Reservoirs of Stability: Flux Tubes in the Dynamics of Cortical Circuits

Abstract: Triggering a single additional spike in a cerebral cortical neuron was recently demonstrated to cause a cascade of extra spikes in the network that is likely to rapidly decorrelate the network's microstate. The mechanisms involved in this extreme sensitivity of cortical networks are currently not well understood. Here, we show in a minimal model of cortical circuit dynamics that exponential state separation after single spike and even single synapse perturbations coexists with dynamical stability to infinitesimal state perturbations. We propose a unifying picture of exponentially separating flux tubes enclosing unique stable trajectories composing the networks' state spaces.