联合信息在价值驱动注意捕获中的作用

发布时间：2018-08-18 21:30

【摘要】：以往奖励学习中和奖励建立联结的刺激特征,当该刺激特征不再能够预测奖励结果时仍能无意识的捕获注意,即价值驱动的注意捕获。个体优先注意与价值相关的刺激具有适应意义,能够避免损失,使奖励最大化。有关价值驱动注意捕获产生的原因,存在着基于空间的注意转移观和基于非空间的过滤观之争。在以往研究中,与奖励联结的特征都是与任务相关的单一视觉特征(如,颜色、形状、方向)。然而对某一特征的注意选择受到空间位置的调节。当与奖励联结时,任务相关特征如何受到空间信息调节还不清楚。为了探讨此问题,本研究采用联合信息(即将任务相关的颜色特征与空间信息结合)来预测奖励,考察价值驱动的注意捕获。本论文包括三项研究,共7个实验,均采用训练-测试两阶段范式。研究一探讨空间位置信息以及颜色和空间位置联合信息在价值驱动注意捕获中的作用,包括3个实验。实验1探讨基于位置的价值驱动的注意捕获效应。20名被试参加实验。训练阶段将两个不相邻的位置出现的颜色刺激和高奖励建立联结,将另外两个不相邻的位置出现的同样刺激和低奖励建立联结,为2(奖励水平:高、低)×2(阶段:1、2)被试内设计。在测试阶段,这个颜色刺激作为分心物出现。测试阶段为单因素被试内设计(分心物的位置:不出现、高奖励位置、低奖励位置、高奖励之间的中性位置、低奖励之间的中性位置、高低奖励之间的中性位置);结果发现,只有当分心物出现在高奖励位置或两个高奖励位置之间的中性位置,才能捕获注意。实验2考察基于位置的价值驱动注意捕获效应是由于整体效应还是由于扩散效应导致的。20名被试参加实验。训练阶段将两个相邻位置分别和高、低奖励建立联结(即:去除掉两个高奖励位置之间的中性位置),为2(奖励水平:高、低)×2(阶段:1、2)被试内设计。测试阶段为单因素被试内设计(分心物的位置:不出现、高奖励位置、低奖励位置、高奖励邻近中性位置、低奖励邻近中性位置、高低奖励之间的中性位置)。结果发现,只有分心物出现在高奖励位置才能够捕获注意,结果支持整体效应。实验3考察颜色和位置信息在价值驱动注意捕获中的交互作用。24名被试参加实验。训练阶段将颜色和位置联合起来预测奖励,为2(奖励水平:高、低)×2(阶段:1、2)被试内设计。测试阶段,分心物的呈现方式为2(颜色:高奖励颜色、低奖励颜色)×5(位置:高奖励位置、低奖励位置、高奖励之间中性位置、低奖励之间中性位置、高低奖励之间中性位置)被试内设计。结果发现,只有高奖励颜色出现在高奖励位置或高奖励之间的中性位置时才能捕获注意。研究二探讨空间场景信息以及颜色和空间场景联合信息在价值驱动注意捕获中的作用。包括2个实验。实验4探讨基于空间场景的价值驱动注意捕获效应。24名被试参加实验。训练阶段将不同的场景下出现的同一颜色刺激分别与高、低奖励建立联结,为2(奖励水平:高、低)×6(阶段:1-6)被试内设计。测试阶段为3(场景:高奖励场景、低奖励场景、中性场景)×2(分心物:出现、不出现)被试内设计。结果表明,只有当分心物出现在高奖励场景下才能捕获注意。实验5考察颜色和空间场景联合信息在价值驱动注意捕获中的交互作用。24名被试参加实验。在训练阶段,将一种场景下出现的颜色A与高奖励建立联结,而将另一种场景下出现的颜色B与低奖励建立联结,为2(奖励水平:高、低)×6(阶段:1-6)被试内设计。测试阶段为3(颜色:高奖励颜色、低奖励颜色、无分心物)×3(场景:高奖励场景、低奖励场景、中性场景)被试内设计。结果表明,高奖励颜色在所有场景中,均能够捕获注意。在研究一和研究二的实验结束后,要求被试完成奖励-联结问卷,实验结果表明,大多数被试在训练阶段能够外显意识到奖励联结规律。研究三探讨任务无关的分心颜色特征在价值驱动注意捕获中的作用及其机制,包括2个实验。实验6考察奖励-联结的任务无关颜色刺激作为突显分心物出现能否捕获注意。20名被试参加实验。训练阶段采用侧抑制任务,将侧面刺激的颜色与奖励建立联结,为2(奖励水平:高、低)×2(侧面与目标数字的奇偶一致性:一致、不一致)。测试阶段为单因素设计(分心物呈现:不出现、高奖励联结分心物出现、低奖励联结分心物出现)。结果发现,当与高奖励联结颜色作为突显分心物出现,相对对低奖励联结颜色来说,能够更大程度的捕获注意。实验7考察奖励-联结的任务无关颜色刺激作为不突显分心物出现能否捕获注意,并结合ERPs的N2pc指标考察价值驱动的注意捕获效应是由于空间转移还是过滤代价导致的。18名被试参加实验。训练阶段与测试阶段的实验设计与实验6完全相同。区别是,测试阶段奖励联结的分心物作为不突显特征出现。行为数据结果表明,与高奖励联结的颜色分心物当不再突显时仍然能够捕获注意。ERPs数据表明,与高奖励联结的分心物并没有诱发N2pc成分。本研究条件下,可以得出以下结论:(1)奖励预测性信息对于价值驱动注意捕获的产生是充分且必要的。奖励预测性信息可以是任务相关的:如目标的颜色和位置,也可以是任务无关的:如目标刺激所存在的场景或和目标无关的分心刺激特征。只要能够预测并区分不同数量的奖励,均能够产生价值驱动的注意捕获效应。(2)位置所产生的价值驱动的注意捕获具有一定的泛化性,表现为分心物出现在两个高奖励之间的中性位置时也能够产生价值驱动的注意捕获效应,这种泛化是由于训练阶段奖励学习的整体效应所导致。(3)当将颜色和位置联合起来预测奖励时,个体在训练阶段能够将颜色和位置的联合特征"捆绑"起来与奖励建立联结;训练阶段建立的联结不能泛化到部分特征上。(4)当将颜色和场景联合起来预测奖励时,颜色会获得优先的注意,产生价值驱动的注意捕获,而场景不发挥作用。(5)在训练阶段,大多数被试能够外显的意识到信息与奖励之间的联结规律,个体在训练阶段进行了外显的学习。(6)训练阶段和高奖励相联结的分心刺激特征,在测试阶段作为突显的分心物出现时或作为不突显的分心物出现时,均能够引发价值驱动的注意捕获效应。ERPs结果表明,注意捕获的产生不是由于注意的空间转移导致的,而是由过滤代价导致的。
[Abstract]:Priority attention to value-related stimuli is adaptive and can avoid loss and maximize reward. In previous studies, the characteristics of reward association were all single visual features related to task (e.g., color, shape, direction). However, the choice of attention for a certain feature was regulated by spatial position. When reward association, task It is not clear how the relevant features are regulated by spatial information. In order to explore this issue, this study uses joint information (i.e. the combination of task-related color features and spatial information) to predict rewards and examine value-driven attention capture. The purpose of this study was to explore the role of spatial location information and color and spatial position information in value-driven attention capture. The same stimulus and low reward in two adjacent locations were linked and designed for 2 (reward level: high, low) *2 (stage: 1, 2) subjects. The results showed that attention could be captured only when distractors appeared in the neutral position between high and low reward positions or between two high reward positions. In the training stage, two adjacent positions were connected with high and low rewards (i.e. removing the neutral position between the two high reward positions) and 2 (reward level: high, low) *2 (stage: 1, 2) subjects were designed. In the test stage, the single factor design (distraction position: no appearance, high) was performed. The results showed that only when distractors appeared in high reward positions could attention be captured. The results supported the overall effect. Experiment 3 examined the interaction between color and location information in value-driven attention capture. Twenty-four subjects participated in the experiment. In the training stage, the color and position were combined to