Seminario a cargo de la Dra. Denise Duarte

Facultad de Ciencias Económicas

Tendrá lugar en nuestra Facultad el Seminario sobre: «Modeling neuronal spikes through Variable Length Memory Chains», a cargo de la Profesora Dra. Denise Duarte. Ph. D. in Statistics, IME, Universidade de São Paulo, Associate Professor at Department of Statistics, Universidade Federal de Minas Gerais.

El Seminario se llevará a cabo el  miércoles 25 de septiembre del presente, a las 11 horas, en el aula 2 de posgrado, primer piso de la Facultad de Ciencias Económicas UNT. Es organizado por el Instituto de Investigaciones Estadísticas.

Resumen 

Modeling neuronal spikes through Variable Length Memory Chains

Dra. Denise Duarte

Ph. D. in Statistics, IME, Universidade de São Paulo

Associate Professor at Department of Statistics, Universidade Federal de Minas Gerais

A neuron can react with an electrical firing when it is exposed to a stimulus. This is called a spike. Looking at the sequence of spikes in time generated by a stimulated neuron, this can be seen as a stochastic process with a nontrivial dependency structure. It is not expected, for example,to see two consecutive spikes, nor a very long time between two spikes. Moreover, it is reasonable to think that with each new spike, the process loses memory, meaning that it marks a signal renewal time. One model capable of capturing these dependence characteristics is the so-called Variable Length Memory Chains, VLMC, which has several applications known in the literature. In this talk, we present a methodology to estimate the probability law of a neuronal spike process and its dependence structure, given a stimulus, considering that it was generated by a VLMC. We can use the estimated model to classify neurons according to the similarity of their laws.