Hi all,
I think this is an spm bug -- has anyone got a paired t-test to work?
I get the same error you describe on several different data sets.
(spm5 rev 411, Matlab 7.0.1 r14)
To get a simple paired t-test, you could just do a one-sample t-test
on difference images (that you'll need to create yourself).
I used avwmaths from FSL to create these, and spm5 was then able to
perform a single-sample test with results that looked right.
e.g. avwmaths A1.nii -sub B1.nii diff1.nii will give diff1 = A1 - B1
(http://www.fmrib.ox.ac.uk/fsl/ though you could probably do this
within Matlab/SPM)
Best,
Ged.
On 16/02/06, Woodard, John <[log in to unmask]> wrote:
>
> I'm trying to analyze a series of PET scans using the paired t-test analysis
> option in SPM 5. There are 2 scans per patient, pre-drug and post-drug. The
> design matrix appears to be appropriate, but when I try to estimate, I get
> the following:
>
>
>
>
>
> SPM5: spm_spm (v$Rev: 301 $) 16:37:45 - 16/02/2006
>
> ========================================================================
>
>
>
> SPM5: spm_spm (v$Rev: 301 $) 16:37:48 - 16/02/2006
>
> ========================================================================
>
> Initialising parameters : ...initialised
>
>
>
> Plane 1/69 , block 1/1 : ...done
>
> Plane 2/69 , block 1/1 : ...done
>
>
>
> … …
>
>
>
> Plane 68/69 , block 1/1 : ...done
>
> Plane 69/69 , block 1/1 : ...done
>
> Temporal non-sphericity (over voxels) : ...REML estimation
>
> ??? Error using ==> plus
>
> Matrix dimensions must agree.
>
>
>
> Error in ==> spm_reml at 84
>
> C = C + Q{i}*h(i);
>
>
>
> Error in ==> spm_spm at 860
>
> [V,h] = spm_reml(Cy,xX.X,xVi.Vi);
>
>
>
> Error in ==> spm_spm at 299
>
> spm_spm(SPM);
>
>
>
> ??? Error while evaluating uicontrol Callback.
>
>
>
>
>
>
>
> Any ideas will be greatly appreciated!
>
>
>
> Best regards,
>
>
>
> John
>
>
>
> John L. Woodard, Ph.D.
>
> Associate Professor of Psychology
>
> Rosalind Franklin University of Medicine and Science
>
> 3333 Green Bay Rd.
>
> North Chicago, IL 60064
>
> Voice: 847-578-3311
>
> Fax: 847-578-8765
>
> e-mail: [log in to unmask]
>
>
|