediction, and instead consistent with a fixed-capacity parallel model. Preliminary results from Experiment 2 also showed an intermediate divided attention effect for the semantic categorization of objects. Thus, it appears that divided attention limits performance for judgments of simultaneous objects, but the magnitude of this effect is smaller than for words.
Acknowledgement: NIH R01 EY012925