A comparison of context-dependent hippocampal place codes in 1-layer and 2-layer recurrent networks

Recently, it has been suggested that attractor networks may provide a mecha nism for context-dependence in hippocampal place codes. We have proposed th at context may be coded by "latent attractors" - mutually competitive and i nternally cooperative cell groups which channel the system's response to af ferent stimuli. We have also argued that it is the disynaptically recurrent dentate gyrus-hilus (DGH) system which embodies these latent attractors in the hippocampus. Others have suggested the CA3 network - with its monosyna ptic recurrence - is the site. While latent attractors land, thus, context- dependent coding) can be implemented by a 1-layer or 2-layer recurrent netw ork, we show that a 2-layer recurrent network can implement a sparse contex t-dependent place-code using latent attractors much more flexibly than a 1- layer network. (C) 2000 Elsevier Science B.V. All rights reserved.