Dear TBSS experts I'm trying to analyze MD maps between 2 groups, after obtaining pretty satisfactory results of FA analysis. I have followed the steps specified in the TBSS website, however when I typed 'tbss_non_fa MD', the following error messages appear in the terminal. kim@kim-desktop:~/TBSS/FA_CONTROL_LHS$ tbss_non_FA MD using pre-chosen registration target: target upsampling alternative images into standard space C01_FA ** ERROR (nifti_image_read): failed to find header file for '../MD/C01_FA' ** ERROR: nifti_image_open(../MD/C01_FA): bad header info Error: failed to open file ../MD/C01_FA ERROR: Could not open image ../MD/C01_FA Image Exception : #22 :: Failed to read volume ../MD/C01_FA terminate called after throwing an instance of 'RBD_COMMON::BaseException' Aborted . . . (the same error messages are replicated in each subject) merging all upsampled MD images into single 4D image projecting all_MD onto mean FA skeleton now run stats - for example: randomise -i all_MD_skeletonised -o tbss_MD -m mean_FA_skeleton_mask -d design.mat -t design.con -n 500 --T2 -V (after generating design.mat and design.con) My questions are 1) From the messages, I quess that, in addition to MD maps, all original FA maps should be copied to the MD directory. Is it right? 2) Despite this message, I found that 'all_MD.nii.gz' and 'all_MD_skeletonised.nii.gz' files were created in the 'stats' directory. It is quite strange that, when I check these two files using mricron, these are not MD maps but exactly same with FA map. After randomising, I found that the results were identical to those of FA analysis, as is expected. I do not understand why this happens, because I strictly complied with the methods indicated in the online manual. I'm using FSL 4.1.5 and Ubuntu 9.10. Please let me know the right way. Thank you in advance for your kindness. Kim