多可能目标条件下背景线索效应的眼动研究
发布时间：2018-10-31 16:01
【摘要】：Chun和Jiang(1998)研究发现在视觉搜索任务中,被试在布局不变的场景(重复场景)中搜索目标的速度快于在布局随机变化的场景(新异场景)中搜索目标。这一学习现象被称为背景线索效应。前人的研究大多是在单目标条件下(每个重复场景的目标·背景一一对应),通过行为实验的结果,对背景线索效应的机制进行推测和分析。然而,在日常生活中,目标与背景的关系往往是非一一对应的。本研究中采用的是多可能目标范式,该范式的优点是,每个重复场景所匹配的目标位置增加,如每个重复场景匹配2个或4个目标位置,同时又不增加总的目标数量。本研究中的多可能目标实验条件的结果与经典的单目标实验条件的结果具有可比性,从而进一步揭示背景线索效应的认知机制。当前,背景线索效应的发生机制存在三种观点:(1)注意引导加工机制;(2)反应选择加工机制;(3)注意引导和反应选择的双加工模型。基于观点(1),如果仅通过注意引导机制,那么眼动结果仅在搜索的中期阶段发现背景线索效应;基于观点(2),如果仅通过反应选择机制,那么眼动结果仅在搜索的晚期阶段发现背景线索效应;基于观点(3),如果注意引导和反应选择机制均发挥作用,那么眼动结果在搜索中期阶段和搜索晚期阶段均发现背景线索效应。本研究包括以下四个实验:实验一,为了验证前人在单目标条件下对背景线索效应及其眼动规律的探讨,并与多可能目标位置条件进行比较。实验一采用经典背景线索效应范式,以15名大学生为被试,实验材料为由12个分心字母L和1个目标字母T随机分布而组成的搜索场景。实验设计是2(场景类型:重复、新异)×7(时间窗口:1-7)的被试内设计。结果发现:(1)在反应时间上,场景类型和时间窗口的交互作用显著,随着不断学习,搜索重复场景的反应时显著短于新异场景,即被试获得了显著地背景线索效应;(2)在搜索早期阶段,重复场景显著短于新异场景;(3)在搜索中期阶段,重复场景显著短于新异场景,并且场景类型和时间窗口的交互作用显著,表现出显著的背景线索效应;(4)在搜索晚期阶段,重复场景显著长于新异场景;(5)在平均注视次数上,重复场景显著少于新异场景;(6)在平均注视时间上,重复场景显著长于新异场景;(7)在扫描模式比率上,重复场景显著小于新异场景。实验二,为了探究当一个重复场景匹配两个可能目标位置时,背景线索效应是否出现,以及被试在该搜索任务下的眼动规律。实验二采用两个可能目标范式,以15名大学生为被试。本实验的实验材料和实验设计与实验一相同。结果发现:(1)在反应时间上,场景类型和时间窗口的交互作用显著,即被试获得了显著地背景线索效应;(2)在搜索早期阶段,场景主效应和交互作用均不显著;(3)在搜索中期阶段,重复场景显著短于新异场景;(4)在搜索晚期阶段,重复场景显著长于新异场景;(5)在平均注视次数上,重复场景显著少于新异场景;(6)在平均注视时间上,重复场景显著长于新异场景;(7)在扫描模式比率上,重复场景显著小于新异场景。实验三,为了探究当一个重复场景匹配四个可能目标位置时,背景线索效应是否出现,以及被试在该搜索任务下的眼动规律。实验三采用四个可能目标范式,以14名大学生为被试。本实验的实验材料和实验设计与实验一相同。结果发现:(1)在反应时间上,场景类型和时间窗口的交互作用边缘显著,被试获得了背景线索效应;(2)在搜索早期阶段,重复场景显著短于新异场景;(3)在搜索中期阶段,重复场景显著短于新异场景;(4)在搜索晚期阶段,重复场景显著长于新异场景;(5)在平均注视次数上,重复场景显著少于新异场景;(6)在平均注视时间上,重复场景显著长于新异场景;(7)在扫描模式比率上,重复场景显著小于新异场景。实验四,为了探究当仅保留重复场景的背景熟悉性,而使场景中目标出现的位置随机时,背景线索效应如何表现,以及搜索的眼动规律。实验四采用随机目标位置条件,以14名大学生为被试。本实验的实验材料和实验设计与实验一相同。结果发现:(1)在反应时间上,场景类型和时间窗口的交互作用不显著,被试未表现出背景线索效应;(2)在搜索早期阶段和搜索晚期阶段,场景类型主效应不显著、场景类型×时间窗口的交互作用不显著;(3)在搜索中期阶段,重复场景显著短于新异场景;(4)在平均注视次数上,重复场景显著少于新异场景;(5)在平均注视时间上,重复场景显著长于新异场景;(6)在扫描模式比率上,重复场景显著小于新异场景。在本研究条件下,得到以下结论:(1)在多可能目标条件下,被试能内隐地习得显著的背景线索效应。(2)目标的可预测性和背景的熟悉性有助于表现出背景线索效应。(3)无论单目标条件还是多可能目标条件下,背景线索效应的优势均主要来自于搜索中期阶段,支持注意引导。(4)在搜索过程中,背景线索效应具体表现为对重复场景平均注视次数的减少,以及搜索路径更加直接。
[Abstract]:Chun and Jiang (1998) study found that in a visual search task, it was tested that the speed of the search target in a layout-unchanged scenario (repetitive scenario) was faster than searching for targets in a randomly changed scene (new-out scenario). This learning phenomenon is called the background clue effect. Previous research is mostly under single target condition (the target and background of each repeat scene one-to-one), and the mechanism of background thread effect is estimated and analyzed through the result of behavioral experiment. However, in daily life, the relationship between the target and the background is often non-one-to-one. A multi-potential target paradigm is used in this study, which has the advantage that the target location that matches each repeat scene increases, such as matching 2 or 4 target locations per repeat scene without increasing the total target number. The results of the multi-potential target experimental conditions in this study are comparable to the results of the classical single-objective experimental conditions, thus further revealing the cognitive mechanism of background cues. At present, there are three viewpoints: (1) attention to guiding processing mechanism; (2) reaction selection processing mechanism; (3) attention to the double-processing model of guidance and reaction selection. Based on the viewpoint (1), if only the guidance mechanism is paid attention, the eye movement results only find the background clue effect in the middle stage of the search, and based on the point of view (2), if only the selection mechanism is adopted, the eye movement results only find the background clue effect in the late stage of the search; Based on the point of view (3), if attention is paid to both the guidance and the response selection mechanisms, the eye movement results both in the middle of the search and in the late stage of the search. This study includes the following four experiments: experiment 1, in order to verify the background clue effect and the eye movement law of the predecessor under single target condition, and compare it with the multi-potential target position condition. In the experiment, a search scene composed of 12 distracted letters L and 1 target letter T was randomly distributed in 15 college students using the classical background cue effect paradigm. The experimental design is the internal design of 2 (Scene Type: Repeat, New, and New) Rule 7 (Time Window: 