High dimensional exploration: A comparison of PCA, distance concentration, and classification performance in two fMRI datasets

Joset A. Etzel; Todd S. Braver

doi:10.1109/CIDM.2014.7008662

High dimensional exploration: A comparison of PCA, distance concentration, and classification performance in two fMRI datasets

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

Abstract

fMRI (functional magnetic resonance imaging) studies frequently create high dimensional datasets, with far more features (voxels) than examples. It is known that such datasets frequently have properties that make analysis challenging, such as concentration of distances. Here, we calculated the probability of distance concentration and proportion of variance explained by PCA in two fMRI datasets, comparing these measures with each other, as well as with the number of voxels and classification accuracy. There are clear differences between the datasets, with one showing levels of distance concentration comparable to those reported in microarray data [1, 2]. While it remains to be determined how typical these results are, they suggest that problematic levels of distance concentration in fMRI datasets may not be a rare occurrence.

Original language	English
Title of host publication	IEEE SSCI 2014 - 2014 IEEE Symposium Series on Computational Intelligence - CIDM 2014
Subtitle of host publication	2014 IEEE Symposium on Computational Intelligence and Data Mining, Proceedings
Publisher	Institute of Electrical and Electronics Engineers Inc.
Pages	157-162
Number of pages	6
ISBN (Electronic)	9781479945191
DOIs	https://doi.org/10.1109/CIDM.2014.7008662
State	Published - Jan 13 2015
Event	5th IEEE Symposium on Computational Intelligence and Data Mining, CIDM 2014 - Orlando, United States Duration: Dec 9 2014 → Dec 12 2014

Publication series

Name	IEEE SSCI 2014 - 2014 IEEE Symposium Series on Computational Intelligence - CIDM 2014: 2014 IEEE Symposium on Computational Intelligence and Data Mining, Proceedings

Conference

Conference	5th IEEE Symposium on Computational Intelligence and Data Mining, CIDM 2014
Country/Territory	United States
City	Orlando
Period	12/9/14 → 12/12/14

Keywords

MVPA
PCA
distance concentration
fMRI
support vector machines

Access to Document

10.1109/CIDM.2014.7008662

Cite this

Etzel, J. A., & Braver, T. S. (2015). High dimensional exploration: A comparison of PCA, distance concentration, and classification performance in two fMRI datasets. In IEEE SSCI 2014 - 2014 IEEE Symposium Series on Computational Intelligence - CIDM 2014: 2014 IEEE Symposium on Computational Intelligence and Data Mining, Proceedings (pp. 157-162). Article 7008662 (IEEE SSCI 2014 - 2014 IEEE Symposium Series on Computational Intelligence - CIDM 2014: 2014 IEEE Symposium on Computational Intelligence and Data Mining, Proceedings). Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.1109/CIDM.2014.7008662

Etzel, Joset A. ; Braver, Todd S. / High dimensional exploration : A comparison of PCA, distance concentration, and classification performance in two fMRI datasets. IEEE SSCI 2014 - 2014 IEEE Symposium Series on Computational Intelligence - CIDM 2014: 2014 IEEE Symposium on Computational Intelligence and Data Mining, Proceedings. Institute of Electrical and Electronics Engineers Inc., 2015. pp. 157-162 (IEEE SSCI 2014 - 2014 IEEE Symposium Series on Computational Intelligence - CIDM 2014: 2014 IEEE Symposium on Computational Intelligence and Data Mining, Proceedings).

@inproceedings{32efaaf261d041a48618a8cb25454d09,

title = "High dimensional exploration: A comparison of PCA, distance concentration, and classification performance in two fMRI datasets",

abstract = "fMRI (functional magnetic resonance imaging) studies frequently create high dimensional datasets, with far more features (voxels) than examples. It is known that such datasets frequently have properties that make analysis challenging, such as concentration of distances. Here, we calculated the probability of distance concentration and proportion of variance explained by PCA in two fMRI datasets, comparing these measures with each other, as well as with the number of voxels and classification accuracy. There are clear differences between the datasets, with one showing levels of distance concentration comparable to those reported in microarray data [1, 2]. While it remains to be determined how typical these results are, they suggest that problematic levels of distance concentration in fMRI datasets may not be a rare occurrence.",

keywords = "MVPA, PCA, distance concentration, fMRI, support vector machines",

author = "Etzel, {Joset A.} and Braver, {Todd S.}",

note = "Publisher Copyright: {\textcopyright} 2014 IEEE.; 5th IEEE Symposium on Computational Intelligence and Data Mining, CIDM 2014 ; Conference date: 09-12-2014 Through 12-12-2014",

year = "2015",

month = jan,

day = "13",

doi = "10.1109/CIDM.2014.7008662",

language = "English",

series = "IEEE SSCI 2014 - 2014 IEEE Symposium Series on Computational Intelligence - CIDM 2014: 2014 IEEE Symposium on Computational Intelligence and Data Mining, Proceedings",

publisher = "Institute of Electrical and Electronics Engineers Inc.",

pages = "157--162",

booktitle = "IEEE SSCI 2014 - 2014 IEEE Symposium Series on Computational Intelligence - CIDM 2014",

