Berlin 2012 – scientific programme

Parts | Days | Selection | Search | Updates | Downloads | Help

SYSD: Symposium SKM Dissertationspreis

SYSD 1: SKM Dissertation Prize

SYSD 1.2: Invited Talk

Monday, March 26, 2012, 12:20–12:40, EW 201

Pulsed communication in neocortical neuronal networks — •Moritz Deger — Bernstein Center Freiburg, University of Freiburg, Freiburg im Breisgau, Germany

Neurons communicate via action potentials, also called spikes. The times of neuronal spike emission define a sequence of point-like events, the spike train. Firstly, I will introduce stochastic point processes that allow to define concise models of spike trains of single neurons. A neuron in a neocortical network typically receives spike trains from thousands of cells as input. This input signal can be modelled as a superposition of point processes. As we found, the effective refractory period of the component spike trains dominates the superposition's statistics. This result suggests superpositions of Poisson process with dead-time as a minimal and analytically tractable model of pooled neuronal input. Secondly, I will discuss the stochastic dynamics of the membrane potential of neurons driven by fluctuating inputs. Our extension of the theory of the integrate-and-fire neuron allows for pulsed inputs instead of diffusive noise and reveals the neuron's non-linear, instantaneous response to perturbations. Finally, the statistics of a neuronal spike train can be described as a function of the statistics of the input. Assuming consistency of neuronal and input (network) activity, these methods enable to understand the spike train statistics of neurons in generic neocortical networks.