Full text
| |
| Author(s): |
Rodrigo Felipe de Oliveira Pena
Total Authors: 1
|
| Document type: | Doctoral Thesis |
| Press: | Ribeirão Preto. |
| Institution: | Universidade de São Paulo (USP). Faculdade de Filosofia, Ciências e Letras de Ribeirão Preto (PCARP/BC) |
| Defense date: | 2018-12-06 |
| Examining board members: |
Antonio Carlos Roque da Silva Filho;
Ariadne de Andrade Costa;
Ricardo Mauricio Xavier Leão;
Elbert Einstein Nehrer Macau;
Renato Tinós
|
| Advisor: | Antonio Carlos Roque da Silva Filho |
| Abstract | |
In cortical network models with spiking neurons, the mechanisms responsible for the emergence and impact of neuronal activity fluctuations are not yet completely understood. In this work, computational models of cortical networks were used to investigate how rhythmic and non-rhythmic fluctuations arise and their possible consequences. Networks with two types of topology were studied: random and hierarchical modular, this latter inspired on experimental evidence about cortical architecture. Three different simplified spiking neuron models were used: integrate-and-fire, Izhikevich, and integrate-and-fire with adaptation. Initially, the types of self-sustained activity patterns that emerge in hierarchical modular networks with mixtures of electrophysiological neuronal classes were studied. In these models, the long-duration self-sustained activity patterns are oscillatory and their lifetime depend on the hierarchical level of the network and its neuronal composition. Next, the effect of the introduction of synaptic noise in random networks was studied. These networks displayed intermittent alternations between rhythmic and non-rhythmic activity patterns with characteristics similar to synchronous and asynchronous cortical states, respectively. A reductionist approach for homogeneous neuronal networks, in which an iterative self-consistent scheme is used so that a single neuron spike train generates second-order statistical properties similar to the ones of a network, was extended to heterogeneous networks. It was shown that this reductionist scheme captures situations in which slow activity fluctuations emerge. Finally, the reductionist scheme and information theoretical tools were used to study the emergence of slow activity fluctuations and their propagation through hierarchical modular networks. The results show that the information propagation in the network depends on the number of modules, suggesting an optimal hierarchical level for information propagation. The studies done contribute to deepen the understanding of the relationship between structure and neuronal composition in cortical network models, and point to mechanisms of emergence and maintenance of activity fluctuations in these networks (AU) | |
| FAPESP's process: | 13/25667-8 - Mechanisms of propagation of epileptiform activity in a large-scale cortical model |
| Grantee: | Rodrigo Felipe de Oliveira Pena |
| Support Opportunities: | Scholarships in Brazil - Doctorate (Direct) |