help
help > RE: tissue-thrpobability maps for aCompcor
Jul 10, 2012 08:07 AM | Alfonso Nieto-Castanon - Boston University
RE: tissue-thrpobability maps for aCompcor
In the default conn settings the
normalization and segmentation steps are performed simultaneously
(using SPM combined normalization/segmentation procedure starting
from the anatomical volumes; SPM uses an iterative procedure that
takes advantage of the tissue probability maps to aid in the
normalization of the anatomical volumes), we use the posterior
probability maps in normalized space resulting from this procedure
as Gray/White/CSF masks, which are then thresholded and eroded. The
resulting masks are then used to extract the BOLD timeseries from
the (normalized) functional volumes. Let me know if this
clarifies.
Best
Alfonso
Originally posted by Pär Flodin:
Best
Alfonso
Originally posted by Pär Flodin:
> Thanks very much for informative
answers!
>
Part of my question remains however: How are the white matter and CSF masks
(used in the compCor procedure) obtained? Do you use the posterior
probability maps for WM and CSF (>50%, one voxel erosion) obtained from
NORMALIZED or NON-NORMALIZED (but then likely co-registered to the
functional volumes) subject-specific anatomical images? In other words, do
you first normalize the anatomical image and then segment it, or do you
first segment subject-specific anatomical scans and use these
non-normalized segmentation outputs for creating the WM and CSF masks? Do
you extract the wm and CSF time-series from normalized or from
non-normalized functional volumes?
Best
P�r
>
Part of my question remains however: How are the white matter and CSF masks
(used in the compCor procedure) obtained? Do you use the posterior
probability maps for WM and CSF (>50%, one voxel erosion) obtained from
NORMALIZED or NON-NORMALIZED (but then likely co-registered to the
functional volumes) subject-specific anatomical images? In other words, do
you first normalize the anatomical image and then segment it, or do you
first segment subject-specific anatomical scans and use these
non-normalized segmentation outputs for creating the WM and CSF masks? Do
you extract the wm and CSF time-series from normalized or from
non-normalized functional volumes?
Best
P�r
Threaded View
| Title | Author | Date |
|---|---|---|
| Pär Flodin | Jun 1, 2012 | |
| Alfonso Nieto-Castanon | Jul 8, 2012 | |
| Pär Flodin | Jul 9, 2012 | |
| Alfonso Nieto-Castanon | Jul 10, 2012 | |
| Pär Flodin | Jul 10, 2012 | |