Network mechanisms underlying the role of oscillations in cognitive tasks

Abstract

Oscillatory activity robustly correlates with task demands during many cognitive tasks. However, not only are the network mechanisms underlying the generation of these rhythms poorly understood, but it is also still unknown to what extent they may play a functional role, as opposed to being a mere epiphenomenon. Here we study the mechanisms underlying the influence of oscillatory drive on network dynamics related to cognitive processing in simple working memory (WM), and memory recall tasks. Specifically, we investigate how the frequency of oscillatory input interacts with the intrinsic dynamics in networks of recurrently coupled spiking neurons to cause changes of state: the neuronal correlates of the corresponding cognitive process. We find that slow oscillations, in the delta and theta band, are effective in activating network states associated with memory recall. On the other hand, faster oscillations, in the beta range, can serve to clear memory states by resonantly driving transient bouts of spike synchrony which destabilize the activity. We leverage a recently derived set of exact mean-field equations for networks of quadratic integrate-and-fire neurons to systematically study the bifurcation structure in the periodically forced spiking network. Interestingly, we find that the oscillatory signals which are most effective in allowing flexible switching between network states are not smooth, pure sinusoids, but rather burst-like, with a sharp onset. We show that such periodic bursts themselves readily arise spontaneously in networks of excitatory and inhibitory neurons, and that the burst frequency can be tuned via changes in tonic drive. Finally, we show that oscillations in the gamma range can actually stabilize WM states which otherwise would not persist.

Bumps and oscillons in networks of spiking neurons

“…We focused here on sinusoidal forcing of bump solutions, which is a proxy of oscillations ubiquitous in neuronal systems. In previous work 26 , the neural mass version of this model was studied in terms of its response to oscillatory forcing in various frequency bands, and the present paper makes this exploration feasible also in the spatially-extended model. We leave this exploration to a future publication.…”

“…In this setting we expect oscillons to emerge without bifurcation from a localised steady state of the QIF mean field model with A = 0, upon imposing a small-amplitude forcing, A 1. We therefore select the default kernel (5), set η = −10, for which the model with A = 0 supports one stable (wide) and one unstable (narrow) bump (see Figure 3), and continue time-periodic solutions to (25) in A > 0 for ω = 4, close to the network's resonant frequency 26 .…”

“…In a recent study we have investigated the effect of periodic forcing on a population of excitatory spiking neurons without spatial interaction 26 , whose solutions correspond to the spatially-homogeneous states of the present model. In that context it was shown that a sufficiently large forcing amplitude is able to suppress homogeneous oscillations.…”

A Mean-Field Description of Bursting Dynamics in Spiking Neural Networks with Short-Term Adaptation

“…These regimes could be well traversed via transient inputs, as depicted in Figure 4B. Driving the microscopic model given by (2) and (20)(21)(22)(23) with the same transient inputs, we found that the spiking dynamics were still attracted to the low-dimensional manifold described by the macroscopic system. This shows, that even with τ A = 10τ , the condition τ A τ is sufficiently satisfied for the macroscopic description of the population bursting dynamics to be valid.…”

“…Forη, the parameter range in which the limit cycle exists corresponds to most of the cells in the population being in an excitable regime and has been reported for a number of models using QIF neurons (e.g. 15,22,23 ). Within this range, the inter-burst period can be varied from 13τ up to 105τ via changes in α andη (see Figure 3B).…”