Dear Doug,
There are two struct variables in this line: SPM (from the SPM.mat file)
and xY (from the VOI_*.mat file). Both need to have a Sess field.
An SPM variable from a first level fMRI analysis will always have a Sess
field, ie SPM.Sess.
A xY variable will have a Sess field if the time series are extracted
from a first level fMRI analysis:
https://github.com/spm/spm12/blob/r6906/spm_regions.m#L137-L145
From what you describe, it seems that the problem is with xY: How do you
create the VOI file (ie script, eigenvariate button, batch interface)?
What is returned by:
>> which spm_regions.m -all
>> spm ver spm_regions.m
It might also be useful to see the content of the problematic VOI.mat file:
>> load(spm_select(1,'^VOI_.*\.mat$'));
>> xY
Best regards,
Guillaume.
On 14/09/17 01:07, Doug Burman wrote:
> Volkmar and others,
> I've made progress on identifying the problem, but I still need help with a fix. The original error message included this line:
>
> Error in spm_peb_ppi (line 241)
> Sess = SPM.Sess(xY(1).Sess);
>
> This actually is looking for a variable SPM.Sess.Sess from the VOI file, which I created from pressing the Eigenvariate option after running some results. The SPM12 version does not create this field. (It does create the SPM.Sess structure, but this does not include another Sess field.)
>
> If I create the VOI file from SPM8 rather than SPM12, there is no problem -- the PPI process runs fine.
>
> So to solve the problem: either the output file from the Eigenvariate function in SPM12 needs to restore this SPM.sess.sess field, or the PPI function in spm_peb_ppi needs to find the necessary information from a different variable.
>
> Doug Burman
>
--
Guillaume Flandin, PhD
Wellcome Trust Centre for Neuroimaging
University College London
12 Queen Square
London WC1N 3BG
|