Dynamics and robustness of familiarity memory

TL;DR

This paper explores the dynamics and robustness of familiarity memory in neural networks, comparing energy-based and energy derivative-based discriminators, and analyzing their performance under noise.

Contribution

It introduces a detailed analysis of two familiarity discriminators' dynamical properties and their robustness to neural noise using signal-to-noise ratio and mean field methods.

Findings

Familiarity signals decay over time after stimulus presentation.

Robustness of discriminators decreases with increasing neural noise.

Maximum discriminable stimuli depends on noise level and discriminator type.

Abstract

When one is presented with an item or a face, one can sometimes have a sense of recognition without being able to recall where or when one has encountered it before. This sense of recognition is known as familiarity. Following previous computational models of familiarity memory we investigate the dynamical properties of familiarity discrimination, and contrast two different familiarity discriminators: one based on the energy of the neural network, and the other based on the time derivative of the energy. We show how the familiarity signal decays after a stimulus is presented, and examine the robustness of the familiarity discriminator in the presence of random fluctuations in neural activity. For both discriminators we establish, via a combined method of signal-to-noise ratio and mean field analysis, how the maximum number of successfully discriminated stimuli depends on the noise level.