Signal processing in local neuronal circuits based on activity-dependent noise and competition

TL;DR

This paper investigates how recurrent neuronal networks with plastic synapses detect weak signals, showing that synaptic noise and depression influence frequency selectivity and information processing.

Contribution

It demonstrates that activity-dependent noise and synaptic depression enable frequency selectivity and optimize information transfer in neural circuits with plastic connectivity.

Findings

Networks acquire frequency selectivity due to synaptic noise.

Synaptic depression optimizes mutual information for non-periodic stimuli.

Correlations in signals decrease the mutual information between input and output.

Abstract

We study the characteristics of weak signal detection by a recurrent neuronal network with plastic synaptic coupling. It is shown that in the presence of an asynchronous component in synaptic transmission, the network acquires selectivity with respect to the frequency of weak periodic stimuli. For non-periodic frequency-modulated stimuli, the response is quantified by the mutual information between input (signal) and output (network's activity), and is optimized by synaptic depression. Introducing correlations in signal structure resulted in the decrease of input-output mutual information. Our results suggest that in neural systems with plastic connectivity, information is not merely carried passively by the signal; rather, the information content of the signal itself might determine the mode of its processing by a local neuronal circuit.