On 7/25/18 8:31 AM, John Anderson wrote:
External Email - Use Caution
Dear Petsurfer experts,
I used partsurfer in FS6 and had applied partial volume correction
(PVC) on PET data. I followed the same steps as in wiki which was very
well explained! Thank you!.
The analyses went fine... I would like to inquire about some issues,
and I highly appreciate any clarification:
1. The flag --psf in "mri_gtmpvc": Is the value of "psf" means the
amount of smoothing (fwhm in mm) that will be applied on the corrected
maps? If I give "psf" a value of "zero", does this means that the
corrected maps are no longer going to be smoothed? The motivation of
the question is that I would like to smooth the images at the end of
the analysis, so I can apply the amount of smoothing on the PVC images
vs analysis involve non-pvc images.
It has nothing to do with applied smoothing but rather the amount of
blurring incurred during the PET acquisition and reconstruction process.
It models the blurring function as an isotropic Gaussian filter of the
given FWHM. Any smoothing you apply will be applied later. Note that
applying volume smoothing to PV-corrected data is invalid. Surface
smoothing is the only valid way to smooth PVCed data.
2. For some of my group analyses, I don't see any difference between
the groups. When I apply PVC using "mri_gtmpvc", I see difference in
pathologically relevant regions, and vice versa for some other
analyses where I see difference between the groups without PVC then
the difference disappear following PVC. I highly appreciate if you
please clarify why PVC can find results or eliminate findings. What is
the idea of partial volume effect PVE in this case?
This is a big topic. See https://www.ncbi.nlm.nih.gov/pubmed/26915497.
That paper shows that atrophy can make the apparent effect of age on FDG
be bigger than it really is. This is because PVEs reduce the FDG signal
in GM which tends to be the same direction as the actual effect (ie, the
PVEs reinforce what is already there making it too big and too
significant). For another tracer where one expects the uptake to go down
with age or disease (eg, PiB), then PVEs can make effects go away.
3. Group analysis: following PVC, I used "mri_concat" to combine the
images "?.mgx.gm.nii.gz", then I smoothed the final combined images
using the flag "fwhm" in "mri_surf2surf". I used fwhm=3mm and then
fwhm=6mm. For all analyses I can see clusters of difference between
the groups at fwhm=3mm. For some analyses the difference disappears
when I use fwhm=6mm . Is this till a partial volume effect even after
applying pvc or this is relevant to an extra smoothing effect which
obscure the PET signal? is there any recommended fwhm value for PET
images?
The FWHM you chose to smooth by for analysis purposes is somewhat
arbitrary. The ideal FWHM depends on the size of the activation you are
expecting, which you don't know until you do the analysis, which you
can't do until you set the FWHM. If you have some idea about the size of
the effect you expect, then you can set the FWHM to that level.
Otherwise, people just fall back to typical values between 5-10.
Thank you for your guidance,
John
_______________________________________________
Freesurfer mailing list
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
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.
