Dear Avijit,
The departure from sphericity is the unequal variance setting, basically
meaning to use Welch's t-test instead of Student's.
As advised by the error message, the recommendation here is to check the
input images - use CheckReg to display the images from all subjects and
see whether there is something peculiar for one or more subjects in the
scaling or mask of the data.
Best regards,
Guillaume.
On 06/12/2019 12:31, Avijit Chowdhury wrote:
> Dear experts,
>
> I have constrasts from 2 different groups of participants (healthy and patient), and each group with unequal participant size.
>
> I have ran the same 1st level design for each participant from both groups, and calculated constrasts of interest for each participant. I am trying to do between group comparison using these contrasts in 2nd level.
>
> I use a two-sample t-test under 2nd level design specification.
>
> The specification is successful. During estimation, however, I get the following error:
>
> ------------------------------------------------------------------------
> Running job #2
> ------------------------------------------------------------------------
> Running 'Model estimation'
>
> SPM12: spm_spm (v6842) 20:15:31 - 06/12/2019
> ========================================================================
>
> SPM12: spm_est_non_sphericity (v6913) 20:15:31 - 06/12/2019
> ========================================================================
> Chunk 1/1 : ...done
> Failed 'Model estimation'
> Error using spm_est_non_sphericity (line 208)
> Please check your data: There are no significant voxels.
> In file "E:\Avi\spm12\spm_est_non_sphericity.m" (v6913), function "spm_est_non_sphericity" at line 208.
> In file "E:\Avi\spm12\spm_spm.m" (v6842), function "spm_spm" at line 431.
> In file "E:\Avi\spm12\config\spm_run_fmri_est.m" (v5809), function "spm_run_fmri_est" at line 33.
>
> The following modules did not run:
> Failed: Model estimation
>
>
>
> I believe this has something to do with temporal auto-correlation? But why would this be the case, since this is 2nd level analysis there is no temporal element. I have used "Independence = yes" and "Variance = unequal". I would really like to know how to fix this.
>
> Any help is greatly appreciated. I can email the data to anyone who may be able to help.
>
> Thank you,
> Avijit Chowdhury
>
--
Guillaume Flandin, PhD
Wellcome Centre for Human Neuroimaging
UCL Queen Square Institute of Neurology
London WC1N 3BG
|