Hi Doug and FreeSurferers,
I came across this email chain and am having a similar problem. I am working on
an interaction between a categorical and continuous variable, covarying for a
categorical variable. (E.g. the interaction between group membership and age,
covarying for gender). I know th
That was the problem!
Thanks!
- Jerry
On Thu, Apr 10, 2008 at 5:55 PM, Doug Greve <[EMAIL PROTECTED]>
wrote:
> I think the problem is that the last line is a blank line. Remove that
> and retry.
>
> doug
>
> Jerry Yeou-Wei Chen wrote:
>
> I am getting the following error when I run mri_glmfit wi
I think the problem is that the last line is a blank line. Remove that
and retry.
doug
Jerry Yeou-Wei Chen wrote:
I am getting the following error when I run mri_glmfit with my custom
design matrix:
WARNING: matlab elment type is 940126218, which is not a matrix.
could not allocate 82268570
I am getting the following error when I run mri_glmfit with my custom design
matrix:
WARNING: matlab elment type is 940126218, which is not a matrix.
could not allocate 822685706 x 805908489 matrix
Cannot allocate memory
unsupported matlab format -849662976 (unknown)
Cannot allocate memory
Belo
Thanks for the clarification, Doug.
The Wiki page for mri_glmfit says to "Specify the design matrix in matlab4
format."
- Jerry
On Wed, Apr 9, 2008 at 5:19 PM, Doug Greve <[EMAIL PROTECTED]>
wrote:
>
> >From mri_glmflit --help :
>
> --X design matrix file
>
> Explicitly specify the design matrix
From mri_glmflit --help :
--X design matrix file
Explicitly specify the design matrix. Can be in simple text or in matlab4
format. If matlab4, you can save a matrix with save('X.mat','X','-v4');
It does not need/expect any extension.
doug
Jerry Yeou-Wei Chen wrote:
Regarding custom design
Regarding custom design matrices, does the --X flag require a design matrix
in matlab format? If so, can I simply create it in a text editor and name it
with the suffix ".mat"?
Also, if I use the --X to specify a design matrix, am I still supposed to
specify the fsgd file?
Thanks,
- Jerry
On Wed,
You can pass custom design and contrast matrices to mri_glmfit, so you
can pass the same matrices to mri_glmfit that you would use with spm.
Alternatively, you can set up an fsgd file with a 3rd continuous
variable that is the produce of the first two.
doug
Jerry Yeou-Wei Chen wrote:
H
Hello,
I would like to know how to examine an interaction between continuous
variables.
In SPM, my understanding is that a new interaction regressor, comprised of
the product of the two continuous variables of interest, can be coded and
entered into the model.
Would this be appropriate for CTA grou
