Abstract
Previous research in multiple-object tracking has provided evidence that the endogenous selection of targets to be tracked requires serial selection and engagement of attention (Pylysyn and Annan, 2006). However, this work was conducted entirely via manipulations of the static pre-trial cues. This raises the possibility that reallocation of attention during task performance is distinct from selecting targets from a static set. To explore this question, a task that required participants to add and remove objects from the currently tracked set was created. Dynamic modifications of the tracked set of objects were accomplished via color cues: previously monochromatic dots were flashed equiluminant red or green to cue that a target was to be removed or added to the tracked set, respectively. While the change in color produces an exogenous shift of attention, processing the target required an endogenous association of color with action to be taken. If this processing requires serial selection, then attention would be predicted to be diverted from the previous set and performance would suffer. In contrast, the results obtained via an ideal-observer analysis suggest that there was little to no penalty on overall performance due to processing the changes in the tracked set, and thus engaging and disengaging attention can happen rapidly and efficiently even when endogenous processing is required.
Consortium Research Fellowship, AFRL 711th HPW.