Dear SPM
list,
I'm running into a major estimation problem with a fixed
effects model, 85 scans each for 11 subjects, block paradigm, with motion
parameters as covariates for each subject, using Matlab 7.0.1 on a Windows
XP machine. (Actually, two different Windows XP machines and two different
models, so the error's reproducible.) The fix that someone recommended to
the list:
feature('jitallow','structs','off');
as in
doesn't help either.
The error message is the same as the one mentioned in that e-mail,
and this one:
>SPM2: spm_spm
(v2.66)
12:09:43 -
07/04/2005
>=======================================================================
>Initialising
parameters
:
>...computing??? Index exceeds matrix dimensions.
>
>Error in ==>
spdiags at 114
>
a((len(k)+1):len(k+1),:) = [i i+d(k) B(i+(m>=n)*d(k),k)];
...
Yes, this is with all spm2 updates; I just
downloaded and installed those but got the same error.
Just to make things more frustrating, Matlab 6.5 will estimate
the same model, but then after four contrasts the SPM.mat gets corrupted and can
no longer be read by spm, marsbar, or even Matlab itself. (The error here
wasn't very informative: "Can't read file.") Unfortunately it's taking
about twelve hours to estimate this model, so it's impractical to re-estimate it
each time for just four contrasts.
Hope somebody has found a fix or has a suggestion!
Alex Dresner
**********************
Alex Dresner, Ph.D.
fMRI
Physicist
Imaging Sciences Department
Hammersmith Hospital
Imperial
College London
020 8383 4516 (w)