Hi Todd,
Can you get the values for sf_X1o, Fc, and sX?
The reason for the error may be that there is always a constant term in
the design matrix already (the intercept for the mean). If you add
another one, the regressors are linearly dependent. You can check this
on the graphics page that is shown after you have created your design.
Under the columns in the design matrix is a coloured bar, which is white
for independent regressors, and grey for dependent regressors. If your
bar shows up dark grey (or black?) then the model cannot be estimated.
all the best
Alle Meije
Todd Penney wrote:
> Hi All,
>
> I tried running the SPM steps with just a constant in the GLM in
> order to retrieve the resulting ResMS.img files for a calculation I
> need to perform. Unfortunately SPM5 is returning an error during the
> estimation:
>
> SPM5: spm_spm (v$Rev: 301 $) 11:08:37 - 24/08/2006
> ========================================================================
> Initialising parameters :
> ...computing??? Error using ==> mtimes
> Inner matrix dimensions must agree.
>
> ??? Error while evaluating uicontrol Callback.
>
>
> I stepped through the code to find the location of the error myself
> and it is occurring in the spm_FcUtil routine at line 301 which reads:
>
> varargout = {sf_X1o(Fc,sX)};
>
> I do not know how to correct it. Any suggestions? Thank you for
> your help.
>
> Todd Penney
>
|