Dear Shoichi,
have you checked that when specifying the DCM manually, you choose
the VOIs in exactly the same order as they are specified by
spm_dcm_voi? Also, are you using the same SPM.mat file for both
analyses or have you changed the order of inputs (regressors) in the
design matrix somehow?
Best wishes,
Klaas
At 03:10 08/12/2006, you wrote:
>Dear Klaas and DCM experts,
>
>Thank you for your response.
>I am sorry, but I cannot succeed if I modified the description of
>spm_dcm_U. However, I found that the error message disappeared. I used the
>current version of SPM2 (original SPM2 overwritten with the current updates
>files, updates.tar.gz, Apr 11 2005,
>ftp://ftp.fil.ion.ucl.ac.uk/spm/spm2_updates/). I also downloaded
>spm_logdet.m. I confirmed that the procedure (3) and (4) are correct.
>
>My results are:
>[a] manually calculated using GUI
>DCM.B =
> 0 0 0.0011
> 0 0 0.0015
> 0 0 0
>
>[b] batch analyse with "spm_dcm_voi"
>(3) spm_dcm_voi('DCM_model1',
>{'VOI_ROI1.mat', 'VOI_ROI2.mat', 'VOI_ROI3.mat'}) ;
>(4) spm_dcm_U('DCM_model1','SPM',1,[2]);
>(5) spm_dcm_estimate('DCM_model1')
>DCM.B =
> 0 0 0.0026
> 0 0 0.0060
> 0 0 0
>
>[c] batch analyse without "spm_dcm_voi"
>(4) spm_dcm_U('DCM_model1','SPM',1,[2]);
>(5) spm_dcm_estimate('DCM_model1')
>DCM.B =
> 0 0 0.0011
> 0 0 0.0015
> 0 0 0
>
>I think the step (3) is wrong, but I do not why. Is it due to the
>discordance among SPM files?
>
>Sincerely,
>
>Shoichi
|