Hi again,
Surprisingly, when we rotate the DXM, i.e. instead of using two-sample
t-test, we use one-sample t-test with group as covariate spm_spm do
manage estimate the model. I guess this may suggest that something in
the choice of the threshold is wrong in the below case.
So we are happy.
Cheers
Pia
-----Original Message-----
From: Pia Rotshtein
Sent: 04 July 2007 11:31
To: 'SPM (Statistical Parametric Mapping)'
Subject: RE: [SPM] Vectors must be the same lengths.
Hi to the list and the experts,
To recap on Judith e-mail from May 2007, on the error massage following
second level model estimation:
>> Error using ==> plot
>> Vectors must be the same lengths.
So basically we got the same error massage when trying to run
two-sample t-test with two groups.
Exploring the source of the error it seems that spm (spm_spm) did not
find any above threshold voxels for any of the beta, and hence ending up
producing this error massage.
In spm_spm the error come form lines 849-857 because the variable 's' is
still zero and is not updated in lines 742-749, where the threshold is
been checked (I assume).
Basically it would have been reasonable, but when we run each group
separately we do manage to get some blobs, and some of these blobs do
overlaps, which suggest that there are some above threshold voxels.
Any help comments, advices, suggestions will be highly appreciated.
Thanks
Pia
|