Durai-
A member of the AFNI team may have a better answer... I ran the 3dvolreg command on about 30 DTI scans that I had "around", but I do not see an intensity change (to that level). It's possible that it's related to a
previous discussed topic on Fourier interpolation having a difficult time with spikes in the data or other distortions. It could be that your particular subject has some large amount of motion during what could be 514 different DTI directions (is that right?). You can get some idea about the motion during the scan with commands like these:
3dvolreg -cubic -verbose -1Dfile motion.1D -prefix volreg.dti.nii.gz -base 0 inputfile.nii.gz
1d_tool.py -infile motion.1D -show_censor_count -censor_prev_TR -censor_motion 0.3 motion_out
1dplot -jpg motion_censored.jpg -censor motion_out_censor.1D motion.1D
Consider posting the image of motion over the 515 sub-briks in your datafile. If you do have motion in your DTI scan (or even if you do not), you may want to look into using
TORTOISE to try and resolve some of artifacts introduced by motion or other noise.
Peter
Edited 2 time(s). Last edit at 06/24/2013 03:36PM by Peter Molfese.