There is usually three different approaches to analyze longitudinal data (and yours is longitudinal as you have 2 time points and are interested in change).
- linear mixed effects models (matlab), recommended
- 2 stage approach (long_mris_slopes), old approach, and superseded by LME. Can be still used if you have the same number of time points per subject and the time deltas are the same (or similar) for all subjects..
- paired difference, for very basic designs (with only 2 time points per subject and same time delta). I have done this once many years ago and did all steps manually (computed the difference, set up my own matrices (no FSDG) and ran glmfit. There is also a repeated measure anova option.
https://surfer.nmr.mgh.harvard.edu/fswiki/LongitudinalStatistics gives an overview with relevant links.
Based on your description I think you are trying the last approach. So you would not need to run long_mris_slopes. You need two FSGD files for that (one with all time points, the second with the pairs.
You should be following :
http://surfer.nmr.mgh.harvard.edu/fswiki/PairedAnalysis
and probably Doug can help with that pipeline.
Best, Martin
On 05/14/2017 11:57 PM, Maria A Infante wrote:
do I need to run long_mris_slopes on the --qdec.long table? the example is not for a longitudinal analysis. I have 2 time points and 3 groups.
M. Alejandra Infante, Ph.D. Postdoctoral Fellow UC San Diego Department of Psychiatry Adolescent Brain Cognitive Development (ABCD) Study
*From:* freesurfer-bounces@nmr.mgh.harvard.edu [freesurfer-bounces@nmr.mgh.harvard.edu] on behalf of Martin Reuter [mreuter@nmr.mgh.harvard.edu] *Sent:* Sunday, May 14, 2017 1:14 PM *To:* Freesurfer support list *Subject:* Re: [Freesurfer] 2 stage model longitudinal data analysis
Hi Maria,
yes, the example on the wiki is a simple “one sample group mean” (osgd). You can find examples with more than one group here:
https://surfer.nmr.mgh.harvard.edu/fswiki/FsgdExamples
Best, Martin
On 13 May 2017, at 00:24, Maria A Infante <minfante@ucsd.edu mailto:minfante@ucsd.edu> wrote:
Hi, I am interested in looking at groups differences (3 groups) in percent thickness change. I want to use mri_glmfit to analyze but I am unclear on what I should include on my fsgd file. Can you please provide an example?
Also The example provided on the fswiki under Option mri_glmfit doesn't call for an fsgd file.
long_mris_slopes --qdec long.testretest.qdec \ --meas thickness \ --hemi lh \ --sd $SUBJECTS_DIR \ --do-pc1 --do-label \ --generic-time \ --fwhm 15 \ --qcache fsaverage \ --stack-pc1 lh.testretest.thickness-pc1.stack.mgh \ --isec-labels lh.testretest.fsaverage.cortex.label
mri_glmfit --osgm \ --glmdir lh.testretest.thickness-pc1.fwhm15 \ --y lh.testretest.thickness-pc1.stack.fwhm15.mgh \ --label lh.testretest.fsaverage.cortex.label \ --surf fsaverage lh
Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu mailto:Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer
Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer