1/ f dynamics adaptable attractor selection and synchronizability in noise-driven multistable neuronal networks

Abstract

We study minimal multistable systems of coupled model neurons with combined excitatory and inhibitory connections. With slow potassium currents, multistability of several firing regimes with distinctively different firing rates is observed. In the presence of noise, there is noise-driven switching between these states of which transient dynamics have 1/ f -type power spectra. The selection between higher- and lower-frequency oscillations depends on external inputs, which results in coherence between the periodic input and the system's firing rate. Without slow potassium currents, there are multistable solutions in which two inhibitory neurons fire synchronously or anti-synchronously. Addition of a small amount of noise results in increased synchronizability of the two neurons depending on the level of external inputs. These results suggest adaptable dynamics of multistable neural attractors to external inputs enhanced by additional noise.