Thank you for the thorough response Steve.

Best,
Krista

On Thu, Jul 14, 2011 at 3:32 AM, Stephen Smith <[log in to unmask]> wrote:
Hi - I strongly recommend that unless you really understand the details of the scripts/programs and GUIs, it is safest and simplest to carry out "standard" and "complete" analyses and not try to move single files around and perform part-analyses - hope that makes sense?

filtered_func_data does not know anything about the spatial transformations that were used to get it into the space it is currently in.

In general it's not good to re-feed preprocessed data again into the main processes such as registration, because it probably has (e.g.) been spatially smoothed, which would then reduce the quality of further registrations.

Hope this helps,
Cheers, Steve.



On 13 Jul 2011, at 06:26, Krista Wisner wrote:

Hi,

This may seem like a rudimentary topic/question, but I have not been able to find a clear answer and wanted to check before running my analyses.

Context:
I completed all of my preprocessing steps (motion correction, slice timing correction, BET, smoothing, high pass filtering, B0 unwarping; FLIRT for linear reg of func to structural; FNIRT for non-linear reg of structural to standard) in FEAT first so that I could check that these steps were completed successfully before moving into my analysis.

Question:
If I were to copy just the filtered_func_data.nii.gz files into a new directory somewhere else in order to run my group analysis on a different server, would all of the FLIRT/FNIRT efforts be lost?  Or is this information somehow embedded in the filtered_func_data.nii.gz file?  {I ask this since (1) upon loading the filtered_func_data.nii.gz into fslview, I see that this image file clearly shows no evidence of the transformations actually being performed ...  and since (2) the /reg folder created in the .feat directory during the preprocessing in FEAT would no longer be found with the filtered_func_data.nii.gz once it has been copied to it's new location on a different server.}

Follow-up Question.
- If indeed the FLIRT/FNIRT efforts are lost after copying the filtered_func_data.nii.gz files into a new directory, if I were to copy the required structurals into the same directory, could this registration (FLIRT/FNIRT) be completed "again" during the group level analysis in MELODIC?
- Or more simply, while copying the filtered_func_data.nii.gz for each person into the new directory, I could copy the /reg directory created during the FEAT preprocessing along with the it into the new location. Would this take care of making sure the registration completed correctly during the group-level analysis?

Thank you in advance,

Krista Wisner



---------------------------------------------------------------------------
Stephen M. Smith, Professor of Biomedical Engineering
Associate Director,  Oxford University FMRIB Centre

FMRIB, JR Hospital, Headington, Oxford  OX3 9DU, UK
+44 (0) 1865 222726  (fax 222717)
[log in to unmask]    http://www.fmrib.ox.ac.uk/~steve
---------------------------------------------------------------------------






--
Krista Wisner

Graduate Student
Clinical Science and Psychopathology Research Program
University of Minnesota
N427 Elliott Hall, 75 E. River Rd.
Minneapolis, MN 55455