Abstract
Multiple object tracking (MOT) has been extensively studied in the laboratory. However, the relevance of this task to performance in the real-life environments is unknown. Here we examine whether the performance in laboratory tasks of MOT with identical objects (Pylyshyn & Storm, 1988) and that of multiple identity tracking (MIT) with distinct objects (Oksama & Hyönä, 2004) correlate with a real-life tracking performance, that is making a passing decision in soccer. Furthermore, we study how much information can be extracted in dynamic real-life environments. It is possible that previous laboratory results with artificial stimuli and novice participants underestimate the capacity in real-life tracking. It is conceivable that soccer experts can track players better than novices. To study these questions, we devised a task battery including a video-based soccer task measuring passing accuracy, a video-based realistic MIT task, a laboratory MIT task (moving faces of FC Barcelona players) and two laboratory MOT tasks with medium and fast speed. The task battery was presented to 27 experienced soccer players and 22 novices. Regression analyses showed that the MIT tasks proved to be significant predictors of the accuracy in making passing decisions, whereas, contrary to our expectations, MOT tasks did not show any significant correlation with the passing task. Moreover, experts performed better than novices in the realistic MIT task; they were able to track more player information than novices. However, no expert – novice –differences were found in the MOT tasks. These results suggest that the cognitive ability needed in the MIT tasks resembles closely abilities needed in the real-life tracking tasks but such resemblance was not found with MOT. The results are consistent with our previous findings that MIT and MOT are performed by separate systems in terms of eye-movements and brain activity (Oksama & Hyönä, 2016; Nummenmaa, Oksama, Glerean, & Hyönä, 2017).
Meeting abstract presented at VSS 2018