I’m trying to follow the code for performing fixel-based analysis (FBA).
This is for DWI data of b=1000, 32-directional, and no reverse phase-encoding readout.
I note that within the related papers to both FBA and AFD, for each subject the spherical harmonic images are to be generated for lmax=4, for enabling co-registration to a population template.
Prior to this step I generated the response functions with the dhollander option and lmax values set to 0,4.
However, when running dwi2fod msmt_csd -lmax 0,4 -shell 0,1000 with the response functions from WM, GM, CSF etc - I receive the following error: “Number of lmaxes specified does not match number of tissues”
If I let the algorithm also go to default and do not specify the lmax, it rather runs smoothly (I’ve passed the -info and -debug options).
I’ve checked everything ranging from the embedded gradient direction files, and the response functions for each tissue appear to be in order. This makes me believe it may be a bug… For an example subject:
3632.371103 0 0
1885.697221 -707.6406179 158.278385