1-7). The results show that: (1) In response time, the interaction between scene type and time window is remarkable. With the continuous learning, the response time of repeated scenes is shorter than the new one, that is, it is tried to obtain a significant background clue effect. (2) In the early stage of searching, the repeated scenes are significantly shorter than the new different scenes; (3) in the middle of the search, the repeated scenes are significantly shorter than the new different scenes, and the interaction between the scene type and the time window is remarkable, the obvious background clue effect is displayed, and (4) at the late stage of searching, the repeated scenes are significantly longer than the new different scenes; (5) the repeated scenes are significantly smaller than the new different scenes at the average watching times; (6) the repeated scenes are significantly longer than the new different scenes at the average watching times; and (7) the repeated scenes are significantly smaller than the new different scenes at the scanning mode ratio. Experiment 2, in order to explore whether the background thread effect appears when one repeating scene matches two possible target positions, and the eye movement rule under the search task is tested. Two possible target paradigms were used in experiment 2, and 15 college students were tested. The experimental materials and experimental designs of this experiment were the same as those of experiment 1. The results show that: (1) The interaction between the scene type and the time window is remarkable in response time, that is, it is tried to obtain a significant background clue effect. (2) In the early stage of the search, the main effect and interaction of the scene are not significant; (3) During the middle stage of the search, (4) repeating the scene significantly longer than the new one in the late stage of the search; (5) repeating the scene significantly less than the new different scene at the average watching times; and (6) repeating the scene significantly longer than the new different scene at the average watching time; (7) In the scan mode ratio, the repetition scene is significantly smaller than the new one. Experiment 3, in order to explore whether the background thread effect appears when a duplicate scene matches the four possible target positions, and the eye movement rules under the search task are tested. The experiment consisted of four possible target paradigms, and 14 college students were tested. The experimental materials and experimental designs of this experiment were the same as those of experiment 1. The results show that: (1) In response time, the interaction edge of scene type and time window is remarkable, and the background clue effect is obtained. (2) In the early stage of searching, the repetitive scene is significantly shorter than the new one; (3) In the middle stage of searching, (4) repeating the scene significantly longer than the new one in the late stage of the search; (5) repeating the scene significantly less than the new different scene at the average watching times; and (6) repeating the scene significantly longer than the new different scene at the average watching time; (7) In the scan mode ratio, the repetition scene is significantly smaller than the new one. Experiment 4: In order to explore how to keep the background of repeated scenes familiar, how to display the background clue effect when the target appearing in the scene is random, and the eye movement law of the search. Experiment 4 adopts the random target position condition, and 14 college students are tested. The experimental materials and experimental designs of this experiment were the same as those of experiment 1. The results show that: (1) The interaction between the scene type and the time window is not significant in the reaction time, and the test does not show the background clue effect; (2) In the early stage of searching and searching the late stage, the main effect of the scene type is not significant, the interaction between the scene types and the time windows is not significant; (3) the repeated scenes are significantly shorter than the new different scenes during the middle of the search; (4) the repeated scenes are significantly smaller than the new different scenes at the average watching times; and (5) on the average gaze time, The repetition scene is significantly longer than the new one; (6) In the scan mode ratio, the repetition scene is significantly