Abstract: It has been a controversial issue for a long time that the hierarchical priority of scene gist processing and its demand for attention resources. The most previous researchers have used several paradigms to study this question, such as rapid visual categorization, dual-task paradigm, inattention blindness, eye movement paradigm, and Stroop paradigm and so on, but no unanimous conclusion was reached. However, there was no research using the attentional blink paradigm to explore this issue. As for the logic of the attentional blink paradigm, when two successively presented targets require attention resources to process, the shorter the time interval between them, the lower the recognition accuracy of the second target. Therefore, this study aimed to provide the attentional blink evidence for hierarchical processing priority of scene gist and its necessity to the attentional resources. In this study, the researchers asked participants to perform the face recognition task (T1) and scene gist recognition task (T2) at basic and superordinate levels. As shown in Figure 1, a trial consisted of 16 interference pictures and two target pictures appearing in rapid serial visual presentation (RSVP). There were 18 pictures in total, each picture presenting for 100 ms and then immediately being masked by the next picture. The first target stimulus was one colored face image. In order to control the preparedness effect of participants, T1 randomly appeared in the position of the fifth to the eighth image of the whole stimulus sequence. The second target stimulus was one scene image after T1. There are four-time intervals between T1 and T2, and T2 appeared in the 2nd, 3rd, 5th and 8th positions after T1, respectively. Hence, stimuli onset asynchrony (SOA) was 200 ms, 300 ms, 500 ms, and 800 ms, respectively. At the end of each trial, the participants were asked to answer two corresponding questions. Firstly, whether the T1 was consistent with the face presented in question 1 by pressing the different keys (M = consistent, Z = inconsistent), and secondly whether the scene gist of T2 is consistent with the categorical words in question 2 (M = consistent, Z = inconsistent). Participants were asked to give priority to the task of face recognition and ensure accuracy as much as possible. There was no limit on response time, and participants did not receive any feedback on task performance. As shown in Figure 2, the findings were in the following: (1) Compared with scene gist categorization at the basic level, the accuracy of scene gist categorization at superordinate level was higher, which suggested that superordinate level processing have priority over the basic level processing in the scene gist categorization. (2) There was no attentional blink effect in the superordinate level scene gist categorization, however, there was a significant attentional blink effect in the basic level scene gist categorization, which suggested that superordinate level scene gist categorization need less attentional resources, however, the basic level scene gist categorization needs many attentional resources. It can be concluded that this study provides evidence for the superiority of the superordinate level processing over the basic level processing in scene gist categorization from the attentional blink paradigm.

Key words: scene gist, hierarchical processing, attentional blink, superordinate level priority

摘要： 场景主旨加工的层级优先性及其对注意资源的需求一直备受关注。本研究采用注意瞬脱范式, 记录被试完成面孔再认任务和四种时间延迟条件下不同层级场景主旨分类任务的正确率。结果发现, 上级水平场景主旨分类任务的正确率没有出现注意瞬脱效应, 而基本水平的场景主旨分类任务的正确率出现了注意瞬脱效应, 前者高于后者; 并且当进行上级水平场景主旨分类任务时, 面孔再认的正确率显著高于进行基本水平场景主旨分类任务时的。由此可见, 场景主旨的层级水平调节场景主旨加工对注意资源的需求, 支持场景主旨加工具有上级水平优势的观点。

关键词: 场景主旨, 层级加工, 注意瞬脱, 上级水平优势