Dear Doug,
Thank you very much! I am ready to start doing my analysis but before I start you were mentioning that for the simulation 10,000 is a good number. You also suggested me to run:
mri_glmfit-sim \ --glmdir lh.AGE.glmdir \ --cache 4 neg \ --overwrite
but what is the number of the iteration if I run this part? Do I need to add this and do I need to mention the threshold in the above coding?
THANKS! Antonella
________________________________ From: Douglas N Greve greve@nmr.mgh.harvard.edu To: Antonella Kis atorok9@yahoo.com Cc: "freesurfer@nmr.mgh.harvard.edu" freesurfer@nmr.mgh.harvard.edu Sent: Wednesday, August 17, 2011 1:10 PM Subject: Re: [Freesurfer] mri_glmfit-sim
Antonella Kis wrote:
Hi Doug,
Thanks again for your help. There are few more things not very clear for me and I will be very graetful if you can advise me:
1). Did you mean that if I did not qcache my data while I was doing pre-processing running the recon-all I should use then lh.AGE.thickness.10B.mgh?
yes
- Are then the mentioned steps the right one to be followed?
yes
- I saw on one tutorial about the mri_glm to perform estimations.
What are this estimations? Do I need to run:
Where did you see reference to mris_glm? This is an old program that we should not even distribute.
# For the left hemisphere mris_glm --surfmeas thickness \ --hemi lh \ --trgsubj average \ --fsgd ./my_gender_age_fsgd.txt doss \ --beta ./beta_doss-thickness-100lh.mgz\ --var ./var_doss-thickness-100lh.mgz \ --y ./y_doss-thickness-100lh_000.mgz \ --nsmooth 100 # For the right hemisphere mris_glm --surfmeas thickness \ --hemi rh \ --trgsubj average \ --fsgd ./my_gender_age_fsgd.txt doss \ --beta ./beta_doss-thickness-100rh.mgz \ --var ./var_doss-thickness-100rh.mgz \ --y ./y_doss-thickness-100rh_000.mgz \ --nsmooth 100 4) If I have less than 80 subjects do I need to run the full MC simulation and I must supply the smoothest of my data as fwhm from the y.fsgd file? Also I would need to mention the threshold like in the following example?
Yes. You are better off using mri_glmfit-sim rather than the command line below. mri_glmfit-sim will run that program giving it the appropriate fwhm.
mri_glmfit --y lh.gender_age.thickness.10.mgh \ --glmdir lh.gender_age.glmdir \ --fsgd gender_age.txt doss \ --surf fsaverage lh \ --fwhm 14.517 --C age.mat \ --sim mc-full 10000 2 lh.gender_age.glmdir/csd1
Do I add the --cache 4 neg \ --overwrite
at the end?
- When I run tksurfer to view the sig.mgh file and I set the
threshold to 2, meaning vertices with p<.01, I have just few vertices coloured in blue. Does this mean a low activation? If yes, what I have to do?
Yes. You can try lowering it to 1.3 (p<.05). If the effect is too subtle, you will need to add more subjects (assuming there is an effect at all). doug
Thank you very much! Antonella
*From:* Douglas N Greve greve@nmr.mgh.harvard.edu *To:* Antonella Kis atorok9@yahoo.com *Cc:* "freesurfer@nmr.mgh.harvard.edu" freesurfer@nmr.mgh.harvard.edu *Sent:* Wednesday, August 17, 2011 11:53 AM *Subject:* Re: [Freesurfer] mri_glmfit-sim
Antonella Kis wrote:
Dear Doug,
Thank you very much for your valuable reply.
I still have few questions on which I would like to ask your advise:
QUESTION 1:
While running the recall-all for all my subjects I also used the -qcache option. I am not sure how this data is further used for my GLM analysis.
I will appreciate if you can confirm if in this case in order to do a thickness-age correlation group study I need to complete the following steps:
STEP 1: Uncached Data: resamples each subjects data to fsaveragea (into a common space), and Concatenating all the subjects' into a single file by running
mris_preproc --fsgd AGE_fsgd.fsgd \ --target fsaverage --hemi lh \ --meas thickness \ --out lh.AGE.thickness.00.mgh
STEP 2: The independent variable is the thickness smoothed at various FWHM (full-width/half-max), usually 0, 5, 10, 10, 20, and 25mm so run:
mris_preproc --fsgd AGE_fsgd.fsgd \ --cache-in thickness.fwhm10.fsaverage \ --target fsaverage --hemi lh \ --out lh.AGE.thickness.10.mgh
STEP 3: If I run the next step (I know is optional) where I can use this data lh.AGE.thickness.10B.mgh? Is this a better data to be use for my further steps?
You would use it in the same place that you would use lh.AGE.thickness.10.mgh. It is just a different way to compute it if you did not qcache.
