Orientation decoding depends on maps, not columns

Jeremy Freeman, Gijs Joost Brouwer, David J. Heeger, Elisha P. Merriam

Research output: Contribution to journalArticle

Abstract

The representation of orientation in primary visual cortex (V1) has been examined at a fine spatial scale corresponding to the columnar architecture. We present functional magnetic resonance imaging (fMRI) measurements providing evidence for a topographic map of orientation preference inhumanV1at amuchcoarser scale, in register with the angular-position component of the retinotopicmap of V1. This coarse-scale orientation map provides a parsimonious explanation for why multivariate pattern analysis methods succeed in decoding stimulus orientation from fMRI measurements, challenging the widely held assumption that decoding results reflect sampling of spatial irregularities in the fine-scale columnar architecture. Decoding stimulus attributes and cognitive states from fMRI measurements has proven useful for a number of applications, but our results demonstrate that the interpretation cannot assume decoding reflects or exploits columnar organization.

Original languageEnglish (US)
Pages (from-to)4792-4804
Number of pages13
JournalJournal of Neuroscience
Volume31
Issue number13
DOIs
StatePublished - Mar 30 2011

Fingerprint

Magnetic Resonance Imaging
Visual Cortex
Multivariate Analysis

ASJC Scopus subject areas

  • Neuroscience(all)

Cite this

Orientation decoding depends on maps, not columns. / Freeman, Jeremy; Brouwer, Gijs Joost; Heeger, David J.; Merriam, Elisha P.

In: Journal of Neuroscience, Vol. 31, No. 13, 30.03.2011, p. 4792-4804.

Research output: Contribution to journalArticle

Freeman, Jeremy ; Brouwer, Gijs Joost ; Heeger, David J. ; Merriam, Elisha P. / Orientation decoding depends on maps, not columns. In: Journal of Neuroscience. 2011 ; Vol. 31, No. 13. pp. 4792-4804.
@article{761c4f95e736419889a00fdc67201872,
title = "Orientation decoding depends on maps, not columns",
abstract = "The representation of orientation in primary visual cortex (V1) has been examined at a fine spatial scale corresponding to the columnar architecture. We present functional magnetic resonance imaging (fMRI) measurements providing evidence for a topographic map of orientation preference inhumanV1at amuchcoarser scale, in register with the angular-position component of the retinotopicmap of V1. This coarse-scale orientation map provides a parsimonious explanation for why multivariate pattern analysis methods succeed in decoding stimulus orientation from fMRI measurements, challenging the widely held assumption that decoding results reflect sampling of spatial irregularities in the fine-scale columnar architecture. Decoding stimulus attributes and cognitive states from fMRI measurements has proven useful for a number of applications, but our results demonstrate that the interpretation cannot assume decoding reflects or exploits columnar organization.",
author = "Jeremy Freeman and Brouwer, {Gijs Joost} and Heeger, {David J.} and Merriam, {Elisha P.}",
year = "2011",
month = "3",
day = "30",
doi = "10.1523/JNEUROSCI.5160-10.2011",
language = "English (US)",
volume = "31",
pages = "4792--4804",
journal = "Journal of Neuroscience",
issn = "0270-6474",
publisher = "Society for Neuroscience",
number = "13",

}

TY - JOUR

T1 - Orientation decoding depends on maps, not columns

AU - Freeman, Jeremy

AU - Brouwer, Gijs Joost

AU - Heeger, David J.

AU - Merriam, Elisha P.

PY - 2011/3/30

Y1 - 2011/3/30

N2 - The representation of orientation in primary visual cortex (V1) has been examined at a fine spatial scale corresponding to the columnar architecture. We present functional magnetic resonance imaging (fMRI) measurements providing evidence for a topographic map of orientation preference inhumanV1at amuchcoarser scale, in register with the angular-position component of the retinotopicmap of V1. This coarse-scale orientation map provides a parsimonious explanation for why multivariate pattern analysis methods succeed in decoding stimulus orientation from fMRI measurements, challenging the widely held assumption that decoding results reflect sampling of spatial irregularities in the fine-scale columnar architecture. Decoding stimulus attributes and cognitive states from fMRI measurements has proven useful for a number of applications, but our results demonstrate that the interpretation cannot assume decoding reflects or exploits columnar organization.

AB - The representation of orientation in primary visual cortex (V1) has been examined at a fine spatial scale corresponding to the columnar architecture. We present functional magnetic resonance imaging (fMRI) measurements providing evidence for a topographic map of orientation preference inhumanV1at amuchcoarser scale, in register with the angular-position component of the retinotopicmap of V1. This coarse-scale orientation map provides a parsimonious explanation for why multivariate pattern analysis methods succeed in decoding stimulus orientation from fMRI measurements, challenging the widely held assumption that decoding results reflect sampling of spatial irregularities in the fine-scale columnar architecture. Decoding stimulus attributes and cognitive states from fMRI measurements has proven useful for a number of applications, but our results demonstrate that the interpretation cannot assume decoding reflects or exploits columnar organization.

UR - http://www.scopus.com/inward/record.url?scp=79955737563&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=79955737563&partnerID=8YFLogxK

U2 - 10.1523/JNEUROSCI.5160-10.2011

DO - 10.1523/JNEUROSCI.5160-10.2011

M3 - Article

C2 - 21451017

AN - SCOPUS:79955737563

VL - 31

SP - 4792

EP - 4804

JO - Journal of Neuroscience

JF - Journal of Neuroscience

SN - 0270-6474

IS - 13

ER -