Dear Carolyn,
if you are really sure about your contrast, in order to test if the
problem is related to the con-manager. you can by-passed it and use a
direct command line instruction with the following
load SPM
xC = spm_FcUtil('Set','mycontrast','T','c', C, SPM.xX.xKXs); % C is
your contrast weights;
if there is no error at this stage you can estimate the contrast by doing
SPM.xCon(end+1)=xC;
spm_contrasts(SPM);
otherwise, there is real problem with your contrast.
Good luck
F.
On Feb 6, 2008 7:28 PM, Carolyn L. Fort <[log in to unmask]> wrote:
> Dear SPM-ers,
>
> I'm receiving the "invalid contrast weights" (or wording close to that) error
> message within the contrast manager window when inputting contrasts
> weights for
> a new t-test contrast. I'm not new to SPM and have used this function
> inumerable
> times, and I'm quite sure that my weights are valid and logical.
>
> Here is what is different this time from others, when creating a new contrast:
> our design in quite complex, requiring 324 values to be entered in this
> window.
> Is there a maximum # of values that can be entered here? I did not see mention
> of this in the con_man manual.
>
> Even entering test values, e.g., 1 -1 0 0 0, and allowing it to fill in the
> remaining 0's returns the same error. Interestingly though, if I choose an
> F-test for the effects of interest and run a similar test (1 -1 0 0 0) it runs
> just fine.
>
> Ideas?
>
> Many thanks,
>
> Carolyn
>
>
> Carolyn L. Fort
> Research Specialist, Senior
> Department of Psychiatry
> University of Arizona
> PO Box 245002
> Tucson, AZ 85724-5002
> v: 520-626-8568
> f: 520-626-6050
>
|