Abstract
In a typical multiple object tracking experiment, observers are asked to track 4 of 8 identical objects as they move independently and unpredictably for several seconds. The striking and robust finding of these experiments is that observers can successfully track 3–5 objects under these conditions. Moreover, tracking is not disrupted when the objects move behind occluders. How is this accomplished? One possibility is that the visual system actively tracks invisible occluded targets. An alternative explanation is that the system is able to rapidly recover the missing targets when they reappear in plain view. Behavioral studies have addressed this question by measuring RT to transient luminance probes. However, this method changes the attentional state of the observer. In the current study, we tried to distinguish between these two explanations by looking at the patterns of neural activity revealed by ERPs. Last year, we demonstrated that the amplitude of sustained contralateral ERP activity is highly sensitive to the number of objects that an observer is currently tracking (Drew & Vogel, VSS 2006). This amplitude provides a convenient online measure of the number of object representations that are currently active on a given trial. Here, we found that the amplitude of this contralateral tracking activity is not disrupted for objects that are occluded during the trial. These results suggest that object representations do indeed remain active for tracked items that fall from view as they pass behind other occluding objects.