Deforming the hippocampal map

David S. Touretzky, Wendy E. Weisman, Mark C. Fuhs, William E. Skaggs, Andre Fenton, Robert U. Muller

Research output: Contribution to journalArticle

Abstract

To investigate conjoint stimulus control over place cells, Fenton et al. (J Gen Physiol 116:191-209, 2000a) recorded while rats foraged in a cylinder with 45° black and white cue cards on the wall. Card centers were 135° apart. In probe trials, the cards were rotated together or apart by 25°. Firing field centers shifted during these trials, stretching and shrinking the cognitive map. Fenton et al. (2000b) described this deformation with an ad hoc vector field equation. We consider what sorts of neural network mechanisms might be capable of accounting for their observations. In an abstract, maximum likelihood formulation, the rat's location is estimated by a conjoint probability density function of landmark positions. In an attractor neural network model, recurrent connections produce a bump of activity over a two-dimensional array of cells; the bump's position is influenced by landmark features such as distances or bearings. If features are chosen with appropriate care, the attractor network and maximum likelihood models yield similar results, in accord with previous demonstrations that recurrent neural networks can efficiently implement maximum likelihood computations.

Original languageEnglish (US)
Pages (from-to)41-55
Number of pages15
JournalHippocampus
Volume15
Issue number1
DOIs
StatePublished - 2005

Fingerprint

Neural Networks (Computer)
Cues
hydroquinone
Place Cells

Keywords

  • Attractor network model
  • Hippocampal map
  • Maximum likelihood estimation

ASJC Scopus subject areas

  • Neuroscience(all)

Cite this

Touretzky, D. S., Weisman, W. E., Fuhs, M. C., Skaggs, W. E., Fenton, A., & Muller, R. U. (2005). Deforming the hippocampal map. Hippocampus, 15(1), 41-55. https://doi.org/10.1002/hipo.20029

Deforming the hippocampal map. / Touretzky, David S.; Weisman, Wendy E.; Fuhs, Mark C.; Skaggs, William E.; Fenton, Andre; Muller, Robert U.

In: Hippocampus, Vol. 15, No. 1, 2005, p. 41-55.

Research output: Contribution to journalArticle

Touretzky, DS, Weisman, WE, Fuhs, MC, Skaggs, WE, Fenton, A & Muller, RU 2005, 'Deforming the hippocampal map', Hippocampus, vol. 15, no. 1, pp. 41-55. https://doi.org/10.1002/hipo.20029
Touretzky DS, Weisman WE, Fuhs MC, Skaggs WE, Fenton A, Muller RU. Deforming the hippocampal map. Hippocampus. 2005;15(1):41-55. https://doi.org/10.1002/hipo.20029
Touretzky, David S. ; Weisman, Wendy E. ; Fuhs, Mark C. ; Skaggs, William E. ; Fenton, Andre ; Muller, Robert U. / Deforming the hippocampal map. In: Hippocampus. 2005 ; Vol. 15, No. 1. pp. 41-55.
@article{a6db285029264cf1bf0eaa56d1e9740a,
title = "Deforming the hippocampal map",
abstract = "To investigate conjoint stimulus control over place cells, Fenton et al. (J Gen Physiol 116:191-209, 2000a) recorded while rats foraged in a cylinder with 45° black and white cue cards on the wall. Card centers were 135° apart. In probe trials, the cards were rotated together or apart by 25°. Firing field centers shifted during these trials, stretching and shrinking the cognitive map. Fenton et al. (2000b) described this deformation with an ad hoc vector field equation. We consider what sorts of neural network mechanisms might be capable of accounting for their observations. In an abstract, maximum likelihood formulation, the rat's location is estimated by a conjoint probability density function of landmark positions. In an attractor neural network model, recurrent connections produce a bump of activity over a two-dimensional array of cells; the bump's position is influenced by landmark features such as distances or bearings. If features are chosen with appropriate care, the attractor network and maximum likelihood models yield similar results, in accord with previous demonstrations that recurrent neural networks can efficiently implement maximum likelihood computations.",
keywords = "Attractor network model, Hippocampal map, Maximum likelihood estimation",
author = "Touretzky, {David S.} and Weisman, {Wendy E.} and Fuhs, {Mark C.} and Skaggs, {William E.} and Andre Fenton and Muller, {Robert U.}",
year = "2005",
doi = "10.1002/hipo.20029",
language = "English (US)",
volume = "15",
pages = "41--55",
journal = "Hippocampus",
issn = "1050-9631",
publisher = "Wiley-Liss Inc.",
number = "1",

}

TY - JOUR

T1 - Deforming the hippocampal map

AU - Touretzky, David S.

AU - Weisman, Wendy E.

AU - Fuhs, Mark C.

AU - Skaggs, William E.

AU - Fenton, Andre

AU - Muller, Robert U.

PY - 2005

Y1 - 2005

N2 - To investigate conjoint stimulus control over place cells, Fenton et al. (J Gen Physiol 116:191-209, 2000a) recorded while rats foraged in a cylinder with 45° black and white cue cards on the wall. Card centers were 135° apart. In probe trials, the cards were rotated together or apart by 25°. Firing field centers shifted during these trials, stretching and shrinking the cognitive map. Fenton et al. (2000b) described this deformation with an ad hoc vector field equation. We consider what sorts of neural network mechanisms might be capable of accounting for their observations. In an abstract, maximum likelihood formulation, the rat's location is estimated by a conjoint probability density function of landmark positions. In an attractor neural network model, recurrent connections produce a bump of activity over a two-dimensional array of cells; the bump's position is influenced by landmark features such as distances or bearings. If features are chosen with appropriate care, the attractor network and maximum likelihood models yield similar results, in accord with previous demonstrations that recurrent neural networks can efficiently implement maximum likelihood computations.

AB - To investigate conjoint stimulus control over place cells, Fenton et al. (J Gen Physiol 116:191-209, 2000a) recorded while rats foraged in a cylinder with 45° black and white cue cards on the wall. Card centers were 135° apart. In probe trials, the cards were rotated together or apart by 25°. Firing field centers shifted during these trials, stretching and shrinking the cognitive map. Fenton et al. (2000b) described this deformation with an ad hoc vector field equation. We consider what sorts of neural network mechanisms might be capable of accounting for their observations. In an abstract, maximum likelihood formulation, the rat's location is estimated by a conjoint probability density function of landmark positions. In an attractor neural network model, recurrent connections produce a bump of activity over a two-dimensional array of cells; the bump's position is influenced by landmark features such as distances or bearings. If features are chosen with appropriate care, the attractor network and maximum likelihood models yield similar results, in accord with previous demonstrations that recurrent neural networks can efficiently implement maximum likelihood computations.

KW - Attractor network model

KW - Hippocampal map

KW - Maximum likelihood estimation

UR - http://www.scopus.com/inward/record.url?scp=14844321263&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=14844321263&partnerID=8YFLogxK

U2 - 10.1002/hipo.20029

DO - 10.1002/hipo.20029

M3 - Article

VL - 15

SP - 41

EP - 55

JO - Hippocampus

JF - Hippocampus

SN - 1050-9631

IS - 1

ER -