关键词: 多身份追踪, 身份交换, 身份−位置绑定, 部分报告法, 语音复述

Abstract:  The Model of Multiple Identity Tracking (MOMIT) proposed by Oksama and Hyönä (2008) models observers’ tracking performance among multiple unique moving items. The model provides a functional explanation for the process for how the “what” and “where” information work together in a dynamic visual environment. However, two main issues in MOMIT are still unclear. First, though MOMIT supposes that the “what” and “where” information can be analyzed separately by the identity processing system and location processing system in the early stage, it is unclear whether these two processing systems share the same attentional resources. Second, MOMIT posits that the what-where bindings are stored in the temporary episodic buffer, but there is no direct evidence of this. Exploring these two issues may help us understand the cognitive mechanism of multiple identity tracking (MIT) better and improve the MOMIT. In this study, we used a variant of the MIT paradigm in which we interfered with the what-where binding by making the objects switch identities during tracking. In Experiment 1, we designated three identity-switch conditions: identity switch within the set of targets, identity switch within the set of distractors, and identity switch within all objects. And in the baseline condition the objects’ identities did not change throughout tracking. Given the limitation of the whole report method, Experiment 2 then tested the hypothesis again using the partial report method. The numbers of 0-9 were used as the objects’ identities in both experiments. The results of Experiment 1 showed that identity switching impaired both location tracking and identity recognition. Specifically, the location tracking and identity recognition was impaired the most in the condition where identities switch for all objects, followed by the condition in which the identities of the targets switched, and then the condition in which the identities of the distractors switched, which was not significantly different from the baseline condition. In addition, this declining trend was the same when participants had to track 4 targets and 5 targets. In other words, the increase of tracking load diminished people’s capacity to track location and identity recognition. The results of Experiment 2 showed the same effect due of interfering with what-where bindings. We also found that the partial report method revealed more data than the whole report method. Finally, in Experiment 3, we completely randomized pronunciations of the letters A-Z in the auditory channel in order to eliminate the interference of phonetic rehearsal. The results were the same as in Experiment 1. That is, the phonetic rehearsal did not affect the effects of identity-location binding. Overall, the results provide deeper understanding of MIT and improve the MOMIT through direct behavioral evidence. (1) The results reveal that the location processing system and identity processing system share a common attention resource pool, and the utilization of “where” information in the visual system seems to take precedence over “what” information. (2) The impairment of what-where binding will damage the tracking performance of MIT. (3) People mainly use attentional resources to enhance visual resolution towards targets (target-oriented) in MIT, rather than processing distractors. (4) The whole report method is less sensitive than the partial report method and may underestimate the capacity of visual working memory. (5) Even after controlling for phonetic rehearsal, people still experience interference from identity-location binding when they are tracking multiple moving objects.

Key words: Multiple Identity Tracking, identity switching, what-where binding, partial report method, phonetic rehearsal