维度的结合与分离对归类不确定性预测的影响

探讨在归类不确定的情境下目标与预测特征两个维度的结合或分离对被试特征预测的影响。共包括 3个实验 :实验 1在Murphy和Ross的研究的基础上进一步提高非靶类型中目标及预测特征的基本概率 ,考察被试的特征预测是否会受非靶类型信息的影响。实验 2探讨非靶类型的目标与预测特征结合与否是否会影响被试预测特征时对非靶类型信息的使用。实验 3探讨提高靶类型中目标与预测特征结合的比例是否影响被试对特征的预测。结果表明 :当非靶类型中目标与关键特征处于分离的状态时 ,被试在进行特征预测时没有利用非靶类型的信息 ,符合单类说的假设 ;而当非靶类型中目标与关键特征结合时 ,被试在进行特征预测时则会利用非靶类型的信息 ,符合Bayesian规则 ;靶类型中的目标与关键特征结合的比例提高 ,被试对特征预测的概率也随之提高。据此 ,本研究将目标与预测特征结合比例这个变量加入Bayesian规则的计算公式 ,对该预测模型进行了修正     

Action representations activated by task‐irrelevant information: Is it really irrelevant?

Accessing action knowledge is believed to rely on the activation of action representations through the retrieval of functional, manipulative, and spatial information associated with objects. However, it remains unclear whether action representations can be activated in this way when the object information is irrelevant to the current judgment. The present study investigated this question by independently manipulating the correctness of three types of action‐related information: the functional relation between the two objects, the grip applied to the objects, and the orientation of the objects. In each of three tasks in Experiment 1, participants evaluated the correctness of only one of the three information types (function, grip or orientation). Similar results were achieved with all three tasks: “correct” judgments were facilitated when the other dimensions were correct; however, “incorrect” judgments were facilitated when the other two dimensions were both correct and also when they were both incorrect. In Experiment 2, when participants attended to an action‐irrelevant feature (object color), there was no interaction between function, grip, and orientation. These results clearly indicate that action representations can be activated by retrieval of functional, manipulative, and spatial knowledge about objects, even though this is task‐irrelevant information.     

特征整合与加工深度对道德概念容器隐喻联结的影响

提出了隐喻提取假说将隐喻联结的形成和提取进行分离, 并通过3个实验探究了道德概念与容器空间的隐喻联结及其受知觉加工深度和特征整合程度的影响。实验1采用空间Stroop范式, 实验2a和实验2b均采用启动范式, 实验3a和实验3b均采用加入任务要求的Stroop范式。结果发现：(1)在经典Stroop范式中未发现道德概念与容器空间的隐喻联结; (2)在启动范式中发现, 较深知觉加工深度下道德概念与容器空间存在较弱的隐喻联结; (3)在较高特征整合程度的Stroop任务中, 道德概念与容器空间存在较强的隐喻联结。结果表明：道德概念与容器空间存在道德为内、不道德为外的隐喻联结, 这种隐喻联结在映射上表现为双向性, 并且受到特征整合程度和知觉加工深度的影响, 同时也为隐喻提取假说提供了证据支持。     

The influence of visual noise in the binding of irrelevant features to responses

ABSTRACT

In order to efficiently control our actions, stimuli and responses made to them are stored and integrated into event files. This integration is not restricted to relevant stimuli. Even irrelevant stimuli or features that co-occur at responding can be integrated with the response. In the presence of only one irrelevant feature, binding effects for that feature are generally observed. In the presence of more than one irrelevant features, empirical evidence is inconsistent. The present experiments tested feature-response binding effects for irrelevant features when more than one irrelevant feature was present. In two experiments feature-response binding effects for three irrelevant features (colour, shape, and location) were tested in a block-wise manner. In each block one feature was orthogonally varied to the response while the others were held constant (Experiment 1) or always changed (Experiment 2). Significant binding effects were observed for each of three features when the other two were held constant (Exp. 1). However, when the other two features were varied, significant binding effects were only observed for colour and location, but not for shape (Exp. 2). The results suggest that strength of feature-response binding effects is influenced by the variation of simultaneously presented features.     

