Dear Sein,
the error messages appear a bit bewildering due to buffering of messages that
are then all displayed at the same time when a "fatal error" occurrs.
> Dear SPM community,
> I am trying to use the unwarp toolbox...
>
> -'creating temporary smooth
> -->completed
> -'computing effects on reference scans iteration 1 -->completed
> ... --
> >completed
>
> -Figure: estimation of EPI
> -->completed
>
> then..... following error message:
>
Fine.
>
> Initialising SPM.........done
> SPM present working directory:
> C:\Programme\Matlab12\work
> ~ - SPM99: Unwarp toolbox -----------------------------------------
> Toolbox directory prepended to MatLab path:
> c:\programme\spm99\toolbox\Unwarp
> ----------------------------------------- 21:53:15 - 09/05/2001
>
> SPM99: unwarp (v%A%) 21:53:18 - 09/05/2001
> ========================================================================
These are all informal messages generated earlier.
>
> Warning: Matrix is close to singular or badly scaled.
> Results may be inaccurate. RCOND = 7.142941e-020.
> > In c:\programme\spm99\toolbox\unwarp\uw_FindFields.m at line 341
> In c:\programme\spm99\toolbox\unwarp\Unwarp.m at line 367
> In c:\programme\spm99\spm.m at line 1092
> uw_FindFields.m terminated abnormally:
>
> Error using ==> >
> Function '>' not defined for variables of class 'ss'.???
I can't figure out exactly what is happening here, but I the first step is
clear. The "design-matrix" used in the linearisation of one iteration of the
problem is very close to singular. This could happen if
a. There is no movement at all, which could happen if you pick the resliced r*
images, instead of the original images with assosciated transformation
matrices.
b. There is a very close correlation between two or more of the movement
parameters. If that is so, you should try to reduce the number of parameters
you model (e.g. go for pitch and roll only).
c. There is truly very little movement in your data, in which case there is no
need for "Unwarp" anyway.
The singularity of the matrix only causes a warning. I can't quite figure out
error that causes FindFields to terminate. It is very likely to be associated
with the first warning though, so try one of the alterantives above.
> Reference to
> non-existent field 'beta'.
> Error in ==> c:\programme\spm99\toolbox\unwarp\uw_undeform_images.m
> On line 90 ==> def_array = zeros(prod(P(1).dim(1:3)),size(ds.beta,2));
> Error in ==> c:\programme\spm99\toolbox\unwarp\Unwarp.m
> On line 379 ==> uw_undeform_images(spm_vol(wifiles),flags,ds);
> Error in ==> c:\programme\spm99\spm.m
> On line 1092 ==> evalin('base',varargin{2}(i).prog)
> ??? Error while evaluating uicontrol Callback.
After uw_FindFields.m is finished calculating the fields, uw_undeform_images.m
starts to reslice the images according to the .mat files from "realign", and
the *_uw.mat file from "Unwarp". Because uw_FindFields.m failed no
deformation-fields have been stored, so the reslicing fails. Hence, this is
just a knock-on from the error above.
If none of a-c above applies perhaps you could get back to me with a more
detailed description of your input to "Unwarp", and perhaps mail me the
resulting *_uw.mat file. Just remember no to send that to the mailbase since we
don't want to litter everyones discs with large attachements.
Good luck Jesper
|