Hi,
> >During analyses of 15 patient subject's data I'm encountering problem in
> the
> >estimation process of one of the subjects. I've checked previous emails
> >regarding this subject and have downloaded updates but still no luck.
> Model
> >estimation produces the following error:
> >
> >
> >Temporal non-sphericity (over voxels) : ...REML estimation
> >
> >Warning: Divide by zero.
> >
> >(Type "warning off MATLAB:divideByZero" to suppress this warning.)
> >
> >> In c:\spm2\spm_spm.m at line 772
> >
> > ReML Block - 1
> >
> >??? Index exceeds matrix dimensions.
This is a recurring problem, so I thought I would send on my progress
on it. I've appended an earlier email I sent to the spm-authors
list. In summary the reply was, that the non-sphericity was
inestimable and the non-sphericity structure xX.V is rank deficient.
I guess you could try and solve this by reducing the covariance model
(not modelling repeated measures for example), or providing more
observations.
Best,
Matthew
>I don't know if y'all noticed a post recently by Boris Suchan, who was
>getting errors for a PET design in spm_spm, like this:
>
>SPM2: spm_spm (v2.66) 16:42:19 - 10/02/2004
>========================================================================
>Initialising parameters :
>...computing???
>Index exceeds matrix dimensions.
>
>Error in ==> D:\MATLAB6p5\toolbox\matlab\sparfun\spdiags.mOn line 102
==> a = [a; i i+d(k) B(i+(m>=n)*d(k),k)];
>
>Error in ==> d:\spm2\spm_spm.mOn line 419 ==> s =
>spdiags(1./sqrt(diag(s)),0,nScan,nScan);
>
>Error in ==> d:\spm2\spm_spm.mOn line 826 ==> SPM =
spm_spm(SPM);
>
>I had had the same thing, so I asked him to send me his xVi
structure(attached). Sure enough, when you run
>
>[u s] = spm_svd(xVi.V);
>
>on his input (which is 36x36), you get a u of 36x16 and an s of 16x16.
>This is because the svd values have been thresholded out in spm_svd.
>This obviously leads to a crash when spm_spm gets to:
>
>s = spdiags(1./sqrt(diag(s)),0,nScan,nScan);
>
>because diag(s) returns only 16 values.
>
>Any thoughts greatly appreciated...
|