Hi Alex, we don't have an elegant solution to this, but you can do it in matlab easily enough. If your data D are in a N-by-327684 matrix (I assume that this is left and right hemis), then
a.vol = permute(D(:,1:163842),[2 3 4 1]); MRIwrite(a,'lh.yourdata.mgh');
b.vol = permute(D(:,163843:end),[2 3 4 1]); MRIwrite(b,'rh.yourdata.mgh');
doug
Alexander Lebedev wrote:
Excuse me for disturbing Douglas,
The matter is that I have the coefficients for each vertex and I would like to find the easiest way to display them into the brain space, showing their importance for the classification. My original ascii file contains N-by-327684 matrix (N subjects, 327 684 vertices - 163 842 for each hemisphere). But I can easily change it if necessary... I'm just wondering if there is a way to do this... I'm asking because it seems like "y.mgh -> y.asc" (mri_convert y.mgh y.ascii --ascii) conversion is available only one-way.
Regards, Alex
- Alexander Lebedev awlman@rambler.ru [Fri, 27 Jan 2012 10:07:11
+0400]:
Yes, sorry...
Alex
Hi Alex, I'm not sure I understand what you mean. Do you mean that
you
have a coefficient for each vertex and you want to display your
results
on a surface and need to get them into a format that FS can read? doug
Alexander Lebedev wrote:
Yes, thanks a lot Douglas! That's what I was looking for. And it's much easier than my approach... I needed to extract the vertices
in
order to perform a multivariate analysis in R. May I also ask
another
question then? I have performed pls-regression and would like to
map
the coefficients into the brain space... Is there such an
opportunity
(at least theoretically speaking) ?
Regards, Alex
Hello Doug! Thank you so much! Sorry, I've just noticed a mistake... My matrix is 327684x2: the 1st column denotes vertex numbers and the 2nd one contains corresponding regression coefficients... I will try to display this tomorrow... Probably, I will have some troubles with this, considering that there are both positive and negative coefficients... But I will try. --- Regards, Alex
* Douglas N Greve greve@nmr.mgh.harvard.edu [Mon, 30 Jan 2012 10:27:36 -0500]:
Hi Alex, we don't have an elegant solution to this, but you can do it in matlab easily enough. If your data D are in a N-by-327684 matrix (I assume
that
this is left and right hemis), then
a.vol = permute(D(:,1:163842),[2 3 4 1]); MRIwrite(a,'lh.yourdata.mgh');
b.vol = permute(D(:,163843:end),[2 3 4 1]); MRIwrite(b,'rh.yourdata.mgh');
doug
Alexander Lebedev wrote:
Excuse me for disturbing Douglas,
The matter is that I have the coefficients for each vertex and I
would
like to find the easiest way to display them into the brain space, showing their importance for the classification. My original ascii file contains N-by-327684 matrix (N subjects, 327 684 vertices - 163 842 for each hemisphere). But I can easily change it if necessary... I'm just wondering if there is a way to do this... I'm asking
because
it seems like "y.mgh -> y.asc" (mri_convert y.mgh y.ascii --ascii) conversion is available only one-way.
Regards, Alex
- Alexander Lebedev awlman@rambler.ru [Fri, 27 Jan 2012 10:07:11
+0400]:
Yes, sorry...
Alex
Hi Alex, I'm not sure I understand what you mean. Do you mean
that
you
have a coefficient for each vertex and you want to display your
results
on a surface and need to get them into a format that FS can read? doug
Alexander Lebedev wrote:
Yes, thanks a lot Douglas! That's what I was looking for. And
it's
much easier than my approach... I needed to extract the
vertices
in
order to perform a multivariate analysis in R. May I also ask
another
question then? I have performed pls-regression and would like
to
map
the coefficients into the brain space... Is there such an
opportunity
(at least theoretically speaking) ?
Regards, Alex
Dear Doug,
I just want to let you know that your solution works perfectly. All the vertices were successfully mapped into the brainspace via MRIwrite and visualized as an "overlay" after selection of the appropriate colormap and thresholds. Thank you very much for your help! --- Regards, Alex
* Alexander Lebedev awlman@rambler.ru [Tue, 31 Jan 2012 01:50:03 +0400]:
Hello Doug! Thank you so much! Sorry, I've just noticed a mistake... My matrix is 327684x2: the 1st column denotes vertex numbers and the 2nd one contains corresponding regression coefficients... I will try to display this tomorrow... Probably, I will have some troubles with this, considering that there are both positive and negative coefficients... But I will try.
Regards, Alex
- Douglas N Greve greve@nmr.mgh.harvard.edu [Mon, 30 Jan 2012
10:27:36
-0500]:
Hi Alex, we don't have an elegant solution to this, but you can do it in
matlab
easily enough. If your data D are in a N-by-327684 matrix (I assume
that
this is left and right hemis), then
a.vol = permute(D(:,1:163842),[2 3 4 1]); MRIwrite(a,'lh.yourdata.mgh');
b.vol = permute(D(:,163843:end),[2 3 4 1]); MRIwrite(b,'rh.yourdata.mgh');
doug
Alexander Lebedev wrote:
Excuse me for disturbing Douglas,
The matter is that I have the coefficients for each vertex and I
would
like to find the easiest way to display them into the brain space, showing their importance for the classification. My original
ascii
file contains N-by-327684 matrix (N subjects, 327 684 vertices -
163
842 for each hemisphere). But I can easily change it if
necessary...
I'm just wondering if there is a way to do this... I'm asking
because
it seems like "y.mgh -> y.asc" (mri_convert y.mgh y.ascii --ascii) conversion is available only one-way.
Regards, Alex
- Alexander Lebedev awlman@rambler.ru [Fri, 27 Jan 2012 10:07:11
+0400]:
Yes, sorry...
Alex
Hi Alex, I'm not sure I understand what you mean. Do you mean
that
you
have a coefficient for each vertex and you want to display your
results
on a surface and need to get them into a format that FS can
read?
doug
Alexander Lebedev wrote:
Yes, thanks a lot Douglas! That's what I was looking for. And
it's
much easier than my approach... I needed to extract the
vertices
in
order to perform a multivariate analysis in R. May I also ask
another
question then? I have performed pls-regression and would like
to
map
the coefficients into the brain space... Is there such an
opportunity
(at least theoretically speaking) ?
Regards, Alex
freesurfer@nmr.mgh.harvard.edu
-
Alexander Lebedev -
Douglas N Greve