The CA3 network as a memory store for spatial representations

Papp, Gergely; Witter, Menno P.; Treves, Alessandro

doi:10.1101/lm.687407

Cited by 58 publications

(55 citation statements)

References 92 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Furthermore, it may well be that the network of collateral connections, if extensive, may be crucial in ensuring the smooth continuity of the putative single attractor state generated by a local network of grid cells, rather than a multiplicity of attractor states as in the hippocampus (Battaglia and Treves, 1998). The departure of attractor states generated by networks of finite size from the ideal notion of a continuous attractor has been noted early (Tsodyks and Sejnowski, 1995), but it has only recently emerged as a key issue in computational neuroscience (Hamaguchi and Hatchett, 2006;Papp et al, 2007;Roudi and Treves, 2008).…”

Section: Discussionmentioning

confidence: 99%

The emergence of grid cells: Intelligent design or just adaptation?

2008

View full text Add to dashboard Cite

ABSTRACT:Individual medial entorhinal cortex (mEC) 'grid' cells provide a representation of space that appears to be essentially invariant across environments, modulo simple transformations, in contrast to multiple, rapidly acquired hippocampal maps; it may therefore be established gradually during rodent development. We explore with a simplified mathematical model the possibility that the self-organization of multiple grid fields into a triangular grid pattern may be a single-cell process, driven by firing rate adaptation and slowly varying spatial inputs. A simple analytical derivation indicates that triangular grids are favored asymptotic states of the self-organizing system, and computer simulations confirm that such states are indeed reached during a model learning process, provided it is sufficiently slow to effectively average out fluctuations. The interactions among local ensembles of grid units serve solely to stabilize a common grid orientation. Spatial information, in the real mEC network, may be provided by any combination of feedforward cortical afferents and feedback hippocampal projections from place cells, since either input alone is likely sufficient to yield grid fields. V V C 2008 Wiley-Liss, Inc. KEY WORDS:hippocampus; entorhinal cortex; firing rate adaptation; attractor network; memory DO GRIDS STEM FROM ATTRACTORS OR FROM OSCILLATIONS?Among the complex memory processes operating within the medial temporal lobe (see e.g., Eichenbaum and Lipton, 2008), the core contribution of the hippocampus may be its capacity to retrieve multiple arbitrary representations, a capacity that has been associated to the 'collateral effect' (Marr, 1971). McNaughton and Morris (1987) and Rolls (1989) proposed that the collateral effect is implemented by the recurrent connections of the CA3 region, which led to the insight that the CA3 network may 'compute' just by following attractor dynamics (Amit, 1989), as described by the simplified Hopfield (1982) model. The striking demonstration of abrupt global remapping, indicative of attractor dynamics, in rat place cells (Wills et al., 2005) has reinforced the notion that attractors are a key to understanding hippocampal memory computation. The concurrent discovery of grid cells in neighboring medial entorhinal cortex (mEC;Fyhn et al., 2004;Hafting et al., 2005) has led to the attractor idea reverberating into mEC networks: grid cells have been interpreted as the stable attractor states of spinglass-like interactions among pyramidal cells, mediated by recurrent connections (Fuhs and Touretzky, 2006). Unlike the multiple representations in CA3, however, which require global remapping transitions from one to the other (Leutgeb et al., 2005), local ensembles of mEC grid cells seem to demonstrate a single representation, which shifts and rotates coherently in different environments (Fyhn et al., 2007). If so, and if attractor computation were its core design principle, what the recurrent network in mEC would produce is merely the recovery of this single representation, e....

show abstract

Section: Discussionmentioning

confidence: 99%

The emergence of grid cells: Intelligent design or just adaptation?