smaller than the new one. Under the conditions of this study, the following conclusions were obtained: (1) Under multi-potential target conditions, a significant background clue effect could be obtained. (2) The predictability of objectives and familiarity with the background contribute to the performance of contextual cues. (3) Under single-objective condition or multi-potential target condition, the advantages of background thread effect mainly come from the middle stage of search, support attention to guide. (4) During the searching process, the background thread effect is embodied as the decrease of the average number of gaze times of the repeated scenes, and the search path is more direct.
【学位授予单位】：天津师范大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：B842.3
本文编号：2302790
[Abstract]:Chun and Jiang (1998) study found that in a visual search task, it was tested that the speed of the search target in a layout-unchanged scenario (repetitive scenario) was faster than searching for targets in a randomly changed scene (new-out scenario). This learning phenomenon is called the background clue effect. Previous research is mostly under single target condition (the target and background of each repeat scene one-to-one), and the mechanism of background thread effect is estimated and analyzed through the result of behavioral experiment. However, in daily life, the relationship between the target and the background is often non-one-to-one. A multi-potential target paradigm is used in this study, which has the advantage that the target location that matches each repeat scene increases, such as matching 2 or 4 target locations per repeat scene without increasing the total target number. The results of the multi-potential target experimental conditions in this study are comparable to the results of the classical single-objective experimental conditions, thus further revealing the cognitive mechanism of background cues. At present, there are three viewpoints: (1) attention to guiding processing mechanism; (2) reaction selection processing mechanism; (3) attention to the double-processing model of guidance and reaction selection. Based on the viewpoint (1), if only the guidance mechanism is paid attention, the eye movement results only find the background clue effect in the middle stage of the search, and based on the point of view (2), if only the selection mechanism is adopted, the eye movement results only find the background clue effect in the late stage of the search; Based on the point of view (3), if attention is paid to both the guidance and the response selection mechanisms, the eye movement results both in the middle of the search and in the late stage of the search. This study includes the following four experiments: experiment 1, in order to verify the background clue effect and the eye movement law of the predecessor under single target condition, and compare it with the multi-potential target position condition. In the experiment, a search scene composed of 12 distracted letters L and 1 target letter T was randomly distributed in 15 college students using the classical background cue effect paradigm. The experimental design is the internal design of 2 (Scene Type: Repeat, New, and New) Rule 7 (Time Window: 1-7). The results show that: (1) In response time, the interaction between scene type and time window is remarkable. With the continuous learning, the response time of repeated scenes is shorter than the new one, that is, it is tried to obtain a significant background clue effect. (2) In the early stage of searching, the repeated scenes are significantly shorter than the new different scenes; (3) in the middle of the search, the repeated scenes are significantly shorter than the new different scenes, and the interaction between the scene type and the time window is remarkable, the obvious background clue effect is displayed, and (4) at the late stage of searching, the repeated scenes are significantly longer than the new different scenes; (5) the repeated scenes are significantly smaller than the new different scenes at the average watching times; (6) the repeated scenes are significantly longer than the new different scenes at the average watching times; and (7) the repeated scenes are significantly smaller than the new different scenes at the scanning mode ratio. Experiment 2, in order to explore whether the background thread effect appears when one repeating scene matches two possible target positions, and the eye movement rule