So, because mris_preproc does not already try to fit the data we use that for LMEs because that's where different model fits are then specified. Thanks a lot!
----- Ursprüngliche Mail ----- Von: "mreuter" mreuter@nmr.mgh.harvard.edu An: "Freesurfer support list" freesurfer@nmr.mgh.harvard.edu Gesendet: Donnerstag, 24. März 2016 18:18:58 Betreff: Re: [Freesurfer] REPOST: registering longitudinal data to template
Hi Clara,
long_mris_slopes is for the 2 stage approach. It computes a linear fit (within each subject) across time (using the base space), then it maps the results to the template (fsaverage) and applies smoothing. It can compute rate (slope of linear fit) or some percent change measures, also the average across time.
mris_preproc is a tool to stack and map data to the global template (fsaverage) and apply smoothing etc. It does not perform any fitting.
Best, Martin
On 03/24/2016 12:44 PM, Clara Kühn wrote:
Thank you, Martin! That clears things up a lot!
Could you explain what the difference is between the commands long_mris_slopes and mris_preproc? As far as I understand they give you a different output. But what do they do differently in respect to mapping longitudinal data?
Cheers, Clara
----- Ursprüngliche Mail ----- Von: "mreuter" mreuter@nmr.mgh.harvard.edu An: "Freesurfer support list" freesurfer@nmr.mgh.harvard.edu Gesendet: Mittwoch, 23. März 2016 19:01:42 Betreff: Re: [Freesurfer] REPOST: registering longitudinal data to template
Hi Clara,
- You'd create a different table for your time point with fsid being
the longitudinal directories s1.long.sbase , no fsid-base column. then you can map and stack these using cross sectional commands (mris_preproc --qdec <thistable> ...) 2) You should use Qdec or glm_fit for cross sectional analysis 3) Yes, only for 2 stage model, so don't use it (for long you use LME anyway and for cross see 2)) 4) that commands create a table with one entry for each subject (referring to the base), this is also for 2 stage model (the second stage works on a subject by subject basis) 5) the -qcache in recon all maps and smoothes surface data to the template, so that it is available later. You can do that even on the long runs, or you can just manually map ans smooth things with preproc. 6) A different preproc call using a fsgd file rather than a qdec table. see the glm_fit tutorial for these things.
Best, Martin
On 03/23/2016 12:28 PM, Clara Kühn wrote:
Dear FreeSurfer experts,
I would like to analyze my longitudinal data also cross-sectionally at single time points. I am a little confused as to how to register my data to the template and smooth it accordingly.
So far, I have registered my longitudinal data to my template to use in LMEs with this command: mris_preproc --qdec-long ./qdec/long.qdec.table.dat --target 30kids_template --hemi lh --meas thickness --out lh.thickness.mg
and then smoothed it with this command: mri_surf2surf --hemi lh --s study_average --sval lh.thickness.mgh --tval lh.thickness_sm10.mgh --fwhm-trg 10 --cortex --noreshape
(1) Do I have to do it once with a cross-sectional command and once with a longitudinal command? (2) Should I do the cross-sectional analyses in QDEC or with LME?
Now the question is what these other commands are for?
(3) long_mris_slopes --qdec ct/qdec/long.qdec.table.dat --meas thickness --hemi lh --do-avg --do-rate --do-pc1 --do-spc --do-stack --do-label --time years --qcache kids_template --sd $SUBJECTS_DIR --> I understood this is only for Two-Stage-Models, is that correct?
(4) long_qdec_table --qdec qdec/long.qdec.table.dat --cross --out qdec/cross.qdec.table.dat
(5) recon-all -s name -qcache -target name_of_template
(6) mris_preproc --fsgd gender_age.fsgd --cache-in thickness.fwhm10.fsaverage --target fsaverage --hemi lh --out lh.gender_age.thickness.10.mgh
Please help clear up my confusion :) Thanks in advance, Clara