Resonant synchronization and information retrieve from memorized Kuramoto network

TL;DR

This paper introduces a resonant synchronization mechanism in neural networks modeled by Kuramoto oscillators, demonstrating how external stimuli can retrieve stored information through distinct synchronized patterns.

Contribution

It proposes a novel collective behavior called resonant synchronization and applies it to model brain memory retrieval using coupled Kuramoto oscillators.

Findings

Different stimuli induce unique synchronized patterns in the neural network.

The model explains how information can be stored and retrieved via synchronization patterns.

Synchronization occurs just below the critical threshold, enabling selective response to stimuli.

Abstract

A new collective behavior of resonant synchronization is discovered and the ability to retrieve information from brain memory is proposed based on this mechanism. We use modified Kuramoto phase oscillator to simulate the dynamics of a single neuron in self-oscillation state, and investigate the collective responses of a neural network, which is composed of $N$ globally coupled Kuramoto oscillators, to the external stimulus signals in a critical state just below the synchronization threshold of Kuramoto model. The input signals at different driving frequencies, which are used to denote different neural stimuli, can drive the coupled oscillators into different synchronized groups locked to the same effective frequencies and recover different synchronized patterns emerged from their collective dynamics closely related to the predetermined frequency distributions of the oscillators (memory). This model is used to explain how brain stores and retrieves information by the synchronized patterns emerging in the neural network stimulated by the external inputs.