There are a number of fields in the Analyze header which are not
consistently interpreted, so we give them default values when
creating an Analyze file.  It is unlikely that anything of interest has
changed.  If you are interested you can pinpoint the things that have
changed by doing the following:
  - to check the header, run "fslhd" on the initial and final Analyze  
	and look for the differences (you can save the result of each in a
	file and then run "diff" on the two files)
  - to check the image contents, run "fslmaths initialimage -sub  
finalimage diff"
	with the appropriate filenames for the initial and final images, and
	then run "fslstats diff -R"
	The output should be values equal to (or very near) zero, if there is
	no difference in the intensities stored by the images.

All the best,

On 25 Aug 2009, at 15:50, Daniel Peterson wrote:

> Hi FSLers,
>   I was playing around with fslchfiletype, and I took an md5 sum of  
> an analyze .img,
> converted it to .nii.gz, then back to analyze and the md5 changed. I  
> couldn't find any
> obvious thing that had changed.
>   Just wondering if anyone out there has a good idea of what might  
> kind of subtle little
> things might be changed when converting from Analyze to g-zipped  
> Nifti and back.
> Thanks!
> -Dan