Cross-validation and hypothesis testing in neuroimaging: An irenic comment on the exchange between Friston and Lindquist et al.

Philip T. Reiss

doi:10.1016/j.neuroimage.2015.04.032

Cross-validation and hypothesis testing in neuroimaging: An irenic comment on the exchange between Friston and Lindquist et al.

Research output: Contribution to journal › Comment/Debate

Abstract

The "ten ironic rules for statistical reviewers" presented by Friston (2012) prompted a rebuttal by Lindquist et al. (2013), which was followed by a rejoinder by Friston (2013). A key issue left unresolved in this discussion is the use of cross-validation to test the significance of predictive analyses. This note discusses the role that crossvalidation- based and related hypothesis tests have come to play in modern data analyses, in neuroimaging and other fields. It is shown that such tests need not be suboptimal and can fill otherwise-unmet inferential needs.

Original language	English
Pages (from-to)	248-254
Number of pages	7
Journal	NeuroImage
Volume	116
DOIs	https://doi.org/10.1016/j.neuroimage.2015.04.032
State	Published - 1 Aug 2015
Externally published	Yes

Bibliographical note

Funding Information:
My sincere thanks are extended to the referees for their invaluable feedback; to Christos Davatzikos for helpful advice; to Martin Lindquist for providing the fMRI data; to Lan Huo and Pei-Shien Wu for their work on the power simulations; and to Brian Caffo, Xavier Castellanos, Ani Eloyan, Pei-Shien Wu and Yuliya Yoncheva for helpful feedback on the manuscript. This work was partially supported by National Institutes of Health grant 1R01MH095836-01A1 , and indirectly through grants R01MH076136-06 and R01DA035484-01 supporting the original pain study.

Publisher Copyright:
© 2015 Elsevier Inc..

ASJC Scopus subject areas

Neurology
Cognitive Neuroscience

Access to Document

10.1016/j.neuroimage.2015.04.032

Cite this

@article{7d4bf65c3cbe4e67a123bcab7581fcd0,

title = "Cross-validation and hypothesis testing in neuroimaging: An irenic comment on the exchange between Friston and Lindquist et al.",

abstract = "The {"}ten ironic rules for statistical reviewers{"} presented by Friston (2012) prompted a rebuttal by Lindquist et al. (2013), which was followed by a rejoinder by Friston (2013). A key issue left unresolved in this discussion is the use of cross-validation to test the significance of predictive analyses. This note discusses the role that crossvalidation- based and related hypothesis tests have come to play in modern data analyses, in neuroimaging and other fields. It is shown that such tests need not be suboptimal and can fill otherwise-unmet inferential needs.",

author = "Reiss, {Philip T.}",

note = "Funding Information: My sincere thanks are extended to the referees for their invaluable feedback; to Christos Davatzikos for helpful advice; to Martin Lindquist for providing the fMRI data; to Lan Huo and Pei-Shien Wu for their work on the power simulations; and to Brian Caffo, Xavier Castellanos, Ani Eloyan, Pei-Shien Wu and Yuliya Yoncheva for helpful feedback on the manuscript. This work was partially supported by National Institutes of Health grant 1R01MH095836-01A1 , and indirectly through grants R01MH076136-06 and R01DA035484-01 supporting the original pain study. Publisher Copyright: {\textcopyright} 2015 Elsevier Inc..",

year = "2015",

month = aug,

day = "1",

doi = "10.1016/j.neuroimage.2015.04.032",

language = "English",

volume = "116",

pages = "248--254",

journal = "NeuroImage",

issn = "1053-8119",

publisher = "Academic Press Inc.",

}

TY - JOUR

T1 - Cross-validation and hypothesis testing in neuroimaging

T2 - An irenic comment on the exchange between Friston and Lindquist et al.

AU - Reiss, Philip T.

N1 - Funding Information: My sincere thanks are extended to the referees for their invaluable feedback; to Christos Davatzikos for helpful advice; to Martin Lindquist for providing the fMRI data; to Lan Huo and Pei-Shien Wu for their work on the power simulations; and to Brian Caffo, Xavier Castellanos, Ani Eloyan, Pei-Shien Wu and Yuliya Yoncheva for helpful feedback on the manuscript. This work was partially supported by National Institutes of Health grant 1R01MH095836-01A1 , and indirectly through grants R01MH076136-06 and R01DA035484-01 supporting the original pain study. Publisher Copyright: © 2015 Elsevier Inc..

PY - 2015/8/1

Y1 - 2015/8/1

N2 - The "ten ironic rules for statistical reviewers" presented by Friston (2012) prompted a rebuttal by Lindquist et al. (2013), which was followed by a rejoinder by Friston (2013). A key issue left unresolved in this discussion is the use of cross-validation to test the significance of predictive analyses. This note discusses the role that crossvalidation- based and related hypothesis tests have come to play in modern data analyses, in neuroimaging and other fields. It is shown that such tests need not be suboptimal and can fill otherwise-unmet inferential needs.

AB - The "ten ironic rules for statistical reviewers" presented by Friston (2012) prompted a rebuttal by Lindquist et al. (2013), which was followed by a rejoinder by Friston (2013). A key issue left unresolved in this discussion is the use of cross-validation to test the significance of predictive analyses. This note discusses the role that crossvalidation- based and related hypothesis tests have come to play in modern data analyses, in neuroimaging and other fields. It is shown that such tests need not be suboptimal and can fill otherwise-unmet inferential needs.

UR - http://www.scopus.com/inward/record.url?scp=84941744520&partnerID=8YFLogxK

U2 - 10.1016/j.neuroimage.2015.04.032

DO - 10.1016/j.neuroimage.2015.04.032

M3 - Comment/Debate

C2 - 25918034

AN - SCOPUS:84941744520

SN - 1053-8119

VL - 116

SP - 248

EP - 254

JO - NeuroImage

JF - NeuroImage

ER -

Cross-validation and hypothesis testing in neuroimaging: An irenic comment on the exchange between Friston and Lindquist et al.

Abstract

Bibliographical note

ASJC Scopus subject areas

Access to Document

Other files and links

Fingerprint

Cite this