The Information Structure Of Neural Representation Of Chinese Concrete Verbs

Human semantic memory is a huge knowledge base,and people can interact efficiently with the surrounding environment according to their own knowledge,such as contrasting and classifying objects.In semantic memory research,there are two different types of information structures used to describe the encoding of concept representation,namely,distributional information and experiential information.Distributional information is derived from computational linguistics and is related to statistical patterns that co-exist between vocabulary concepts in natural language use,which encode information about word meaning.Experiential information comes from embodied cognitive theory,which suggests that the meaning of a concept is a summary of a particular experience.These two types of information structures have their own cognitive models for constructing,but there is still controversy over the brain areas in which they play a role in the neural representation of concept knowledge.In previous studies,most of the stimuli were letter languages,and relatively few studies were conducted on the organization of verb representations.Therefore,this study uses Chinese verbs as stimuli to explore whether the two cognitive models have neural representations in the brain and to what extent they can explain the neural responses of semantic understanding.In this study,a total of 90 Chinese single-word concrete verbs were included,and f MRI was used to obtain the whole-brain activation pattern of each concept as its own unique spatial pattern of neural activity.Then,according to the natural language processing model of the corpus,the distributional model was established,and the behavioral score of semantic features was used to establish the experience model.Representational similarity analysis was used to find the brain structures that embody the two cognitive models.In the analysis results,the brain areas related to the distributional model indicate that concepts are represented or encoded according to the statistical regularity of co-occurrence in the text,while the brain areas related to the experience model represent that concepts are organized by sensory and motor experience in these areas.The results showed that both cognitive models of semantics have their own representation regions in the brain.The natural language processing model based on corpus is significantly correlated not only with the brain regions of semantic processing,but also with the sensory and motor cortex.The semantic space based on high-dimensional semantic features is significantly correlated with the neural patterns of multi-channel semantic processing.The results showed that different types of semantic relations are expressed by different brain regions.This study,for the first time,used Chinese single-word concrete verbs as stimuli,and the results supported the two cognitive theories of semantic processing,providing new evidence for exploring the essence of Chinese concept representation.