How do unique objects affect multiple-object tracking? Recent research has catalogued seemingly contradictory findings, varying from enhanced to impaired tracking performance. In this study, we explore the role of object complexity in this broad range of phenomena. In a series of three experiments, we demonstrate that unique objects of varying complexity can produce both costs and benefits on tracking performance. These experiments show that the key effects of uniqueness in object tracking are results of a tradeoff between tracking operation and processing of object identity information within the capacity limit of working memory.
The multiple-object tracking (MOT) paradigm has been widely used as a tool for studying attention and visual cognition (Scholl, 2009). Although most MOT tasks have employed identical objects as tracking stimuli, there is now a growing interest in visual tracking of unique objects (Horowitz et al., 2007, Makovski and Jiang, 2009a, Makovski and Jiang, 2009b and Ren et al., 2009). As most authors are well aware, accurate identity-location binding serves meaningful functions in many real-world tracking tasks. For example, a living organism has to identify which moving objects pose a potential danger; a basketball point guard has to be aware of where every teammate and opponent is in order to make a successful pass, etc. Given the ecological significance of object features for identity tracking, it seems rather unlikely that people would ignore this information in multiple object tracking. Intuitively, individuating objects should benefit tracking performance of MOT, because by remembering object identities at the target tagging phase, observers should be able to recover the lost targets during tracking. In other words, we may find “where” an object is by knowing “what” it is. However, one of the surprising conclusions from prior MOT research is that identity processing of the objects differentiated by features appears to be rather detached from tracking. For example, there is evidence that individuating objects by color does not help tracking performance (Klieger, Horowitz, & Wolfe, 2004). Moreover, people are often quite poor at recalling the identity of correctly tracked objects (Pylyshyn, 2004), or noticing when target properties change (Bahrami, 2003). These findings suggest that tracking is accomplished entirely by updating an object’s spatiotemporal information, and that the identity of the object is largely ignored during this operation. Pylyshyn, 1989, Pylyshyn, 2001 and Pylyshyn, 2004 model, Fingers of INSTantiation (FINST), was created to accommodate this line of findings. According to FINST, MOT is implemented by early vision that picks out a small number of objects while ignoring their visual properties. This early vision is “feature-blind” because the object identity differentiated by visual properties is not encoded or accessible for higher level cognitive processes.
In this study, we have shown that identity processing runs in parallel with tracking process regardless of whether object identities are task relevant. Our results show that individuating objects can impair, facilitate, or produce no difference on tracking performance. The results provide evidence that identity processing in MOT is to some extent involuntary under most tracking conditions. The direction of these effects is likely to be affected by the complexity of object information and the capacity limit of working memory.
