Unstable Attractors: Existence and Robustness in Networks of Oscillators With Delayed Pulse Coupling

Abstract

We consider unstable attractors; Milnor attractors $A$ such that, for some neighbourhood $U$ of $A$, almost all initial conditions leave $U$. Previous research strongly suggests that unstable attractors exist and even occur robustly (i.e. for open sets of parameter values) in a system modelling biological phenomena, namely in globally coupled oscillators with delayed pulse interactions. In the first part of this paper we give a rigorous definition of unstable attractors for general dynamical systems. We classify unstable attractors into two types, depending on whether or not there is a neighbourhood of the attractor that intersects the basin in a set of positive measure. We give examples of both types of unstable attractor; these examples have non-invertible dynamics that collapse certain open sets onto stable manifolds of saddle orbits. In the second part we give the first rigorous demonstration of existence and robust occurrence of unstable attractors in a network of oscillators with delayed pulse coupling. Although such systems are technically hybrid systems of delay differential equations with discontinuous `firing' events, we show that their dynamics reduces to a finite dimensional hybrid system system after a finite time and hence we can discuss Milnor attractors for this reduced finite dimensional system. We prove that for an open set of phase resetting functions there are saddle periodic orbits that are unstable attractors.

arXiv comment29 pages, 8 figures,submitted to Nonlinearity

References in corpus (3)

• Prevalence of unstable attractors in networks of pulse-coupled oscillators
• Long Chaotic Transients in Complex Networks
• Unstable attractors induce perpetual synchronization and desynchronization

Cited by in corpus (20)

• Mathematical frameworks for oscillatory network dynamics in neuroscience
• Coupling functions: Universal insights into dynamical interaction mechanisms
• Revealing networks from dynamics: an introduction
• Dynamical Entropy Production in Spiking Neuron Networks in the Balanced State
• Stable Irregular Dynamics in Complex Neural Networks
• On designing heteroclinic networks from graphs
• The simplest problem in the collective dynamics of neural networks: Is synchrony stable?
• Potentials and Limits to Basin Stability Estimation
• Speed of synchronization in complex networks of neural oscillators Analytic results based on Random Matrix Theory
• On Dynamics of Integrate-and-Fire Neural Networks with Conductance Based Synapses
• Designing Complex Networks
• On local attraction properties and a stability index for heteroclinic connections
• A view of Neural Networks as dynamical systems
• How Chaotic is the Balanced State?
• Controlled Perturbation-Induced Switching in Pulse-Coupled Oscillator Networks
• Bifurcation From Networks of Unstable Attractors to Heteroclinic Switching
• Synchrony for weak coupling in the complexified Kuramoto model
• Isochronous Dynamics in Pulse Coupled Oscillator Networks with Delay
• Partially unstable attractors in networks of forced integrate-and-fire oscillators
• Stochastic facilitation in heteroclinic communication channels