Estudo de redes complexas neuronais com estrutura paramétrica heterogênea
Resumo
Resumo: Este trabalho investigou a dinâmica de um neurônio simulado a partir das equações de Huber-Braun, com foco nas transições dinâmicas para estados caóticos e nas métricas utilizadas para caracterizar essas transições, como o expoente de Lyapunov, a entropia e os microestados de recorrência. A análise do expoente de Lyapunov revelou transições entre comportamentos periódicos e caóticos, com o sistema exibindo comportamento caótico em uma faixa específica de gsa entre 0,35 e 0,40, e retornando à periodicidade fora dessa faixa. A entropia forneceu uma perspectiva adicional, permitindo identificar regiões de alta variabilidade e regularidade nas dinâmicas do sistema. A análise dos microestados de recorrência, utilizando a medição de seis grupos no espaço de recorrência 2 × 2, mostrou como os padrões de recorrência variam entre estados caóticos e periódicos. Ao comparar as três métricas, conclui-se que os microestados de recorrência podem ser utilizados de forma eficaz para caracterizar a periodicidade e as transições para o caos de sistemas dinâmicos, sem a necessidade de conhecer explicitamente as equações geradoras do sistema, permitindo a análise de sistemas com dados empíricos Abstract: This work investigated the dynamics of a simulated neuron based on the Huber-Braun equations, focusing on the dynamic transitions to chaotic states and the metrics used to characterize these transitions, such as the Lyapunov exponent, entropy, and recurrence microstates. The analysis of the Lyapunov exponent revealed transitions between periodic and chaotic behaviors, with the system exhibiting chaotic behavior in a specific range of gsa between 0.35 and 0.40, and returning to periodicity outside of this range. Entropy provided an additional perspective, allowing the identification of regions with high variability and regularity in the system’s dynamics. The analysis of recurrence microstates, using the measurement of six groups in the 2 × 2 recurrence space, showed how recurrence patterns vary between chaotic and periodic states. By comparing the three metrics, it was concluded that recurrence microstates can be effectively used to characterize the periodicity and transitions to chaos in dynamical systems, without the need to explicitly know the generating equations of the system, thus enabling the analysis of systems using empirical data
Collections
- Física [53]