Bistability of attractor neural network model for inferior temporal cortex

摘要

We study a neural network model for inferior temporal cortex proposed by Griniasty et al. in the finite memory loading. We show that uncorrelated Hopfleld type and correlated attractors coexist, and examine the retrieval processes for these attractors when the initial state is set to a noisy-degraded stored pattern. We found that there exists a critical value of initial overlap, that is, the system converges to the correlated attractor when the degree of noise is large, otherwise to the Hopfleld type attractor. We resolve controversy of previously obtained experimental findings regarding neuron properties in the inferior temporal cortex, and propose a novel experimental paradigm based on these theoretical results.