When binaries (libraries and commands) are built with different (newer) compilers and/or run on different (newer) versions of an operating system, then the numerical results aren’t guaranteed to be the same between between releases. See the IMPORTANT NOTE at the bottom of the downloads page, https://surfer.nmr.mgh.harvard.edu/fswiki/rel7downloads

Apart from being built with newer compilers and operating systems, how specific commands perform computations may also have changed in version 7 compared to version 5. See the release notes for versions 6 and 7 on this page, https://surfer.nmr.mgh.harvard.edu/fswiki/ReleaseNotes

If you are interested in the development history of an individual command, then you could look at the source on GitHub and read the log of changes. For example, to look at the development history for mri_glmfit, you could route to https://github.com/freesurfer/freesurfer/tree/dev/mri_glmfit and then click on the text for “History” near the top of the page on the left hand side.

On Nov 18, 2021, at 14:03, Brusaferri, Ludovica <LBRUSAFERRI@mgh.harvard.edu> wrote:

Hi!

I am using a pipeline that was implemented and validated using fs 5.3 , which does the following:

* compute transformation matrix for linear registration of SUV with T1-2mm, with spm
* use tkergister to transform the matrix into .mat, and then invert the matrix
* use flirt to move T1 images into PET space

The last step gives me different results if I enable FS vs 7, and the registration seems to be off. Do you know why this would be the case?

Thank you!
Ludovica_______________________________________________
Freesurfer mailing list
Freesurfer@nmr.mgh.harvard.edu
https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer