Advanced search
1 file | 1.21 MB

Improving selection stability of multiple testing procedures for fMRI

Joke Durnez (UGent) and Beatrijs Moerkerke (UGent)
Author
Organization
Abstract
In search of an appropriate thresholding technique in the analysis of functional MRI-data, several methods to prevent an inflation of false positives have been proposed. Two popular (voxelwise) methods are the Bonferroni procedure (BF), which controls the familywise error rate (FWER), and the Benjamini-Hochberg procedure (BH), which controls the false discovery rate (FDR) (Benjamini & Hochberg 1995). Multiple testing procedures are typically evaluated on their average performance with respect to error rates, ignoring the aspect of variability. Resampling techniques allow to assess the selection variability of individual features (voxels). Following the approach of Gordon, Chen, Glazko & Yakovlev (2009) in the context of gene selection, we investigated whether variability on test results for BF and BH can be reduced by including both the significance and selection variability of the voxels in the decision criterion.
Keywords
Stability, Multiple testing problem, fMRI

Downloads

  • DurnzeMoerkerke HBM2010.pdf
    • full text
    • |
    • open access
    • |
    • PDF
    • |
    • 1.21 MB

Citation

Please use this url to cite or link to this publication:

Chicago
Durnez, Joke, and Beatrijs Moerkerke. 2010. “Improving Selection Stability of Multiple Testing Procedures for fMRI.” In 16th Annual Meeting of the Organization for Human Brain Mapping, Abstracts.
APA
Durnez, J., & Moerkerke, B. (2010). Improving selection stability of multiple testing procedures for fMRI. 16th annual meeting of the organization for human brain mapping, Abstracts. Presented at the 16th Annual Meeting of the Organization for Human Brain Mapping.
Vancouver
1.
Durnez J, Moerkerke B. Improving selection stability of multiple testing procedures for fMRI. 16th annual meeting of the organization for human brain mapping, Abstracts. 2010.
MLA
Durnez, Joke, and Beatrijs Moerkerke. “Improving Selection Stability of Multiple Testing Procedures for fMRI.” 16th Annual Meeting of the Organization for Human Brain Mapping, Abstracts. 2010. Print.
@inproceedings{1220843,
  abstract     = {In search of an appropriate thresholding technique in the analysis of functional MRI-data, several methods to prevent an inflation of false positives have been proposed. Two popular (voxelwise) methods are the Bonferroni procedure (BF), which controls the familywise error rate (FWER), and the Benjamini-Hochberg procedure (BH), which controls the false discovery rate (FDR) (Benjamini \& Hochberg 1995). Multiple testing procedures are typically evaluated on their average performance with respect to error rates, ignoring the aspect of variability. Resampling techniques allow to assess the selection variability of individual features (voxels). Following the approach of Gordon, Chen, Glazko \& Yakovlev (2009) in the context of gene selection, we investigated whether variability on test results for BF and BH can be reduced by including both the significance and selection variability of the voxels in the decision criterion.},
  author       = {Durnez, Joke and Moerkerke, Beatrijs},
  booktitle    = {16th annual meeting of the organization for human brain mapping, Abstracts},
  keyword      = {Stability,Multiple testing problem,fMRI},
  language     = {eng},
  location     = {Barcelona, Spain},
  title        = {Improving selection stability of multiple testing procedures for fMRI},
  year         = {2010},
}