Presentation 2010-07-27
On prediction accuracy in fMRI multi-voxel pattern analysis
Atsushi WADA, Yuichi SAKANO, Hiroshi ANDO,
PDF Download Page PDF download Page Link
Abstract(in Japanese) (See Japanese page)
Abstract(in English) Multi-voxel fMRI analyses are recently-developed techniques that extract information from local activity patterns across voxels to provide sub-voxel sensitivity. However, the nature of the prediction accuracy, a performance measure mainly used in multi-voxel approach to evaluate successful pattern classification, has not yet been clearly understood in relation with fMRI experimental design and preprocessing points of view. In this report, we discuss the effect of preprocessing and pattern classification method on the resulting prediction accuracy.
Keyword(in Japanese) (See Japanese page)
Keyword(in English) Multi-voxel pattern analysis / fMRI / prediction accuracy
Paper # NC2010-33
Date of Issue

Conference Information
Committee NC
Conference Date 2010/7/20(1days)
Place (in Japanese) (See Japanese page)
Place (in English)
Topics (in Japanese) (See Japanese page)
Topics (in English)
Chair
Vice Chair
Secretary
Assistant

Paper Information
Registration To Neurocomputing (NC)
Language JPN
Title (in Japanese) (See Japanese page)
Sub Title (in Japanese) (See Japanese page)
Title (in English) On prediction accuracy in fMRI multi-voxel pattern analysis
Sub Title (in English)
Keyword(1) Multi-voxel pattern analysis
Keyword(2) fMRI
Keyword(3) prediction accuracy
1st Author's Name Atsushi WADA
1st Author's Affiliation Universal Media Research Center, National Institute of Information and Communications Technology(NICT)()
2nd Author's Name Yuichi SAKANO
2nd Author's Affiliation Universal Media Research Center, National Institute of Information and Communications Technology(NICT)
3rd Author's Name Hiroshi ANDO
3rd Author's Affiliation Universal Media Research Center, National Institute of Information and Communications Technology(NICT)
Date 2010-07-27
Paper # NC2010-33
Volume (vol) vol.110
Number (no) 149
Page pp.pp.-
#Pages 3
Date of Issue