Abstract
Many everyday-life situations require to track multiple visual object simultaneously. This encompasses situations in which it is sufficient to monitor the locations of objects (e.g., avoiding collisions in traffic) as well as situations in which the identity of the tracked objects is highly relevant (e.g. monitoring your own children on a playground). Here, we use an eye-tracking approach to study whether the process of location tracking (multiple object tracking; MOT) differs from the process of tracking of identities (multiple identity tracking; MIT). We take advantage of the observation that observers fail to distinguish repeated MOT trials, but their eye movements are similar between repeated presentations. In two experiments, we compared the gaze similarity in pairs of (1) repeated MOT trials, (2) repeated MIT trials and (3) repeated MOT-MIT trials (order counterbalanced). We evaluated the similarity using Pearson correlations for spatiotemporal scanpatterns smoothed with a Gaussian filter. In Experiment 1 (N=20), the tasks were presented in separate blocks, whereas they were intermixed in Experiment 2 (N=20). In both experiments, we consistently observed that the gaze patterns were most similar in repeated MOT trials (Exp.1: r = .443; Exp.2: r = .395). The similarity was lower when we compared repeated MIT trials (r = .333; r = .276) or repeated MOT-MIT trials (r = .305 or r = .251). In Exp.1, the similarity was higher if the participant started with the MOT trials than when they started with the MIT trials. Nevertheless, the intermixed trials in Exp.2 revealed analogous results. The results confirm the gaze patterns in repeated MOT trials are similar. Identical trials presented first as MOT and later as MIT (or vice versa) yield more different gaze patterns suggesting that MOT and MIT rely on distinct processes with the MOT process being more consistent across trials than the MIT process.