Cognitive and cortixal correlates and computational modeling of spatial belief revision

当面对相互矛盾的证据时，个人经常不得不改变他们的信念。他们必须选择应该保留哪种信念，应该放弃哪种信念。到目前为止，这个问题只在空间领域之外进行了调查。在项目的第一个资助期，我们引入了空间信念修正这一新的研究课题，并能够证明：(1)个体的空间信念是心理模型表征和检验的结果；(2)个体持有的信念只是与该模型一致的；(3)替代模型是基于初始模型的变化；(4)个体对初始构建的模型进行了最小程度的改变；(5)大多数情况下，矛盾事实的待定位对象被重新定位，而参照对象的初始位置通常保持不变。此外，我们开发了一个计算模型，重建了大部分实验结果。目前，我们使用相同的实验范式进行脑成像研究，以探索信念修正的神经基础。在第二个资助期，我们希望通过考虑具身认知、语言的影响、呈现形式的影响和多通道等方面来继续和扩大我们在空间信念修正方面的工作。主要的研究问题将是：(1)信念修正中的心理模型是如何基于知觉、运动或触觉经验的；(2)描述空间配置的句子的不同语法结构如何影响模型和信念的巩固和修正；(3)特定通道和非通道空间表征是如何被修正的；(4)个体如何处理“伪不一致”：一种配置被认为是不一致的，因为很难形成对前提的可能的一致解释。我们将系统地结合认知实验、脑功能成像研究和计算建模。该项目将产生第一个详细的空间信念修正认知理论，解释如何在认知层面上修正具体化的心理模型，如何在算法上实施修正，以及如何在人脑中实现这些过程。

Individuals often have to revise their beliefs when confronted with contradicting evidence. They have to select which belief they should retain and which one they should give up. The problem has so far only been investigated outside the spatial domain. In the first funding period of the project, we introduced the new research topic of spatial belief revision and were able to show that (1) individual's spatial beliefs result from the representation and inspection of a mental model; (2) individuals hold a belief just as long as it coheres with this model; (3) alternative models are based on the Variation of the initial model; (4) individuals apply "minimal changes" to the initially constructed model; (5) Mostly, the to-be-located objects of the contradicting facts are relocated while the reference objects' initial locations are usually maintained. Further, we developed a computational model reconstructing most of the experimental findings. Currently, we run brain imaging studies using the same experimental paradigms to explore the neural basis of belief revision. In the second funding period, we want to continue and extend our work on spatial belief revision by taking into account aspects of embodied cognition, the influences of language, the effects of presentation format, and multimodality. The main research questions will be: (1) how mental models in belief revision are grounded in perceptual, motoric, or haptic experience; (2) how different grammatical structures of sentences describing spatial configurations effect the entrenchment and revision of models and beliefs; (3) how modality-specific and amodal spatial representations are revised; (4) how individuals deal with "pseudo-inconsistency": a configuration is considered to be inconsistent, because possible consistent interpretations of the premises are difficult to form. We will systemically combine cognitive experiments, functional brain imaging studies, and computational modeling. The project shall result in the first elaborated cognitive theory of spatial belief revision that explains how embodied mental models are revised on the cognitive level, how revision is algorithmically implemented, and how the processes are realized in the human brain.