Full text
| |
| Author(s): |
Paulo Henrique Lima de Paula
Total Authors: 1
|
| Document type: | Master's Dissertation |
| Press: | Ribeirão Preto. |
| Institution: | Universidade de São Paulo (USP). Faculdade de Filosofia, Ciências e Letras de Ribeirão Preto (PCARP/BC) |
| Defense date: | 2022-04-13 |
| Examining board members: |
Zhao Liang;
João Roberto Bertini Júnior;
Ricardo Marcondes Marcacini
|
| Advisor: | Zhao Liang; Renata Ferranti Leoni |
| Abstract | |
The human brain can be considered a complex network, as it is divided into structural and functional regions that are interconnected. Functional connectivity has been used to model brain regions that may be physically (or anatomically) separated, through temporal dependence on neural activation patterns. An important task in unsupervised learning is the detection of communities in networks (SILVA; ZHAO, 2016). Although many community detection techniques have been proposed, there are still some challenging issues such as unbalanced community detection and low efficiency. Finding community structures in networks is a competence that has been attracting interest from several areas of study. Neuroscience includes an interest in this study, in which communities in brain networks have considerable factors in brain development and functionality, making it favorable to carry out analyzes of community structures at various scales. The theory of complex networks has been a useful mechanism for the study of the brain, due to its possibility to deal with systems in which they have properties of high complexity. In this context, the theory of complex networks together with the techniques of Functional Magnetic Resonance Imaging (fMRI) are used to create and map the brain network. The connectivity patterns of the brain network were analyzed through community detection. Specifically, we proposed a community detection technique that is inspired by the sequential propagation of signals from the particle competition model and by the parallel propagation inspired by the Self-Organizing Maps (SOM). The model has two salient features: 1) It can detect unbalanced communities; 2) It is much more efficient than the original particle competition model due to the introduction of parallel propagation. Our results demonstrate a high detection precision of modules in networks, in addition it is useful to find unbalanced communities. It has also been shown to be an effective method for characterizing communities in brain networks. (AU) | |
| FAPESP's process: | 19/09319-6 - Identification of activity patterns of brain networks in stroke by community detection |
| Grantee: | Paulo Henrique Lima de Paula |
| Support Opportunities: | Scholarships in Brazil - Master |