Open Conference Systems, StatPhys 27 Main Conference

Font Size: 
Macroscopic description of neural ensembles
Gonzalo Uribarri, Santiago Boari, Ana Amador, Gabriel B. Mindlin

##manager.scheduler.building##: Edificio Santa Maria Auditorio San Agustin
Date: 2019-07-10 12:00 PM – 03:45 PM
Last modified: 2019-06-14


Neurons can be thought of as highly non-linear dynamic units whose main characteristic is their excitability. In a nervous system, thousands of neurons interact, generating activation patterns that define different functions in the individual. In many cases, the ultimate output of a nervous system involves the control of a macroscopic biomechanical device, acting as an intermediary between the individual and the environment. The study of neural populations can then be framed within the open question of how to build a statistical dynamics for out-of-equilibrium ensemble of units.

In recent years, the study of this kind of systems has largely benefited from new analytical techniques. These techniques allow to express the dynamics of certain macroscopic observables in large ensembles of coupled oscillators as a reduced set of ordinary differential equations. This makes it possible to build low-dimensional models for neural systems from "first principles". One of the biggest challenges in this kind of modeling is to be able to link the analytical observable for which the dynamic equations are obtained, usually the order parameter, with some measurable biological quantity of the system.

In this presentation I will comment on some of the work done in our laboratory in this direction. We were able to write macroscopic models in terms of firing rates and the total synaptic currents involved. We further reported experimental data suggesting that, in certain conditions, the total synaptic current, as defined in our models, provides a good estimate of the Local Field Potential (LFP). We believe that the introduction of measurable macroscopic variables, such as the LFP, suggest a path to build a bridge between experimental data and low-dimensional models for neural populations.