Hello all,

I am attempting to run a basic PPI using a VOI I recently created.  The bulk of what I am using has worked previously, including contrasts, VOI creation, estimation, and the like.  However, I get the following error:
Error in file_array/subsasgn>subfun (line
132)
mat2file(sobj,dat,va{:});

Error in file_array/subsasgn (line 70)
subfun(sobj,double(dat),args{:});
Error in spm_write_plane (line 33)
V.private.dat =
subsasgn(V.private.dat,S,dat);
Error in spm_write_vol (line 83)
V = spm_write_plane(V,Y,':');
Error in create_mask_image (line 67)
spm_write_vol(spmhdr,maskimg);
Error in set_mask (line 94)
xY.mask=create_mask_image(SPM,region1,xY.XYZmm,maskdir);
Error in timeseries_extract (line 191)
[xY,errorchk]=set_mask(SPM,P.VOI,region1,P.maskdir);
Error in PPPI (line 395)
timeseries_extract;

Any suggestions?  Even knowing what the problem is in more detail would help in making repairs

Do you have permission to write to the SPM directory?

Can you send me the VOI file?

I should, it's not in any place that requires permissions.  It's also on a Windows computer, so usually fewer issues with that.
I've attached the VOI

Ok. I'll see if I can replicate the issue locally.

Akos,

It works fine on my macbook. I couldn't replicate the error.

Have you changed computers, operating systems, or the SPM installation? I did a quick search on the SPM list and it seems that the problem is due to the mat2file mex file. One person reported that the continuous backup of his computer caused an issue. I would check the mex-file and SPM installation.

Best,
Donald
Originally posted by Akos Szekely:

Hello Donald,

Thanks for getting back to me about this.  I ended up testing a few different options in the mean time to get around this, particularly since I haven't changed computers, OS, or SPM.  I don't think I even backup my computer that often since there's not much space on the drive for it.
In any case, I tried making VOIs manually using the eigenvariate button, which worked to an extent, but kept given "VOI is bigger than the data set" issues.  Only recently did I recheck the gPPI manual and find out to multiply the VOI by the group mask.  By that time I'd begun using create_sphere_image, which was a lot easier to change with imcalc.  I ended up creating VOIs using each subject's individual SPM.mat file, and then multiplying by the subject's mask.img file.  Finally it seems to work.

Best,
Akos Szekely