fdr_threshold

Compute False Discovery Rate (FDR) threshold for statistical maps.

fdr_threshold [options] <input> <df> <significance %> [output]

DESCRIPTION

fdr_threshold computes the False Discovery Rate (FDR) threshold for a statistical map (t-statistic or F-statistic volume). The method is based on Genovese, Lazar, and Nichols (2002), “Thresholding of Statistical Maps in Functional Neuroimaging Using the False Discovery Rate” (NeuroImage, 15:870-878).

The tool reads an input statistical map, the degrees of freedom, and the desired significance level (as a percentage), and outputs the corresponding FDR threshold value. Optionally, it can write a corrected p-value volume.

OPTIONS

--verbose

Print progress information during processing.

--clobber

Overwrite existing output files.

--mask mask

Restrict computation to voxels within the specified binary mask.

--twosided

Perform a two-sided test (consider both tails of the distribution).

--negative

Invert t-statistics for single-sided tests (use negative tail).

--pval

Print the corrected p-value rather than the threshold.

--fstats

Treat the input as F-statistics instead of t-statistics.

--df2 n

Specify the second degree of freedom (for F-statistics).

--independent

Use the less conservative independence assumption for FDR correction.

--out output-pval.mnc

Write a corrected p-value volume to the specified file.

EXAMPLES

Compute FDR threshold at 5% significance for a t-map with 20 degrees of freedom:

fdr_threshold tmap.mnc 20 5

Compute two-sided FDR threshold with a brain mask:

fdr_threshold --twosided --mask brainmask.mnc tmap.mnc 20 5

Compute threshold for F-statistics with two degrees of freedom:

fdr_threshold --fstats --df2 100 fmap.mnc 3 5

Save corrected p-value map:

fdr_threshold --out pval_corrected.mnc tmap.mnc 20 5

AUTHOR

Vladimir S. Fonov - McConnell Brain Imaging Centre, Montreal Neurological Institute, McGill University.

COPYRIGHTS

Copyright © 2011 by Vladimir S. Fonov

SEE ALSO

mincstats, minccalc