Convert_cdiflist_to_grads.py Problem

AFNI Message Board
1

Hi,

I tried using dcm2niix_afni to convert my GE DTI blip up / blip down dicom data to NIFTI format. In doing so, I realized that the .bval files that were created did not seem accurate. Every shell had the maximum bval of 1100 except for the first one, which was zero. I think this may be due to a problem with the GE Dicom headers. I found that a previous user in my group extracted the information about bvalues using a cdiflist file. Seeing that AFNI has a function called convert_cdiflist_to_grads.py to extract the gradients from this cdiflist file, I ran this command:


convert_cdiflist_to_grads.py -cdiflist cdiflist45 -bval_max 1100 -prefix test

My file called cdiflist45 looks like this:


45
0.00    0.00    0.00 
0.426    0.00    0.00 
0.19    0.381    0.00	
0.19    0.117    0.363 
0.00    0.00    0.00 
0.674    0.00    0.00 
0.301    0.603    0.00 
0.301    0.185    0.574 
1.00      0.00    0.00 
0.166 0.986 0.00	
-0.110 0.664 0.740 
0.901 -0.419 -0.110 
-0.169-0.601 0.781 
-0.815 -0.386 0.433 
0.656 0.366 0.660 
0.582 0.800 0.143 
0.900 0.259 0.350 
0.693 -0.698 0.178 
0.357 -0.924 -0.140 
0.543 -0.488 -0.683 
-0.525 -0.396 0.753 
-0.639 0.689 0.341 
-0.330 -0.013 -0.944 
0.00    0.00    0.00 
0.191    -0.308    -0.224 
0.19    -0.308    0.224 
-0.191    -0.118    0.362 
0.302    -0.487    -0.354 
0.301    -0.488    0.355 
-0.303    -0.187    0.573 
-0.524 -0.783 0.335 
0.609 -0.065 -0.791 
0.220 -0.233 -0.947 
-0.004 -0.910 -0.415 
-0.511 0.627 -0.589 
0.414 0.737 0.535 
-0.679 0.139 -0.721 
0.884 -0.296 0.362 
0.262 0.432 0.863 
0.088 0.185 -0.979 
0.294 -0.907 0.302 
0.887 -0.089 -0.453 
0.257 -0.443 0.859 
0.086 0.867 -0.491 
0.863 0.504 -0.025

When I run this, I receive this error:


++ ver : 0.1 (libver : 0.3)
** ERROR: Cannot have a ragged cdiflist

Do you know how I can fix this issue? Is there a way to use dcm2niix_afni so that I don’t need the cdiflist45 file or is this the current workaround?

Thanks!
Price

2

Hi, Price-

There is a line in your cdiflist text file where two columns have been merged:


-0.169-0.601 0.781

instead of


-0.169 -0.601 0.781

(line 14 in one-based counting)

If you fix that text file, then you should be good, methinks.

And me-also-thinks that you need to use the cdiflist if you have it, to have the correct gradients.

–pt

3

Huge help! That fixed the problem, thanks.

4

Great to know, thanks.

And, just for reference, I have now updated convert_cdiflist_to_grads.py to provide more useful and specific information to pinpoint where the “raggedness” is occurring, to aid in troubleshooting. That will be a part of the next AFNI build.

–pt