}

Etzel, JA & Braver, TS 2015, High dimensional exploration: A comparison of PCA, distance concentration, and classification performance in two fMRI datasets. in IEEE SSCI 2014 - 2014 IEEE Symposium Series on Computational Intelligence - CIDM 2014: 2014 IEEE Symposium on Computational Intelligence and Data Mining, Proceedings., 7008662, IEEE SSCI 2014 - 2014 IEEE Symposium Series on Computational Intelligence - CIDM 2014: 2014 IEEE Symposium on Computational Intelligence and Data Mining, Proceedings, Institute of Electrical and Electronics Engineers Inc., pp. 157-162, 5th IEEE Symposium on Computational Intelligence and Data Mining, CIDM 2014, Orlando, United States, 12/9/14. https://doi.org/10.1109/CIDM.2014.7008662

High dimensional exploration: A comparison of PCA, distance concentration, and classification performance in two fMRI datasets. / Etzel, Joset A.; Braver, Todd S.
IEEE SSCI 2014 - 2014 IEEE Symposium Series on Computational Intelligence - CIDM 2014: 2014 IEEE Symposium on Computational Intelligence and Data Mining, Proceedings. Institute of Electrical and Electronics Engineers Inc., 2015. p. 157-162 7008662 (IEEE SSCI 2014 - 2014 IEEE Symposium Series on Computational Intelligence - CIDM 2014: 2014 IEEE Symposium on Computational Intelligence and Data Mining, Proceedings).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

TY - GEN

T1 - High dimensional exploration

T2 - 5th IEEE Symposium on Computational Intelligence and Data Mining, CIDM 2014

AU - Etzel, Joset A.

AU - Braver, Todd S.

PY - 2015/1/13

Y1 - 2015/1/13

N2 - fMRI (functional magnetic resonance imaging) studies frequently create high dimensional datasets, with far more features (voxels) than examples. It is known that such datasets frequently have properties that make analysis challenging, such as concentration of distances. Here, we calculated the probability of distance concentration and proportion of variance explained by PCA in two fMRI datasets, comparing these measures with each other, as well as with the number of voxels and classification accuracy. There are clear differences between the datasets, with one showing levels of distance concentration comparable to those reported in microarray data [1, 2]. While it remains to be determined how typical these results are, they suggest that problematic levels of distance concentration in fMRI datasets may not be a rare occurrence.

AB - fMRI (functional magnetic resonance imaging) studies frequently create high dimensional datasets, with far more features (voxels) than examples. It is known that such datasets frequently have properties that make analysis challenging, such as concentration of distances. Here, we calculated the probability of distance concentration and proportion of variance explained by PCA in two fMRI datasets, comparing these measures with each other, as well as with the number of voxels and classification accuracy. There are clear differences between the datasets, with one showing levels of distance concentration comparable to those reported in microarray data [1, 2]. While it remains to be determined how typical these results are, they suggest that problematic levels of distance concentration in fMRI datasets may not be a rare occurrence.

KW - MVPA

KW - PCA

KW - distance concentration

KW - fMRI

KW - support vector machines

UR - http://www.scopus.com/inward/record.url?scp=84925064663&partnerID=8YFLogxK

U2 - 10.1109/CIDM.2014.7008662

DO - 10.1109/CIDM.2014.7008662

M3 - Conference contribution

AN - SCOPUS:84925064663

T3 - IEEE SSCI 2014 - 2014 IEEE Symposium Series on Computational Intelligence - CIDM 2014: 2014 IEEE Symposium on Computational Intelligence and Data Mining, Proceedings

SP - 157

EP - 162

BT - IEEE SSCI 2014 - 2014 IEEE Symposium Series on Computational Intelligence - CIDM 2014

PB - Institute of Electrical and Electronics Engineers Inc.

Y2 - 9 December 2014 through 12 December 2014

ER -

Etzel JA, Braver TS. High dimensional exploration: A comparison of PCA, distance concentration, and classification performance in two fMRI datasets. In IEEE SSCI 2014 - 2014 IEEE Symposium Series on Computational Intelligence - CIDM 2014: 2014 IEEE Symposium on Computational Intelligence and Data Mining, Proceedings. Institute of Electrical and Electronics Engineers Inc. 2015. p. 157-162. 7008662. (IEEE SSCI 2014 - 2014 IEEE Symposium Series on Computational Intelligence - CIDM 2014: 2014 IEEE Symposium on Computational Intelligence and Data Mining, Proceedings). doi: 10.1109/CIDM.2014.7008662

High dimensional exploration: A comparison of PCA, distance concentration, and classification performance in two fMRI datasets

Abstract

Publication series

Conference

Keywords

Access to Document

Other files and links

Fingerprint

Cite this