Hi Ajay,
Make sure you have to run palm_hemisplit on the outputs first, i.e.:
palm_hemisplit bh.thickness_*
Note that the "-o bh.thickness" was a somewhat poor choice of outputs as now it's necessary to use the underscore ("_") to avoid overwriting original inputs; for future runs, consider something as "-o bh.results" or similar.
Then load these split files on top of the surfaces.
The different results shown using white or pial make no sense I'm afraid. The surfaces don't contain data, just the geometry on top of which the results are displayed. Using a different surface on Freeview doesn't change the results.
All the best,
Anderson
On 19 September 2016 at 17:38, Ajay Kurani dr.ajay.kurani@gmail.com wrote:
Hi Anderson, My full design contrasts are below:
/ContrastName1 HC > Grp1 /ContrastName2 HC < Grp1 /ContrastName3 HC > Grp2 /ContrastName4 HC < Grp2 /ContrastName5 Grp1 > Grp2 /ContrastName6 Grp1 < Grp2 /ContrastName7 M > F /ContrastName8 M < F /ContrastName9 HC/Grp1 M/F Interaction /ContrastName10 HC/Grp2 M/F Interaction /ContrastName11 Grp1/Grp2 M/F Interaction /NumWaves 9 /NumPoints 11 /Matrix 1 1 -1 -1 0 0 0 0 0 -1 -1 1 1 0 0 0 0 0 1 1 0 0 -1 -1 0 0 0 -1 -1 0 0 1 1 0 0 0 0 0 1 1 -1 -1 0 0 0 0 0 -1 -1 1 1 0 0 0 1 -1 1 -1 1 -1 0 0 0 -1 1 -1 1 -1 1 0 0 0 1 -1 -1 1 0 0 0 0 0 1 -1 0 0 -1 1 0 0 0 0 0 1 -1 -1 1 0 0 0
My colums correspond to the following: EV1:HC-M EV2:HC-F EV3:Grp1-M EV4:Grp1-F EV5:Grp2-M EV6:Grp2-F EV7:Age EV8:Education EV9:Disease Severity
In the folder I was running the analysis I put the lh.thickness.10mm.mgz, rh.thickness.10mm.mgz, lh.white (fsaverage), rh.white (fsaverage), lh.mask.mgh (taken from running qdec initially), rh.mask.mgh (taken from running qdec initially).
I initially ran palm_hemimerge lh* within matlab Then from a terminal I ran the following command:
palm -i bh.thickness.10mm.mgz -d design.mat -t design.con -o bh.thickness -n 500 -approx tail -corrcon -s bh.white -T -tfce2D -logp -m bh.mask.mgz -nouncorrected
The command ran fully. When I loaded contrast 6 I found no results on the pial surface, however the white matter surface (where the mask was) was speckled all over with no cluster.
If there is a location,I can upload the stats file if that is easier.
Thanks, Ajay
On Sat, Sep 17, 2016 at 10:03 PM, Ajay Kurani dr.ajay.kurani@gmail.com wrote:
Hi Anderson, Thanks for the help. When viewing my results they looked very strange. Upon further investigation it looks as though the mask I supplied to PALM was a white matter mask (mask.mgh from running qdec initially) created when I ran qdec. I assumed this would be the whole cortex but I was wrong. Therefore it seems to only run permutation testing on the surface of the white matter. Due to the fact that it is unsmoothed white matter, I think this is why we see some speckling bleeding through near the boundaries
In order to do permutation testing accurately for surface based cortical thickness, would the mask need to be a volume file which is between the pial and white matter surfaces or would it just need to be the pial surface (lh.pial / rh.pial), or something else? Any suggestions on the best way to create this?
Thanks, Ajay
On Sat, Sep 17, 2016 at 1:03 PM, Ajay Kurani dr.ajay.kurani@gmail.com wrote:
Hi Anderson, Thanks for the help. When viewing my results they looked very strange. Upon further investigation it looks as though the mask I supplied to PALM was a white matter mask (mask.mgh from running qdec initially) created when I ran qdec. I assumed this would be the whole cortex but I was wrong. Therefore it seems to only run permutation testing on the surface of the white matter as seen in the attached photo. Due to the fact that it is unsmoothed white matter, I think this is why we see some speckling bleeding through near the boundaries
In order to do permutation testing accurately for surface based cortical thickness, would the mask need to be a volume file which is between the pial and white matter surfaces or would it just need to be the pial surface (lh.pial / rh.pial), or something else? Any suggestions on the best way to create this?
Thanks, Ajay
On Thu, Sep 15, 2016 at 1:46 PM, Ajay Kurani dr.ajay.kurani@gmail.com wrote:
Hello Freesurfer Experts, I was running permutation simulations on cortical thickness data and I had an issue with non-orthogonal covariates with mri_glmfit-sim -perm. I then tried FSL's PALM which is an extension of randomize to calculate threshold free stats. I saved the output as logp(which is similar to qdec I believe), however I have not been able to load the stats files correctly. The output of palm is lh.thickness_tfce.mgz for my various contrasts.
- Is .mgz the proper format for the stats files or do I need to
convert this to another type like .mgh etc?
- Can I display this in freeview or is another program needed? I also
tried tksurfer but when I loaded the stats file as an overlay nothing displayed. I want to make sure that the stats is loaded as an overlay in freeview/tksurfer and if so, do I need to select anything special so that it scales the logp values correctly?
Thanks, Ajay
Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer
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.