predict the reward, which was 2 (reward level: high, low)*2 (stage: 1, 2). In the test stage, the presentation of distraction was 2 (color: high reward color, low reward color)*5 (position: high reward position, low reward position, neutral position between high reward, low reward position). The results showed that attention could be captured only when the high reward color appeared in the neutral position between the high reward position and the high reward position. Experiment 4 explored the value-driven attention capture effect based on spatial scenarios. Twenty-four subjects participated in the experiment. The results showed that attention could be captured only when distractors appeared in high-reward scenarios. Experiment 5 examined the interaction between color and spatial joint information in value-driven attention capture. 24 subjects participated in the experiment. During the training phase, one scenario was used. Color A was associated with a high reward, while color B in another scenario was associated with a low reward, and was designed for 2 (reward level: high, low) *6 (stage: 1-6). The test phase was 3 (color: high reward color, low reward color, distraction free)*3 (scenario: high reward scenario, low reward scenario, neutral scenario). Design. The results showed that high-reward colors captured attention in all scenarios. After the experiment of study 1 and study 2, subjects were asked to complete the reward-association questionnaire. The experimental results showed that most of the subjects could be aware of the reward-association rule explicitly during the training phase. The role and mechanism of value-driven attention capture included two experiments. Experiment 6 examined whether the reward-linked task-independent color stimuli as prominent distractors could capture attention. Parity Consistency of Side and Target Numbers: Consistency, Inconsistency. The test phase was designed as a single factor (distractors present: absence, high-reward distractors present, low-reward distractors present). The results showed that, when high-reward colors appear as prominent distractors, the range is larger than that of low-reward colors. Experiments 7 examined whether the task-independent color stimuli of reward-link could capture attention as non-prominent distractors, and whether the value-driven attention-capture effect was due to spatial transfer or filtering costs. 18 participants participated in the experiment. The experimental setup of training and testing phases The results of behavioral data showed that the color distractors associated with high rewards could still capture attention when they were no longer highlighted. ERPs data showed that the distractors associated with high rewards did not induce N2pc components. The following conclusions can be drawn: (1) The reward predictive information is sufficient and necessary for the production of value-driven attention capture. The reward predictive information can be task-related, such as the color and location of the target, or task-independent, such as the scene in which the target stimulus exists or the distracted stimulus features unrelated to the target. Value-driven attention capture effect can be produced by measuring and distinguishing different amount of rewards. (2) Value-driven attention capture produced by location has a certain generalization, which shows that distractors can also produce value-driven attention capture effect when they appear in the neutral position between two high rewards. This generalization is due to training. (3) When color and position were combined to predict reward, individuals were able to "bind" the joint features of color and position to establish a link with reward at the training stage; the link established at the training stage could not be generalized to some features. (4) When color and scene were combined to predict reward. In the training stage, most of the subjects were able to perceive the association between information and reward, and the individuals were able to learn the explicit relationship between information and reward. (6) The distracted stimulus characteristics associated with high reward in the training stage were tested. The results of ERPs show that attention capture is not caused by spatial shift of attention, but by filtering cost.
【学位授予单位】：天津师范大学
【学位级别】：博士
【学位授予年份】：2017
【分类号】：B842.3

【相似文献】