An Oscillator Neural Network Retrieving Sparsely Coded Phase Patterns

TL;DR

This paper introduces an extended oscillator neural network model that encodes information in firing timings and non-firing states, enhancing understanding of sparse coding and memory retrieval in neural systems.

Contribution

It proposes a new model allowing sparse phase pattern storage with non-firing states, analyzing its equilibrium and dynamic properties for associative memory.

Findings

The system exhibits good associative memory capabilities.

Timing and non-firing states improve storage capacity.

Model extends understanding of sparse neural coding.

Abstract

Little is known theoretically about the associative memory capabilities of neural networks in which information is encoded not only in the mean firing rate but also in the timing of firings. Particularly, in the case that the fraction of active neurons involved in memorizing patterns becomes small, it is biologically important to consider the timings of firings and to study how such consideration influences storage capacities and quality of recalled patterns. For this purpose, we propose a simple extended model of oscillator neural networks to allow for expression of non-firing state. %which is able to memorize sparsely coded phase patterns including non-firing states. Analyzing both equilibrium states and dynamical properties in recalling processes, we find that the system possesses good associative memory.