视觉特征捆绑：基于时间同步还是基于相同空间位置？

通过操纵客体呈现的空间位置和时间顺序,来考察视觉特征捆绑是基于相同空间位置线索还是基于时间邻近线索。在两个实验中,给被试在固定的注视点位置（固定呈现条件）或在注视点上/下位置（变换呈现条件）快速地系列呈现刺激,记录靶子前后各两个时间位置（-2、-1、+1、+2）的错觉性结合（IC）率。结果发现,在固定呈现条件下,IC更多地出现在-1和+1的时间位置上;而在变换呈现条件下,错觉性结合更多地出现在-2和+2时间位置上。为了进一步排除特定实验范式的影响,在实验3中,我们采用特征分离范式,使被分离的刺激特征同时呈现在相同（相同位置同步）或不同（不同位置同步）的空间位置上。结果表明,这两种条件下的IC率无显著差异,这说明错觉性结合不受时间同步模式的影响。综上所述,来自三个实验的结果表明,时间范围内的特征捆绑机制是按空间位置,而不是按时间邻近,线索进行的：若特征的空间位置相同,时间邻近的特征才被捆绑起来;若空间位置不同,即使时间邻近的特征也不能被捆绑到一起。     

类别标签与典型性程度对类别特征推理的影响

本研究采用学习-测试二阶段实验范式,探讨了类别标签及典型性程度对类别特征推理任务的共同影响,结果表明,类别标签及典型性程度都会影响类别特征推理任务,而且,类别标签及典型性程度之间存在交互作用,典型性程度为高条件下类别标签对类别特征推理任务的影响要高于典型性程度为低条件,类别标签匹配条件下,典型性程度对类别特征推理任务的影响要高于类别标签不匹配条件。     

用文字概率衡量不确定性：特征和问题

文字概率是衡量不确定性的方式之一, 即人们使用诸如“也许”、“未必”的词汇来描述特定事件发生的可能性。文字概率不同于数字概率, 主要体现在文字概率的模糊性、非概率运算性和语义特性上。这使得相对于数字概率, 用文字概率衡量不确定性既有优势也有问题, 进而对人们的不确定信息沟通和风险决策造成影响。虽然文字概率与数字概率存在特征上的差异并且人们在日常交流中偏爱文字概率, 但是大部分风险领域的研究却仅局限于数字概率, 今后有必要研究使用文字概率测量的风险决策。在已有文字概率特征研究的基础上, 还可以进一步探究其不同于数字概率的其它特征(文字/数字概率与双系统模型的联系、文字概率的跨文化差异等)及其对风险决策的影响。     

任务维度和任务无关维度干扰信息的同质性影响视觉搜索的神经机制

在三个实验中, 我们利用功能性磁共振成像(functional magnetic resonance imaging, fMRI)技术考察了在简单特征搜索任务和客体内联合搜索任务中任务维度干扰项的同质性(或称一致性, homogeneity)和任务无关维度干扰信息的同质性对视觉搜索行为的影响及其神经机制。主要探讨了三个问题：(1) 任务维度干扰信息的同质性与目标存在性之间交互作用的神经机制; (2) 任务维度干扰信息的同质性与任务无关维度干扰信息的同质性之间的交互作用及其神经机制; (3) 上述神经机制在不同类型的视觉搜索(特征搜索和联合搜索)中的异同。     

多感觉整合的时间再校准

跨通道刺激的时间同步性是多感觉整合的必要条件, 但是由于刺激物理传导与神经传导的差异, 它们在时间上并非完全匹配。时间再校准指大脑能够适应跨通道刺激间很短的时间延迟的现象, 反映了多感觉整合在时间维度上的可塑性, 表现为适应相继呈现的跨通道刺激后, 主观同时点向时间延迟方向的偏移。本文主要介绍了时间再校准的通道效应与潜在机制, 其初始加工阶段, 它与刺激内容加工的关系及主要的影响因素。今后的研究应当进一步探索时间再校准能否发生于早期加工阶段, 检验其认知过程是否具有双向性, 探讨空间选择性注意的作用, 并结合神经机制的研究, 从综合的视角进行更完善的理论建构。