Dear List,
I might have been a bit hasty when I confirmed yesterday that there
weren't any problems with SPM2 and Matlab R14SP1 student version under
Linux.
I encountered an error and a warning under R14SP1 that did not occur
under R13 and these make me think that sprintf.m and svd.m have been
changed in the current release.
The warning concerns sprintf. It appears that sprintf under R14SP1 no
longer accepts constructs like following (as repeatedly used in
spm_spm.m):, e.g. in line 474)
fprintf('%s%30s\n',sprintf('\b')*ones(1,30),'...done')
for printing status messages repeatedly in one line, because sprintf
compains that the input must be a char. (This can be solved by a
character conversion, e.g.
fprintf('%s%03s\n',char(sprintf('\b')*ones(1,30)),'...done')
The error concerns the singular value decomposition (svd.m)
When I first start R14SP1 and estimate a model under SPM2 (within
subject ANOVA with 4 subject and 6 conditions), I receive the following
error message under R14SP1:
------------------------------------------------------------------------
??? Error using ==> svd
NaN or Inf prevents convergence.
Error in ==> spm_sp>sf_set at 1122
[x.u, s, x.v] = svd(full(X),0);
Error in ==> spm_sp at 221
varargout = {sf_set(X)};
Error in ==> spm_SpUtil at 437
varargout = { spm_sp('r:',spm_sp('set',c0),c1) };
Error in ==> spm_FcUtil at 263
Fc.c = spm_SpUtil('i0->c',sX,iX0);
Error in ==> spm_spm at 857
xCon = spm_FcUtil('Set',Fcname,'F','iX0',iX0,xX.xKXs);
??? Error while evaluating uicontrol Callback.
-----------------------------------------------------------------------
When I look into the directory I see that the ResI-images are still
there, while the beta-images have already been created. When I now
re-estimate the model, it runs through without an error and the
ResI-images are also removed. The error message is reproducable when
I remove the ResI-images manually before re-estimation.
Strangely enough, the same analysis simply runs through under Matlab R13
(student version).
Has anybody encountered similar errors?
Any comments are greatyl appreciated.
Thanks,
Jan
--
Jan Gläscher Neuroimage Nord
+49-40-42803-7890 (office) Dept. of Neurology, Bldg S10
+49-40-42803-9955 (fax) University Hospital Hamburg-Eppendorf
[log in to unmask] Martinistr. 52
20246 Hamburg
Germany
http://www.uke.uni-hamburg.de/zentren/neuro/neurologie/mitarbeiter/glaescher_jan.html
--------------------------------------------------------------------------------------
GnuPG/PGP key id: FEC4B55C
fingerprint: 5A36 1EF6 8472 117E 805A F240 3146 A410 FEC4 B55C
--------------------------------------------------------------------------------------
|