Multiplexing rhythmic information by spike timing dependent plasticity

TL;DR

This study investigates whether spike-timing-dependent plasticity (STDP) can enable a neuron to multiplex rhythmic information across different frequency channels, revealing conditions under which multiple rhythms can be simultaneously represented.

Contribution

The paper demonstrates through theoretical and numerical analysis that STDP can support multiplexing of rhythmic signals without leading to dominance of one rhythm, highlighting a novel functional role for STDP.

Findings

STDP allows neurons to respond to multiple rhythmic inputs simultaneously.

Interactions between different rhythmic channels are limited and mainly mediated by shared activity.

Synaptic weights remain dynamic, enabling continuous adaptation rather than fixed responses.

Abstract

Rhythmic activity has been associated with a wide range of cognitive processes. Previous studies have shown that spike-timing-dependent plasticity can facilitate the transfer of rhythmic activity downstream the information processing pathway. However, STDP has also been known to generate strong winner-take-all like competitions between subgroups of correlated synaptic inputs. Consequently, one might expect that STDP would induce strong competition between different rhythmicity channels thus preventing the multiplexing of information across different frequency channels. This study explored whether STDP facilitates the multiplexing of information across multiple frequency channels, and if so, under what conditions. We investigated the STDP dynamics in the framework of a model consisting of two competing subpopulations of neurons that synapse in a feedforward manner onto a single postsynaptic neuron. Each sub-population was assumed to oscillate in an independent manner and in a different frequency band. To investigate the STDP dynamics, a mean field Fokker-Planck theory was developed in the limit of the slow learning rate. Surprisingly, our theory predicted limited interactions between the different sub-groups. Our analysis further revealed that the interaction between these channels was mainly mediated by the shared component of the mean activity. Next, we generalized these results beyond the simplistic model using numerical simulations. We found that for a wide range of parameters, the system converged to a solution in which the post-synaptic neuron responded to both rhythms. Nevertheless, all the synaptic weights remained dynamic and did not converge to a fixed point. These findings imply that STDP can support the multiplexing of rhythmic information and demonstrate how functionality can be retained in the face of continuous remodeling of all the synaptic weights.