Most of my analysis tasks place in R or Python, with some AFNI at the start and a pinch of SPM. I currently have two 3D matrices, one containing magnitudes and the other containing p-values pertaining to those magnitudes.
I was looking for a better way to do cluster thresholding, and stumbled upon TFCE. Soon after found that FSL implements it. But looking at the fslmaths documentation, I'm not at all clear about how to apply it to my data. It takes two values (H and E, which I found quick descriptions of in the Randomise documentation) and "connectivity". What is connectivity? Do I just need to dump my pvalues into a nifti file (easy to do)? Any special header details need set, like intent codes? Or is it a 4D dataset, with magnitudes followed by pvalues? 1 minus pvalues?
|