Skip to content
Standalone MATLAB implementation of permutation TFCE correction
Branch: master
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Type Name Latest commit message Commit time
Failed to load latest commit information.
demos Fixed demo.m error rate calculation to deal with cases with >1 false … Jul 13, 2018


Standalone MATLAB implementation of permutation TFCE correction

This package offers a standalone implemetation of multiple comparison correction for fMRI data. It achieves this through a permutation testing approach which controls familywise error rate by comparing voxelwise statistics to the maximal statistics obtained from repeating the analysis with randomized data. See Nichols & Holmes (2002) for a detailed treatment of this approach.

This maximal permuted statistic correction technique is combined with the threshold free cluster enhancement (TFCE) transformation due to Smith & Nichols (2009), which obviates the need for arbitrary voxelwise cluster-forming thresholds and instead produces continuous correct p-values for all voxels. Although some spatial specifity is lost relative to purely voxelwise approach, this approach, like cluster corrections, is substantially less conservative due to the fact that it capitalizes on spatial dependency in the data.

All functionality in the package can be accessed via matlab_tfce.m and that file also provides a description of the relevant parameters. The file matlab_tfce_gui.m provides a convenient user interface.

The following statistical tests can be computed:

'onesample' -- tests one sample hypothesis mean > 0

'paired' -- paired (dependent samples) test mean(imgs) > mean(imgs2)

'independent' -- independent (two sample) test mean(imgs) > mean(imgs2)

'correlation' -- correlation across subjects of imgs with covariate

'rm_anova1' -- one-factor repeated measures ANOVA

'rm_anova2' -- two-factor (with interaction) repeated measures ANOVA

'regression' -- multiple linear regression

Other than the ANOVAs, all tests permit both one- and two-tailed modes. See the demo script to observe proof of familywise error rate control under trivial conditions. The TFCE transformation itself and the paired/one-sample t-test routines have been validated directly against FSL's randomise using real fMRI data. Default parameter settings were also adopted from FSL.

See here for additional information:

You can’t perform that action at this time.