Dear Dara,
I think it can be due to :
1- You missed one or more steps in the model design.
2- It can be due to a problem with the conversion to the Analyze_SPM format (you can check it by displaying any image from your data, using SPM to make sure that the image is SPM compatible).
Salem
-----Message d'origine-----
De : SPM (Statistical Parametric Mapping) [mailto:[log in to unmask]] De la part de Dara Cannon
Envoyé : lundi 25 mai 2009 14:58
À : [log in to unmask]
Objet : [SPM] no significant voxels spm_spm.m line 861
Any help with this would be greatly appreciated.
We have set up a flexible factorial design, very simple: factor 1 = subject
(n=62), factor 2 = group (3 groups), ME of factor 1 and main effect of factor
2 specified. Specified all subjects and defined an array for Matrix I (4 x
nscans: column 1: 62x1, column 2: 1 through 62, column 3: 20x1's 21x2's and
21x3's and column 4: 62x1's). There is no apparent problem with the data I
can see, and I've already tried lowering the threshold from 0.001 to 0.05 line
482 of spm_spm.m under 'threshold for voxels entering non-sphericity
estimates' as suggested on the list previously for a similar error reported. This
didn't change the error. The model estimation does not run, no beta files are
generated. Here's the error and associated lines from spm files are given below
the error:
Error running job: Error using ==> spm_spm at 861
Please check your data: There are no significant voxels.
In file "/usr/local/spm5/spm_spm.m" (v946), function "spm_spm" at line 861.
In file "/usr/local/spm5/spm_config_fmri_est.m" (v832), function "run_est" at
line 394.
spm_spm line 861:
%-If not defined, compute non-sphericity V using ReML Hyperparameters
%
==========================================================
================
if ~isfield(xVi,'V')
%-check there are signficant voxels
%----------------------------------------------------------------------
if ~s
spm('FigName','Stats: no sign voxels',Finter); spm('Pointer','Arrow')
figure(Finter);
if isfield(SPM.xGX,'rg')&&~isempty(SPM.xGX.rg)
plot(SPM.xGX.rg)
errordlg({'Please check your data'; ...
'There are no significant voxels';...
'The globals are plotted for diagnosis'});
else
errordlg({'Please check your data'; ...
'There are no significant voxels'});
end;
error('Please check your data: There are no significant voxels.');
return
end
Line 394 of spm_config_fmri_est is the line SPM = spm_spm(SPM) of this
section:
%
==========================================================
=============
% R E M L E S T I M A T I O N
%
==========================================================
=============
if isfield(job.method,'Classical'),
SPM = spm_spm(SPM);
%-Automatically set up contrasts for factorial designs
%-------------------------------------------------------------------
if isfield(SPM,'factor')
if SPM.factor(1).levels > 1
% don't both if you've only got 1 level and 1 factor
cons = spm_design_contrasts(SPM);
I also tried using Baysian 1st level instead of classical and got a similar error
but on different lines.
Running "Model estimation"
Error running job: Reference to non-existent field 'Sess'.
In file "/usr/local/spm5/spm_spm_vb.m" (v911), function "spm_spm_vb" at line
233.
In file "/usr/local/spm5/spm_config_fmri_est.m" (v832), function "run_est" at
line 596.
Any suggestions as to what could be causing this would be greatly
appreciated.
thank you
Dara
|