Abstract

Analysis methods in cognitive neuroscience have not always matched the richness of fMRI data. Early methods focused on estimating neural activity within individual voxels or regions, averaged over trials or blocks and modeled separately in each participant. This approach mostly neglected the distributed nature of neural representations over voxels, the continuous dynamics of neural activity during tasks, the statistical benefits of performing joint inference over multiple participants and the value of using predictive models to constrain analysis. Several recent exploratory and theory-driven methods have begun to pursue these opportunities. These methods highlight the importance of computational techniques in fMRI analysis, especially machine learning, algorithmic optimization and parallel computing. Adoption of these techniques is enabling a new generation of experiments and analyses that could transform our understanding of some of the most complex - and distinctly human - signals in the brain: acts of cognition such as thoughts, intentions and memories.

Original languageEnglish (US)
Pages (from-to)304-313
Number of pages10
JournalNature Neuroscience
Volume20
Issue number3
DOIs
StatePublished - Feb 23 2017

Fingerprint

Magnetic Resonance Imaging
Cognition
Joints
Brain

All Science Journal Classification (ASJC) codes

  • Neuroscience(all)

Cite this

@article{258d5fe8542f4dce94c0f1a6851a5d98,
title = "Computational approaches to fMRI analysis",
abstract = "Analysis methods in cognitive neuroscience have not always matched the richness of fMRI data. Early methods focused on estimating neural activity within individual voxels or regions, averaged over trials or blocks and modeled separately in each participant. This approach mostly neglected the distributed nature of neural representations over voxels, the continuous dynamics of neural activity during tasks, the statistical benefits of performing joint inference over multiple participants and the value of using predictive models to constrain analysis. Several recent exploratory and theory-driven methods have begun to pursue these opportunities. These methods highlight the importance of computational techniques in fMRI analysis, especially machine learning, algorithmic optimization and parallel computing. Adoption of these techniques is enabling a new generation of experiments and analyses that could transform our understanding of some of the most complex - and distinctly human - signals in the brain: acts of cognition such as thoughts, intentions and memories.",
author = "Cohen, {Jonathan D.} and Daw, {Nathaniel Douglass} and {Engelhardt Martin}, Barbara and Uri Hasson and Kai Li and Yael Niv and Norman, {Kenneth Andrew} and Pillow, {Jonathan William} and Ramadge, {Peter Jeffrey} and Nicholas Turk-Browne and Willke, {Theodore L.}",
year = "2017",
month = "2",
day = "23",
doi = "10.1038/nn.4499",
language = "English (US)",
volume = "20",
pages = "304--313",
journal = "Nature Neuroscience",
issn = "1097-6256",
publisher = "Nature Publishing Group",
number = "3",

}

TY - JOUR

T1 - Computational approaches to fMRI analysis

AU - Cohen, Jonathan D.

AU - Daw, Nathaniel Douglass

AU - Engelhardt Martin, Barbara

AU - Hasson, Uri

AU - Li, Kai

AU - Niv, Yael

AU - Norman, Kenneth Andrew

AU - Pillow, Jonathan William

AU - Ramadge, Peter Jeffrey

AU - Turk-Browne, Nicholas

AU - Willke, Theodore L.

PY - 2017/2/23

Y1 - 2017/2/23

N2 - Analysis methods in cognitive neuroscience have not always matched the richness of fMRI data. Early methods focused on estimating neural activity within individual voxels or regions, averaged over trials or blocks and modeled separately in each participant. This approach mostly neglected the distributed nature of neural representations over voxels, the continuous dynamics of neural activity during tasks, the statistical benefits of performing joint inference over multiple participants and the value of using predictive models to constrain analysis. Several recent exploratory and theory-driven methods have begun to pursue these opportunities. These methods highlight the importance of computational techniques in fMRI analysis, especially machine learning, algorithmic optimization and parallel computing. Adoption of these techniques is enabling a new generation of experiments and analyses that could transform our understanding of some of the most complex - and distinctly human - signals in the brain: acts of cognition such as thoughts, intentions and memories.

AB - Analysis methods in cognitive neuroscience have not always matched the richness of fMRI data. Early methods focused on estimating neural activity within individual voxels or regions, averaged over trials or blocks and modeled separately in each participant. This approach mostly neglected the distributed nature of neural representations over voxels, the continuous dynamics of neural activity during tasks, the statistical benefits of performing joint inference over multiple participants and the value of using predictive models to constrain analysis. Several recent exploratory and theory-driven methods have begun to pursue these opportunities. These methods highlight the importance of computational techniques in fMRI analysis, especially machine learning, algorithmic optimization and parallel computing. Adoption of these techniques is enabling a new generation of experiments and analyses that could transform our understanding of some of the most complex - and distinctly human - signals in the brain: acts of cognition such as thoughts, intentions and memories.

UR - http://www.scopus.com/inward/record.url?scp=85013772631&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85013772631&partnerID=8YFLogxK

U2 - 10.1038/nn.4499

DO - 10.1038/nn.4499

M3 - Review article

VL - 20

SP - 304

EP - 313

JO - Nature Neuroscience

JF - Nature Neuroscience

SN - 1097-6256

IS - 3

ER -