Synchrony code in recurrent networks


Neurons in the brain encode information about sensory signals in sequences of stereotypic pulses, called spike trains. In the first period of this project, we studied this encoding process in populations of uncoupled cells that are driven by a common stimulus and are also subject to some intrinsic noise and heterogeneity (a variability of their cellular properties). We derived mathematical formulas for the statistics of the population, in particular, its activity and its partial synchronous output and showed that noise can be beneficial in different respects for the transmission and processing of time-dependent stimuli. In particular, we demonstrated in a successful collaboration of experiment and theory, that both intrinsic noise and leak currents are necessary for a synchrony code to work. In the continuation of the project, analysis of information encoding in the synchronous activity will be extended to the much broader class of spiking neurons with recurrent feedback and address two basic situations. In problem A, a recurrent network in the asynchronous irregular state is driven by a common broadband input signal; in problem B, a three-layered network of populations is considered, in which the first population of uncoupled neurons receives a common broadband input, the second population is subject to feedforward input from the first and to delayed inhibitory feedback from the third population. Problem A is a generic setup that applies to a single propagation step in many cortical pathways, while problem B is more typical for the sensory periphery and reflects in particular the architecture of the electrosensory system in weakly electric fish (a model organism, for which we have access to spike data). For the two models, I want to study (i) whether the recurrent feedback can potentially enhance the synchrony coding, i.e. sharpen the information high-pass filter seen in the synchronous output of a population of uncoupled cells; (ii) what is the effect of network noise and neural heterogeneity on information transmission in the case of a strong recurrent feedback. Besides these general theoretical aspects, the project is devoted to a specific phenomenon: the transmission and processing of faint signals in the weakly electric fish.


Principal Investigators
Lindner, Benjamin Prof. Dr. (Details) (Theoretical Physics / Theory of Complex Systems and Neurophysics)

Duration of Project
Start date: 10/2018
End date: 01/2022

Last updated on 2022-20-01 at 21:15