Hi Doug,
My issue is I am not intending to use a labeled segmented ROI. My main goal
is to calculate the connectivity strength between two binarized ROIs that
are functionally defined. Can fcseed-config accept non Freesurfer segmented
ROIs?
Thanks,
JIahe
On Fri, Sep 12, 2014 at 11:08 AM, Douglas
You need to have your ROI as a segmentation in the anatomical
(conformed) space. If it is a label on the surface, you can use
mri_label2vol to turn it into a volume (make sure to use --proj frac 0 1
.1 to fill the ribbon). When you run funcroi-config specify -seg
yourseg.mgz 1 where yourseg.m
Hi Doug,
I was able to get the surface connectivity analysis to run successfully for
the L.Cingulate example that you posted online (
http://surfer.nmr.mgh.harvard.edu/fswiki/FsFastFunctionalConnectivityWalkthrough
). For my own analysis, I'm hoping to use my own functionally defined ROIs
instead
How did you look at fmcpr.up.sm6.fsaverage.lh.nii.gz? If you did not do
this, try it
tksurfer fsaverage lh inflated -ov fmcpr.up.sm6.fsaverage.lh.nii.gz -t
fmcpr.up.sm6.fsaverage.lh.nii.gz
This will bring up an overlay as well as a time course
On 08/11/2014 05:05 PM, Jiahe Zhang wrote:
> Both
Both looked fine. First value in .mincost is .45
On Mon, Aug 11, 2014 at 4:51 PM, Douglas N Greve
wrote:
> two ways:
>
> 1. run tkregister-sess to visually check them
> 2. look at the first value in register.dof6.dat.mincost. It is hard to
> say what a threshold is for a good registration, but
two ways:
1. run tkregister-sess to visually check them
2. look at the first value in register.dof6.dat.mincost. It is hard to
say what a threshold is for a good registration, but typical values are
about .5 or so. If it is .8 or more, it is probably a problem (and
probably the initialization i
I'm not sure how to check the registration. I see a register.dof6.dat being
produced for each run but I can't tell if it contains the correct
parameters. How can I tell if the registration was accurate?
On Mon, Aug 11, 2014 at 4:01 PM, Douglas N Greve
wrote:
> did you check the registration?
>
did you check the registration?
On 08/11/2014 03:53 PM, Jiahe Zhang wrote:
> There's nothing even when the threshold is set to 0.
>
> I just checked the preprocessing outputs. I can see normal-looking
> volumes
> fmcpr.nii.gz
> fmcpr.up.nii.gz
> fmcpr.up.sm6.mni305.2mm.nii.gz
>
> But there is not
There's nothing even when the threshold is set to 0.
I just checked the preprocessing outputs. I can see normal-looking volumes
fmcpr.nii.gz
fmcpr.up.nii.gz
fmcpr.up.sm6.mni305.2mm.nii.gz
But there is nothing when I open fmcpr.up.sm6.fsaverage.lh.nii.gz or
fmcpr.up.sm6.fsaverage.rh.nii.gz
On
When you say that they are "empty" what do you mean? That there is
nothing above threshold or that there is nothing regardless of how low
the threshold is? See if you get something if you set the thresh close
to 0. Does this happen in the volume and the surface? Have you checked
the registrati
Hi Lingqiang, I did not have a strong reason not to do the low pass
filtering. I think I tried it once and it did not seem to make much
difference, so I did not implement it in FSFAST. The differences may be
due to filtering, but there are a lot of other places where things can
be different too
Thanks Doug. The labels work perfectly now.
I also have a methodological question about the fsfast connectivity
analysis: functional connectivity analyses in the literature (Biswal, etc)
only looks at low-frequency fluctuations from BOLD signals. After
preprocessing and regressing out nuisance re
Hi Lingqiang, convert your .label file to a volume using mri_label2vol
for each subject, something like
cd $SUBJECTS_DIR/$subject/mri
mri_label2vol --label lh.yourlabel.label --temp orig.mgz --regheader
orig.mgz --subject $subject --hemi lh --proj frac 0 1 .1 --o yourlabel.mgz
When you run fcse
Hi Maya, the problem is that the wm.dat file, by default, only has one
column (just the mean of the WM). It is possible to have the top N
principle components in the wm.dat file, but this needs to be set up
when you configure the WM fcseed (it does not really help much). The
same goes for VCSF.
14 matches
Mail list logo
