Hi,
I am looking into the results of my functional data analysis, and would like to compute results changing from cluster-wise to TFCE correction. Since I am changing some preprocessing parameters, I am for the moment considering a single subject.
I am trying:
fslmaths zstat1.nii.gz -tfce outputimage --> in the terminal appears "Segmentation fault (core dumped)".
Could I use
randomise -i <4D_input_data> -o <output_rootname> -d design.mat -t design.con -n 500 -T, with preprocessed data, already demeaned by the grand-mean scaling automatically run by feat to see clusters that survive tfce correction?
If I want to use cluster command instead, just for cluster corrected results;
cluster -i zstat1.nii.gz -t 2.3 --pthresh=0.05 --dlh=? --volume=?
dlh= smoothing I used in my preprocessing in number of voxels? e.g voxel3X3X3 smoothing 6mm dlh=2?
volume= the multiplication of dim1Xdim2Xdim3 found with fslinfo command for the 4Dimage?
Do these commands account for the multiple testing problem in the group analysis, or how could I do if they don't?
Best,
Marco
|