under the search task is tested. Two possible target paradigms were used in experiment 2, and 15 college students were tested. The experimental materials and experimental designs of this experiment were the same as those of experiment 1. The results show that: (1) The interaction between the scene type and the time window is remarkable in response time, that is, it is tried to obtain a significant background clue effect. (2) In the early stage of the search, the main effect and interaction of the scene are not significant; (3) During the middle stage of the search, (4) repeating the scene significantly longer than the new one in the late stage of the search; (5) repeating the scene significantly less than the new different scene at the average watching times; and (6) repeating the scene significantly longer than the new different scene at the average watching time; (7) In the scan mode ratio, the repetition scene is significantly smaller than the new one. Experiment 3, in order to explore whether the background thread effect appears when a duplicate scene matches the four possible target positions, and the eye movement rules under the search task are tested. The experiment consisted of four possible target paradigms, and 14 college students were tested. The experimental materials and experimental designs of this experiment were the same as those of experiment 1. The results show that: (1) In response time, the interaction edge of scene type and time window is remarkable, and the background clue effect is obtained. (2) In the early stage of searching, the repetitive scene is significantly shorter than the new one; (3) In the middle stage of searching, (4) repeating the scene significantly longer than the new one in the late stage of the search; (5) repeating the scene significantly less than the new different scene at the average watching times; and (6) repeating the scene significantly longer than the new different scene at the average watching time; (7) In the scan mode ratio, the repetition scene is significantly smaller than the new one. Experiment 4: In order to explore how to keep the background of repeated scenes familiar, how to display the background clue effect when the target appearing in the scene is random, and the eye movement law of the search. Experiment 4 adopts the random target position condition, and 14 college students are tested. The experimental materials and experimental designs of this experiment were the same as those of experiment 1. The results show that: (1) The interaction between the scene type and the time window is not significant in the reaction time, and the test does not show the background clue effect; (2) In the early stage of searching and searching the late stage, the main effect of the scene type is not significant, the interaction between the scene types and the time windows is not significant; (3) the repeated scenes are significantly shorter than the new different scenes during the middle of the search; (4) the repeated scenes are significantly smaller than the new different scenes at the average watching times; and (5) on the average gaze time, The repetition scene is significantly longer than the new one; (6) In the scan mode ratio, the repetition scene is significantly smaller than the new one. Under the conditions of this study, the following conclusions were obtained: (1) Under multi-potential target conditions, a significant background clue effect could be obtained. (2) The predictability of objectives and familiarity with the background contribute to the performance of contextual cues. (3) Under single-objective condition or multi-potential target condition, the advantages of background thread effect mainly come from the middle stage of search, support attention to guide. (4) During the searching process, the background thread effect is embodied as the decrease of the average number of gaze times of the repeated scenes, and the search path is more direct.
【学位授予单位】：天津师范大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：B842.3
【参考文献】
相关期刊论文 前4条
1 闫国利;熊建萍;臧传丽;余莉莉;崔磊;白学军;;阅读研究中的主要眼动指标评述[J];心理科学进展;2013年04期
2 白学军;魏玲;沈德立;;背景线索对室内场景中目标搜索的注意引导[J];心理学探新;2011年02期
3 葛列众,王松琴,何灿群;背景线索效应的研究综述[J];心理科学;2004年06期
4 储衡清,周晓林;注意捕获与自上而下的加工过程[J];心理科学进展;2004年05期
相关博士学位论文 前2条
1 赵光;情景线索效应加工机制的探讨[D];西南大学;2013年
2 魏玲;真实场景中视觉搜索的背景线索效应[D];天津师范大学;2011年
相关硕士学位论文 前2条
1 王超;目标位置的预测性对于背景线索效应的影响[D];天津师范大学;2015年
2 赵光;情景线索效应双加工模型机制的研究[D];西南大学;2010年
，本文编号：2302790
本文链接：https://www.wllwen.com/shekelunwen/xinlixingwei/2302790.html
