These are random bash shell scripts for FSL that I find useful.
CompROI.sh ROItemplate "3:L_Brain 4:R_Brain" ImgList.csv Result.csvor, can merge multiple values into a single ROI
CompROI.sh ROItemplate "3+4:Brain" ImgList.csv Result.csv
easythresh
, this script takes two statistic images and does a conjunction analysis, testing the 'Conjunction Null Hypothesis. The interpretation of significant regions is that there is evidence of effects in both contrasts tested (not just either/or). There is no assumption of independence required between the two effects tested. For more see my conjunction research page. fsl_fdr.sh stats/zstat1 mask stats/zstat1_fdrcorrp
will produce a 1-PFDR image called zstat1_fdrcorrp in the stats directory. To additionally create an rendered image, use the -rend
option, like fsl_fdr.sh -rend example_func rendered_thresh_zstat1_fdrcorrp \
stats/zstat1 mask stats/zstat1_fdrcorrp
For 1-P images from randomise, use the -1mp
option, as in: fsl_fdr.sh -1mp results_vox_p_tstat1 mask results_vox_corrp_tstat1
Finally, if you've got SPM T images, and $dof
is the degrees-of-freedom, you can use fsl_fdr.sh -Tdf $dof spmT_0001 0 spmT_0001_Pfdr
where setting the mask name to "0" has the effect of using an implicit mask (<>0 means in the brain). fslmax.sh *corrp*nii*
which will show you the best 1-minus-P for each corrected P-value image.fslstats
tool with multiple files (generalization of fslmax.sh
).fslinfo
tool with multiple files, creating a tabular report for the multiple files (optionally a CSV report).PlotFeatMFX.sh
is directly called by the user, like PlotFeatMFX.sh Nback.gfeat 23 37 57
or PlotFeatMFX.sh -usemm Nback.gfeat 44 -52 42
but of course requires R.