Research On Language Vocabulary Frequency Network And Brain Episodic Memory Network Based On Complex System

Management Science cannot develop without research of human behavior characteristics. Human behavior is based on the human brain neural operating process. Language and memory are two cognitive functions with which people can encode, store, retrieve and reencode information. These two functions are working interdependently and collaboratively, which contain people’s external expression and internal accumulation. With the development and improvement of cognitive neuroscience, neuroimaging and complex system theory, more people recognize the importance of the relation among brain, cognition and management. In this dissertation, two main sections analysed these two functions based on complex system theory:models of vocabulary frequency dynamically generated in language text, were constructed and discussed; using the advanced magnetic resonance imaging technique, a putative human brain episodic memory small-world network was constructed and analysed on the level of human brain neural activity essence. The theoretical results on the mechanism of language vocabulary network can be not only applied to vocabulary text generated network, but also to a greater range of human activities quantitative models. It has economic, social and academic meanings to better understand statistical properties of human behavioral principle in management, and to analyse and predict network evolution. As the aging of society in China becoming more and more serious, Alzheimer’s disease has become one of the four biggest threats to human’s health and gradually spread to younger adult people. Because of the untreatable, latent properties and the symptoms of Alzheimer’s disease produce memory loss and thinking problems, change of personal behaviors, influence normal decision, it brings great emotional, financial and time burdens to society and families. In this dissertation, the research on changes of resting-state episodic memory brain network in early stage of the disease has importance academic, economic and social meanings. The main contents of the dissertation are as follows:(1) Vocabulary distribution networks of language text dynamic generating process were established using complex system theory. Several key complex network distribution models’ equivalence and other characteristics were discussed using mathematical derivation. Some easily confused concepts were given strict mathematical definition. An application was demonstrated which perfectly fit the theoretical results. Based on the preferential attachment mechanism of the original Simon’s model, Heaps’ law, an important law in linguistics, is equivalent to power law, the core distribution law in complex network theory. The sufficient and necessary relations between Waring distribution, a class of power law distribution, and a conditional expectation that describes sublinear growth pattern of nodes’ average degree in network, were derived. The relation between power law and densification power law (describes superlinear growth pattern of network) was discussed. The strict mathematical definition, a property of scale-free function and its relation to scale-free network were given or discussed. Finally, a text of full-length novel was used to fit and test our theoretical models as an example.(2) A new method was used to establish resting-state human brain episodic memory network using the technique of functional magnetic resonance imaging (fMRI). The rationality of the putative network was verified. Abnormal increased functional connectivity were found in this network in both early and dementia stages of Alzheimer’s disease, which is predictive of decline of memory performance. Diffrences of small-world properties and network resilience under random failure and targeted attack were tested between different stages of the disease. The possible explainations of the abnormal increased functional connectivity were given.The main contributions are as follows:(1) The relations between some core distribution laws in complex system were derived sufficiently and necessarily, which can easily expand to more real complex network applications. (2) The strict mathematical definition of some easily confused concepts such as scale-free distribution, scale-free function and scale-free network were given. (3) Using actual experimental research method to analyse changes of personal behavior and cognitive function on the level of human brain neural activities. A new method was used to establish human brain episodic memory network of resting-state fMRI, which against the difficulty that episodic memory is not an independent, distinct function that has ambiguous resting-state brain locations. The changes of the episodic network structure were analysed by using small-world method. (4) Abnormal increased functional connectivity were found in episodic memory network of resting-state fMRI in early stage of Alzheimer’s disease, which can predict decline of memory performance. Resting-state imaging has its advantages of easy operating and wide range of application.