2008

View full text Add to dashboard Cite

show abstract

“…In previous reports (see e.g. Papp et al 2007) we have discussed the main properties of simplified models of the CA3 recurrent network, which are extremely interesting even at a rather abstract level of analysis: fragmentation of would-be continuous attractors, tendency to coalesce, storage capacity.…”

Section: Discussionmentioning

confidence: 99%

The role of competitive learning in the generation of DG fields from EC inputs

Treves

2009

Cogn Neurodyn

View full text Add to dashboard Cite

We follow up on a suggestion by Rolls and coworkers, that the effects of competitive learning should be assessed on the shape and number of spatial fields that dentate gyrus (DG) granule cells may form when receiving input from medial entorhinal cortex (mEC) grid units. We consider a simple non-dynamical model where DG units are described by a threshold-linear transfer function, and receive feedforward inputs from 1,000 mEC model grid units of various spacing, orientation and spatial phase. Feedforward weights are updated according to a Hebbian rule as the virtual rodent follows a long simulated trajectory through a single environment. Dentate activity is constrained to be very sparse. We find that indeed competitive Hebbian learning tends to result in a few active DG units with a single place field each, rounded in shape and made larger by iterative weight changes. These effects are more pronounced when produced with thousands of DG units and inputs per DG unit, which the realistic system has available, than with fewer units and inputs, in which case several DG units persists with multiple fields. The emergence of single-field units with learning is in contrast, however, to recent data indicating that most active DG units do have multiple fields. We show how multiple irregularly arranged fields can be produced by the addition of non-space selective lateral entorhinal cortex (lEC) units, which are modelled as simply providing an additional effective input specific to each DG unit. The mean number of such multiple DG fields is enhanced, in particular, when lEC and mEC inputs have overall similar variance across DG units. Finally, we show that in a restricted environment the mean size of the fields is unaltered, while their mean number is scaled down with the area of the environment.

show abstract

“…the development of more elaborate computational models, which however still have to satisfactorily find their way around the spin-glass limit on memory retrieval (see Papp et al 2007). In fact, training virtual rats on several virtual environments, correlated or not, requires them to be endowed with large virtual brains.…”

Section: Correlated Environments Stimulate Orthogonal Ideasmentioning

confidence: 99%

Spatial Cognition, Memory Capacity, and the Evolution of Mammalian Hippocampal Networks

Treves¹

2009

Cognitive Biology

View full text Add to dashboard Cite

The discovery of grid cells in the medial entorhinal cortex (MEC) of the rat (Fyhn et al. 2004) and of the precise triangular pattern of their firing fields (Hafting et al. 2005) requires a substantial reformulation of the questions relating to spatial cognition. It now appears, more clearly than before, that spatial computations per se are largely performed by the rat brain before the hippocampus is ever accessed, and culminate in a sort of universal map of allocentric space, in MEC layer II. Only a portion of the EC participates in such a map, which is applied and used irrespective of context, but in combination with context-specific signals that determine the activity of other parts of entorhinal cortex. The hippocampus operates on the universal map and on context-specific signals to create context-specific metric representations of space, which are stored in memory. The capacity of the hippocampus to rapidly switch between the representations of different contexts is illustrated by hippocampal global "remapping," i.e., the transition to new, unrelated arrangements of place fields by the same population of recorded cells, after suitable behavioral manipulations, and without concurrent MEC remapping (Fyhn et al. 2007). Consequently, understanding the circuitry of the hippocampus crucially involves understanding this capacity for decorrelating spatial representations, at least in rodents. It could well be that in other species complex memories of a less spatial nature take a more prominent role, in which case it would be even more appropriate to approach hippocampal decorrelation and memory processes at an abstract level, and independent of the possibly species-specific spatial processes so finely investigated with the rat model (see also chapter 2 of this volume, by Lucia F. Jacobs).Which approaches can take us beyond a mere functional description of the role of different networks in the brain, and lead us to understand, in evolutionary terms, their design principles? In recent years I have studied three apparently disparate topics from a computational viewpoint: the lamination of the sensory cortex, the differentiation into subfields of the mammalian hippocampus, and the neuronal dynamics that might underlie the faculty for language in the human frontal lobes. These studies share a common perspective: they

show abstract

The CA3 network as a memory store for spatial representations

Cited by 58 publications

References 92 publications

The emergence of grid cells: Intelligent design or just adaptation?

The emergence of grid cells: Intelligent design or just adaptation?

The role of competitive learning in the generation of DG fields from EC inputs

Spatial Cognition, Memory Capacity, and the Evolution of Mammalian Hippocampal Networks

Contact Info

Product

Resources

About