Research Interests

Time correlations for control

We propose to drive any complex system in terms of it's time autocorrelation. In particular, we find that the first autocorrelation coefficient, AC(1) may be a suitable observable for running a closed loop control scheme. We tested the method on three numerical models of very different nature: Ising paramagnetic-ferromagnetic model, a neuronal network model, and Vicsek's bird flock model.

Spatial Correlation Scaling

We study methods for characterizing the dynamical state of a system based on (the scaling of) their spatial correlation, suitably defined. We have characterized one of such methods, known as Box-Scaling, on numerical simulations. We then employed it on different datasets coming from the brains of behaving mice, registered through optogenetic methods.

We plan to extend the method to different animals and registering technologies.

Image from PhysRevE 108, 034302

Image from Phys. Rev. E 104, 064309

Neuronal Network simulations

We study numerical simulations of neuronal network models, where as simple as possible descriptions of each element are used. As a consequence, analytical expression for their dynamics can usually be obtained.

This simple models allow us to understand whether and how the collective effects emerge from simple rules. They also allow us to test and characterize the different tools that will be used with real data, and the expected system response on a variety of situations.

Recently, we have started running simulations on connectomes from a variety of animals and conditions.

Suggested reading: Phys. Rev. E 104, 064309, PhysRevE 106, 054313,

Learning

We study

Image from PhysRevE 108, 034302

Page updated

Google Sites

Report abuse