OPTIONAL: THIS WILL TAKE ABOUT 5 MINUTES
mri_surf2surf --hemi lh \ --s fsaverage \ --sval lh.AGE.thickness.00.mgh \ --fwhm 10 \ --cortex \ --tval lh.AGE.thickness.10B.mgh
STEP 4: GLM Analysis (mri_glmfit)
mri_glmfit \ --y lh.AGE.thickness.10.mgh \ --fsgd AGE_fsgd.fsgd dods\ --C lh-Avg-thickness-age-Cor.mtx \ --surf fsaverage lh \ --cortex \ --glmdir lh.AGE.glmdir
STEP 5: View the uncorrected significance map with tksurfer:
tksurfer fsaverage lh inflated \ -annot aparc.annot -fthresh 2 \ -overlay lh.AGE.glmdir/lh-Avg-thickness-age-Cor/sig.mgh
STEP 6: Viewing the medial surface, change the overlay threshold to something very, very low (say, .01), View --> Configure --> Overlay, set "Min" .01
STEP 7: CSD (Cluster Simulation Data)- Run the Simulation
mri_glmfit-sim \ --glmdir lh.AGE.glmdir \ --sim mc-z 5 4 mc-z.negative \ --sim-sign neg \ --overwrite
QUESTION 2: You mentioned that I should use the --cache option to mri_glmfit-sim. How I should do this? Do I need to change STEP 7?
Yes, use mri_glmfit-sim \ --glmdir lh.AGE.glmdir \ --cache 4 neg \ --overwrite
QUESTION 3: How I can check my data to make sure that I have actual activation above the voxel-wise threshold and how I can change my threshold? Based on what I need to change my threshold? Do I change my treshold until I find some activation?
Run tksurfer to view the sig.mgh file. Set the threshold to the one you used in mri_glmfit-sim using the view->configure->overlay. There are no hard and fast rules to setting thresholds, but people usually use p<.05 or p<.01.
doug
Thank you very much for your valuable help. Antonella
*From:* Douglas N Greve <greve@nmr.mgh.harvard.edu
mailto:greve@nmr.mgh.harvard.edu>
*To:* Antonella Kis <atorok9@yahoo.com mailto:atorok9@yahoo.com> *Cc:* "freesurfer@nmr.mgh.harvard.edu
mailto:freesurfer@nmr.mgh.harvard.edu" <freesurfer@nmr.mgh.harvard.edu mailto:freesurfer@nmr.mgh.harvard.edu>
*Sent:* Tuesday, August 16, 2011 12:04 PM *Subject:* Re: [Freesurfer] thickness-age correlation
Antonella Kis wrote:
Dear Freesurfer experts,
I am doing a thickness-age correlation group difference study
(patients versus controls. I would like to know if:
- my contrast vector defined as 0 0 0.5 0.5 is correct in order to
test the change in thickness with age
Yes.
2)what is the best iteration number for the simulation? Should be
5000 or 1000 or greater?
10,000 -- note that if you're doing a whole-hemisphere correction with the monte-carlo simulation then you should use the pre-computed results (use the --cache option to mri_glmfit-sim)
3)why when I run the simulation for 5 iterations I've got zero
clusters?
You should look at your data to make sure that you have actual activation above the voxel-wise threshold you have set. Also, only 5 clusters would not be enough to get a significant cluster anyway.
4)can I derive the thickness from the significant clusters of glm
analysis? Do I need to run QDEC after GLM?
If you mean that you want the average thickness from each subject for a given cluster, then this information is generated with mri_glmfit-sim in the csdbase.y.ocn.dat file.
doug
Thanks for any enlightenment. Antonella
Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu
mailto:Freesurfer@nmr.mgh.harvard.edu <mailto:Freesurfer@nmr.mgh.harvard.edu mailto:Freesurfer@nmr.mgh.harvard.edu>
https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer
-- Douglas N. Greve, Ph.D. MGH-NMR Center greve@nmr.mgh.harvard.edu mailto:greve@nmr.mgh.harvard.edu
<mailto:greve@nmr.mgh.harvard.edu mailto:greve@nmr.mgh.harvard.edu>
Phone Number: 617-724-2358 Fax: 617-726-7422
Bugs: surfer.nmr.mgh.harvard.edu/fswiki/BugReporting FileDrop: www.nmr.mgh.harvard.edu/facility/filedrop/index.html
The information in this e-mail is intended only for the person to whom it is addressed. If you believe this e-mail was sent to you in error and the e-mail contains patient information, please contact the Partners Compliance HelpLine at http://www.partners.org/complianceline . If the e-mail was sent to you in error but does not contain patient information, please contact the sender and properly dispose of the e-mail.
-- Douglas N. Greve, Ph.D. MGH-NMR Center greve@nmr.mgh.harvard.edu mailto:greve@nmr.mgh.harvard.edu Phone Number: 617-724-2358 Fax: 617-726-7422
Bugs: surfer.nmr.mgh.harvard.edu/fswiki/BugReporting FileDrop: www.nmr.mgh.harvard.edu/facility/filedrop/index.html
Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu mailto:Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer