Busca avançada
Ano de início
Entree

Cortical networks

Processo: 12/51301-8
Linha de fomento:Auxílio à Pesquisa - Regular
Vigência: 01 de fevereiro de 2013 - 31 de janeiro de 2015
Área do conhecimento:Ciências Exatas e da Terra - Ciência da Computação - Metodologia e Técnicas da Computação
Convênio/Acordo: BAYLAT/StMBW - Bavarian Academic Center for Latin America and Bavarian State Ministry of Science and the Arts
Pesquisador responsável:Francisco Aparecido Rodrigues
Beneficiário:Francisco Aparecido Rodrigues
Pesq. responsável no exterior: Cristiane Thielemann
Instituição no exterior: Aschaffenburg University of Applied Sciences (UAS), Alemanha
Instituição-sede: Instituto de Ciências Matemáticas e de Computação (ICMC). Universidade de São Paulo (USP). São Carlos , SP, Brasil
Assunto(s):Redes complexas  Rede nervosa  Modelos matemáticos  Algoritmos 

Resumo

The main goal of the respective project is a profound analysis of cortical cell cultures with a focus on signal transmission and communication patterns inside such a neuronal network. Due to the enormous complexity of these networks, in which one single neuron can address a multitude of other cells, while being influenced by more than just a few cells as well, it is extremely complicated to track down the resulting communication paths of the network itself. Hence it is currently very difficult to understand the general behavior of these cell cultures in particular detail, which in consequence greatly reduces the ability to assess the effect of extern stimuli such as medical drugs. In vitro culturing techniques, in which the network complexity can be influenced directly, as well as the usage of Micro-Electrode Arrays (MEA), that allow the measurement of the signals sent by various neurons, facilitate the assessment process. Yet, the interpretation of the resulting data and hence the characterization of the respective neuronal network still faces certain non-trivial difficulties. As the derived electrode signals can generally originate from multiple neurons in the vicinity of a given MEA electrode, spike sorting methods are necessary to discriminate the respective neuron signals from each other. Although various algorithms have already been designed for this purpose, the lack of reliability of these methods is still significant. To improve these pattern recognition techniques it is essential to acquire a certain amount of a-priori knowledge in order to decrease the degrees of freedom that need to be handled. One promising approach to address this task is the development of a mathematical model that can be used to investigate principles in network structure and signaling behavior. Yet, instead of limiting the model to the neuronal network itself, one unique aspect of this project is the fact that it also includes the surrounding measurement setup, in other words a detailed mathematical description of the Micro-Electrode Array system. Such an improved approach could not only simulate specific behavior of the MEA, but also investigate any influence between the measurement system and the neuronal network culture itself. In a first step, the basic structure of a computational model could be determined and the design requirements could be identified. Once this model can be established successfully, its properties could be compared to data measured in real-life in vitro experiments in order to find possible divergences. In a second step the findings of the mathematical model can be used to improve the structure of the current spike sorting algorithm, which could further significantly increase the accuracy of neuronal network analysis. (AU)