Dear Jie
The error suggests some problem in orthogonalizing the regressors. I don't
know if this is a problem with the batch code or your regressors. By manual
do you mean that if you entered each condition separately and the parametric
regressors separately you did not see this error? If that is the case then
perhaps the batch multiple parameter code has a problem. In that case could
you please send me your batch script and the mat files you are using for
multiple conditions and parameters. I will have a look.
darren
> -----Original Message-----
> From: Jie Zhuang [mailto:[log in to unmask]]
> Sent: Wednesday, November 28, 2007 5:51 AM
> To: d gitelman
> Subject: Re: batch of multiple modulators per condition
>
> Darren,
>
> Thank you very much indeed for your great help! I re-started
> the updated version of SPM5 several times, and it finally
> works. When I do the fixed effect analysis for a single
> subject with this batch method, I found some error messages
> (see below for example) which did not appear in the same
> analysis by hand. The two results are almost identical except
> some tiny difference in the activated voxel size. What do the
> error messages mean? Do they matter to my analysis? Many thanks!
>
> Cheers,
> Jie
>
> ===================================================
> Warning: Matrix is close to singular or badly scaled.
> Results may be inaccurate. RCOND = 5.398838e-47.
> > In spm_orth at 21
> In spm_get_ons at 229
> In spm_fMRI_design at 271
> In spm_fmri_spm_ui at 213
> In spm_config_fmri_spec>run_stats at 971
> In spm_jobman>run_struct1 at 1585
> In spm_jobman>run_struct1 at 1595
> In spm_jobman>run_struct1 at 1595
> In spm_jobman>run_struct at 1514
> Warning: Matrix is close to singular or badly scaled.
> Results may be inaccurate. RCOND = 3.819570e-51.
> > In spm_orth at 21
> In spm_get_ons at 229
> In spm_fMRI_design at 271
> In spm_fmri_spm_ui at 213
> In spm_config_fmri_spec>run_stats at 971
> In spm_jobman>run_struct1 at 1585
> In spm_jobman>run_struct1 at 1595
> In spm_jobman>run_struct1 at 1595
> In spm_jobman>run_struct at 1514
> ===================================================
>
>
> d gitelman wrote:
>
> >Jie:
> >
> >if you look at the batch script for fmri analysis click on multiple
> >conditions and scroll down the help text. Here is the summary
> >
> >1) If you enter each condition by hand then the time and parametric
> >modulators are entered for each condition under the
> Condition subtree.
> >(fmri model spec-> data & design -> subject/session ->
> Conditions -> condition).
> >
> >2) If you use the multiple conditions facility then the mat
> file will
> >always contain 3 variables (cell arrays) to specify the conditions
> >(names, onsets and durations). So as an example let's say you have 2
> >conditions, 5 events per condition (onsets in seconds)
> >
> >names{1} = 'rest';
> >names{2} = 'visual stimulus';
> >onsets{1} = [0 12 38 49 55];
> >onsets{2} = [9 18 25 34 42];
> >durations{1} = [0 0 0 0 0]; (or you can have just a single 0)
> >durations{2} = [0 0 0 0 0];
> >
> >Let's say you want to model a quadratic time effect on the second
> >condition then include the cell array tmod tmod{1} = [];
> (could also
> >put in a 0 meaning no time modulation) tmod{2} = 2;
> >
> >If you wanted a parametric effect include a variable called pmod.
> >This is a structure array of cell arrays.
> >Let's say you want a linear (first order) parametric effect on the
> >second condition pmod(2).name{1} = 'visstim-param';
> pmod(2).param{1} =
> >[1 1 2 3 5]; (this must be the same length as #onsets in that
> >condition) pmod(2).poly{1} = 1; (i.e., first order).
> >
> >If you wanted 2 parametric modulators on the second condition you
> >include the second one as follows pmod(2).name{2} =
> >'another-modulator'; pmod(2).param{2} = [48 52 8 -43 -55];
> >pmod(2).poly{2} = 1;
> >
> >and so on. Note that for the time modulator it is the cell
> index that
> >relates to the condition, while for the parametric modulator
> it is the
> >structure index that relates to the condition while the cell index
> >relates to the number of parametric modulators for that condition.
> >
> >hth,
> >darren
> >
> >
> >
> >
> >
> >>-----Original Message-----
> >>From: Jie Zhuang [mailto:[log in to unmask]]
> >>Sent: Friday, November 23, 2007 4:21 AM
> >>To: [log in to unmask]
> >>Subject: batch of multiple modulators per condition
> >>
> >>Dear Darren,
> >>
> >>According to Jan's email to the SPM mailing list, the
> problem of the
> >>batch of multiple modulators per condition has been solved by you.
> >>However, I still could not find the help text or other related
> >>information. Could you be kind to give me some tips?
> >>Thank you very much!
> >>
> >>Cheers,
> >>Jie
> >>
> >>
> >>
> >
> >
> >
> >
>
|