Abstract
Purpose: In standard multiple-object tracking experiments, Os attend to specific objects (targets) while ignoring other objects (distractors). Typically, the targets are identified at the start of each trial. Then, all objects become identical and move about the screen. After some period of time, memory for target items is assessed. Real life is more continuous. You monitor two cars on the highway. Then a new car becomes relevant, an old one moves away, and so on. We wish to assess the ability to change the tracked set during a trial.
Method: In the critical condition, a set of identical items begins moving. Os are instructed to begin to track targets that briefly change color. They are instructed to stop tracking targets that are briefly marked with an X. This condition was compared with two others. In one condition, all targets were defined at the beginning of each trial as in a typical tracking task. In another, targets were added dynamically, once the items were moving. However, no targets were dropped during a trial. Total tracked items were never greater than four. The overall set size was eight objects. The tracking duration was twenty seconds.
Results: As soon as you see a demonstration of the display, you will know what the data must look like. Performance was very proficient in all conditions. The juggling (pick up and drop off) condition produced 88% accuracy compared to 90% when targets were added throughout the experiment and 87% accurate in the standard tracking experiment.
Conclusions: It is not necessary to establish the tracking set before the start of tracking nor even to have a fixed set. Os can proficiently add and delete items from the tracked set. In on-going experiments, Os appear to be able to do this for trials lasting many minutes.