Abstract:
Neural rhythms and collective oscillations are ubiquitously recorded in neural structures. Oscillations with 10-45Hz frequency (in the so-called « beta/low gamma » range) are thought to arise from reciprocal interactions between excitatory (E) and inhibitory (I) neurones. Most modelling studies assume networks with random unstructured connectivities. However, several experimental results point out the need to model and analyze the spatial organization of oscillatory neuronal activity.
After describing some of these data, I will describe our recent work in this direction. I will show that long-range excitation can either synchronize or desynchronize oscillatory activity of distinct E-I modules, depending on its strength and the specific considered connectivity. I will also show that stochastic action potential emission by individual neurones gives rise to noise at the module level that needs to be taken into account.
As a result, the oscillatory phase dynamics in a 1D chain of E-I modules on large scales is described by generalization of classic stochastic equations for coupled oscillator phases or growing interfaces, with computable coefficients. On scales relevant to the brain, the phase mode is however found to be insufficient to describe the results and modes beyond the phase mode have to be taken into account.
Finally, I will discuss how the results may help to explain recording data in the motor cortex during movement preparation.
Sponsored by the NYU-ECNU Institute of Brain and Cognitive Science at NYU Shanghai