fmn is the mean image averaging over all time points, so it should have
the same dimensions as one of your images. The repmat() call is supposed
to replicate the image over the number of time points, but there is a
bug. It should be f.vol,4 not 3. Try that
On 04/03/2018 02:33 PM, Sarah Cole wr
Doug,
Could you please let me know what this line is doing?
"fmnrep = repmat(fmn,[1 1 1 size(f.vol,3)]);"
When I try to assign f.vol/fmnrep to s.vol (this line s.vol =
f.vol./fmnrep;), matrices do not match in dimension.
Thanks,
S
On Thu, Feb 15, 2018 at 1:05 PM, Douglas N Greve
wrote:
> That
That does not get computed. You can do it in matlab with something like
f = MRIread('f.nii.gz');
fmn = mean(f.vol,4);
fmnrep = repmat(fmn,[1 1 1 size(f.vol,3)]);;
s = f;
s.vol = f.vol./fmnrep;
MRIwrite(s,'percentchange.nii.gz')
On 02/15/2018 02:03 PM, Sarah Cole wrote:
> Yes, that's exactl
Yes, that's exactly what I am looking for. Raw waveform as a percent.
On Thu, Feb 15, 2018 at 1:00 PM, Douglas N Greve
wrote:
> h-offset is the intensity offset (basically the mean signal). h.nii.gz
> is a complicated combination of the betas (regression coefficients) that
> is included for back
h-offset is the intensity offset (basically the mean signal). h.nii.gz
is a complicated combination of the betas (regression coefficients) that
is included for backwards compatibility. What do you mean by % signal
change? Do you want the raw waveform as a percent?
On 02/15/2018 01:52 PM, Sarah
Thank you, Doug.
Sorry for not being clear. Could you please tell me what h.nii.gz and
h-offset.nii.gz are?
Also, do you have any recommendation on how to get % signal change on the
signal intensity as oppose to % signal change on contrasts?
Thank you
On Thu, Feb 15, 2018 at 12:47 PM, Douglas N
It divides the contast (ces) by voxel-wise mean. I'm not sure I
understand your quesstion about h and h-offset
On 02/15/2018 01:04 PM, Sarah Cole wrote:
> Hi,
>
> Could someone please let me know how to get the percent signal change
> for raw signal intensity?
>
> I think the cespct.nii.gz give
Hi,
Could someone please let me know how to get the percent signal change for
raw signal intensity?
I think the cespct.nii.gz gives me the contrast as %. However, I am not
sure about dividing h.nii and h-offset.nii. What are they?
Thank you
___
Freesu
I think you'll need to contact the FSL list to ask how to convert the
COPE volume to a voxel-wise percent signal change. Once you have that,
then you can just project that data onto the surface instead of the
original COPE.
doug
On 07/24/2015 05:11 PM, Mariam Sood wrote:
> Dear Doug,
>
> Thank
Dear Doug,
Thank you for the reply. I want to use the percent change in signal as a
measure of effect size for various regions of interest, so would ideally like
an average percent signal change in that ROI. I want it at the group data
though. Much appreciate your suggestions.
Many thanks,
Ma
what kind of percent output do you want? FSL does a global
normalization, effectively computing percent change (actually 10th of
percent change I think). If you want voxel-wise percent, then it is more
difficult.
On 7/24/15 11:11 AM, Mariam Sood wrote:
> Dear Doug,
>
> I am using the FSL-Freesu
Dear Doug,
I am using the FSL-Freesurfer pipeline, with single subject analysis done in
Feat and group analysis done using mri_glmfit. Is there a way I can calculate
percent signal change from the group analysis output (or any other method which
combines single subject output)?
Many thanks,
M
percent relative to what? FEAT will automatically rescale to 1 (or
something like that)
doug
On 03/25/2014 05:51 PM, Manuel Spitschan wrote:
> Dear FreeSurfer experts,
>
> I've performed a few first-level analyses of BOLD data with FSL, and have
> followed the instructions on the FS Wiki
Dear FreeSurfer experts,
I've performed a few first-level analyses of BOLD data with FSL, and have
followed the instructions on the FS Wiki
(http://surfer.nmr.mgh.harvard.edu/fswiki/FsTutorial/FslGroupFeat) to sample
the COPEs of each run to a common surface (fsaverage) and run a fixed-effects
No, sorry, the raw timecourse would have to be analyzed.
Elif SIKOGLU wrote:
Hello.
I obtained the percent signal change using roisummary-sess. (divided
the BOLD amplitude by the mean functional baseline offset to get the
percent signal change.) Is there a way to get an error estimate (i.e.
Hello.
I obtained the percent signal change using roisummary-sess. (divided the
BOLD amplitude by the mean functional baseline offset to get the percent
signal change.) Is there a way to get an error estimate (i.e. standard
deivation) on this value?
Thanks
Elif
Yes, it is. The h.nii file has a somewhat tortured format. If you just
want the contrast as a percent, then look in the contast directory for
cespct.
Elif SIKOGLU wrote:
Hello,
We want to get the percent signal change for the activated regions and
for that we want to use the h.nii and h-of
Hello,
We want to get the percent signal change for the activated regions and for
that we want to use the h.nii and h-offset.nii files.
We plan to divide the h.nii by h-offset.nii .
First of all is this a correct approach?
Second, h.nii has 4 frames in it, which one do we really need?
Than
Hi,
We have calculated a FIR group analysis with FS-FAST and have the
corresponding ces, cesvar, f, fsig, iminsig, minsig, sig, and t maps in
the contrast folders.
Which one of these should I use for presenting % signal change? We want to
do some further analyses of the group analysis results wit
Thanks for the suggestions. I'm trying to read in the h-offset volumes,
but I can't figure out the correct matlab script to use. To read in the
h volumes I'm using fast_ldsxabfile, but that assumes there is a file
.dat and there's no h-rh-offset.dat. I could probably hack
something together, bu
What I do in "froi" is take the value at the particular voxel and
divide by the h-offset value for that voxel and multiply by 100. For
ROIs, that's done individually for each voxel, and then averaged to get
the value for the ROI.
To get the standard deviation is similar, except of course you c
Greetings,
I'm trying to write a script to generate hemodynamic timecourses for
particular voxels. I want to imitate the feature in tkmedit where these
graphs can be viewed as Percent Signal Change.
I know how to generate this for averages across an ROI, but that
involves computing the Average
22 matches
Mail list logo
