Dear FS team,

I followed the BuildYourOwnMonteCarlo guide and received an error running mri_glmfit-sim. This might be a simple problem but I have troubles finding the flaw. Hopefully you can help me out.

Commands:

% step 1: If not done already - create the label:

mri_annotation2label --subject fsaverage --hemi rh --outdir labeldir

% step 2: mri_mcsim using my own label on the fsaverage surface:

mri_mcsim --o .../S1M1_GroupAnalysis/S1M1 --base mc-z --save-iter --surf fsaverage rh --nreps 10000 --label .../S1M1_GroupAnalysis/S1M1.label

% step 3: mri_glmfit investigating the correlation between an AUC value and t1t2ratio

mri_glmfit --y rh.t1t2ratio_fmed_3sm_decurv.fsavg.5.mgh --fsgd AUC.fsgd --surf fsaverage rh --C intercept.mtx --C slope.mtx --label S1M1.label --o t1t2vsAUC_S1M1.5.glmdir

% step 4: mris_glmfit-sim

mri_glmfit-sim --glmdir t1t2vsAUC_S1M1.5.glmdir --cache-dir .../S1M1_GroupAnalysis/S1M1 --cache 2 pos --cwp 0.05

Error:

ERROR: cannot find .../S1M1_GroupAnalysis/S1M1/fsaverage/rh/cortex/fwhm09/pos/th20/mc-z.csd

Question:

Is there any obvious flaws in my commands?

In the guide I'm suggested to place my results into $FREESURFER_HOME/... is this a necessity?

How does mri_glmfit-sim recognizes that I'm investigating the label S1M1 and not cortex?

Could i simply just copy-paste my results from mri_mcsim into .../S1M1_GroupAnalysis/S1M1/fsaverage/rh/cortex/ ?

Best, Silas

From: freesurfer-bounces@nmr.mgh.harvard.edu <freesurfer-bounces@nmr.mgh.harvard.edu> on behalf of Douglas Greve <greve@nmr.mgh.harvard.edu>
Sent: Monday, February 8, 2016 4:35 AM
To: freesurfer@nmr.mgh.harvard.edu
Subject: Re: [Freesurfer] FreeSurfer group analysis: mri_glmfit-sim

On 2/7/16 1:34 PM, Silas wrote:

Dear FS team,

I'm currently making a group analysis using the "command-line" group analysis stream in freesurfer.

https://surfer.nmr.mgh.harvard.edu/fswiki/FsTutorial/GroupAnalysisDng

FsTutorial/GroupAnalysisDng - Free Surfer Wiki

surfer.nmr.mgh.harvard.edu

Introduction. This tutorial is designed to introduce you to the "command-line" group analysis stream in FreeSurfer (as opposed to QDEC which is GUI-driven), including ...

1) I would like to do a surface based analysis where i'm only investigating a certain area or label - instead of doing a whole brain analysis. This can be done by introducing --label surface_area_of_interest.label when running mri_glmfit. Would this (in theory) give me different results from doing a whole brain analysis - or would it just exclude the results from outside the label?

No. However, if your activation is near the boundary of the ROI, then you may want to constrain the spatial smoothing to be within the ROI.

And how about after clusterwise correction for multiple comparisons?

If your new ROI is much smaller than whole hemisphere, then you will need to generate new correction tables (otherwise it won't help to constrain it to an ROI). To do this, see https://surfer.nmr.mgh.harvard.edu/fswiki/BuildYourOwnMonteCarlo

-> is there a better way to do this than following the command-line group analysis stream in freesurfer?

2) When running mri_glmfit-sim i'm investigating the output ending with -voxel.mgh - how do i interpret the results of this file? When investigating the file in freeview what is the meaning of the units when configuring the overlay?

This is the voxel-wise correction for multiple comparisons. The value of the voxel (and overlay config units) are -log10(p), where p is the p-value corrected for multiple comparisons. So p=.01 means 2.

Best, Silas

_______________________________________________
Freesurfer mailing list
Freesurfer@nmr.mgh.harvard.edu
https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer