- We should have the goal to develop this guide to generate a detailed documentation how the DWI data from the CBD clinical trial was processed
First follow the batman tutorial for preprocessing DWI data, the steps described here assume that you have just finished dwifslpreproc and the output is DWI_preproc.mif
dwi2tensor DWI_preproc.mif DWI_tensor.mif
Next calculate the FAC (fractional anisotropy color) . The step involves two processes, specifically, first calculating the direction from the tensor and the second taking the absolute value of that and writing this into a nii.gz file, which can be viewed in ITKSnap
tensor2metric DWI_tensor.mif -vec - | mrcalc - -abs DWI_FAC.nii.gz
calculate a response function: we want wm.txt
dwi2response dhollander DWI_preproc.mif wm.txt gm.txt csf.txt
Do the deconvolution (we want an FOD file i.e. fibre density distribution)
dwi2fod msmt_csd DWI_preproc.mif wm.txt DWI_FOD.mif
A FOD file has about 46 volumes it can be inspected using mrview
For visual inspection using ITKSnap but also aligment and such, extract the first volume of the FOD file
mrconvert DWI_FOD.mif -coord 3 0 -axes 0,1,2 DWI_FOD_firstVol.nii.gz
The last step would be to extract the DEC from thiss FOD image.
In principle that command is fod2dec
but there are a few other things, such as having a T1.nii.gz ready and aligned. Also, while in principle you can make a decent DEC from a single DWI scan, we ultimately want to derive it from BOTH DWI scans. To do that we need calculate an FOD from each, align the two FODs and then average.
Almost the same but sorta different. FAC is derived from a tensor and DEC can be derived from anything, possebly a tensor, but here in this case from the FOD.