Hi Nils,
This is something that had been fixed just after the last SPM version was
released.
It should work fine if you change Line 79 in spm_dcm_ind_results.m into:
TF{i,j} = TF{i,j} + xY.xf{i,j}(:,1:nf)*xY.U';
It will be fixed in the new SPM release but for now you could just change
it in your code.
Hope this solved the problem!
Best,
Bernadette
2014-09-30 12:14 GMT+01:00 Nils Rosjat <[log in to unmask]>:
> Dear SPM experts,
>
> I tried to use the "Wavelet transform" button in the DCM for M/EEG window,
> but I get an error message. The same happens when I try to save the files
> as images after inverting the DCM. I think that is because it uses the same
> function from spm_dcm_ind_results.m (case{lower('Wavelet')}).
>
> The error occurs, because the matrix dimensions of xY.xf{i,j} and xY.U'
> don't match. The problem seems to be that in line 57 you remove the
> unmodelled frequency modes in xY.U. That leads to a reduced dimension if
> you use less then 8 modes (for example dimension 4x45 instead of 8x45). At
> the same time the dimension of xY.xf{i,j} stays the same (for example
> 144x8). At the moment I replaced xY.U' by xY.Uold' (that is a copy of xY.U
> before removing the unmodelled modes). Is this a possible way to do it or
> would that lead to a wrong time-frequency output?
>
> Best,
> Nils
> __________________
> Dipl.-Math. Nils Rosjat
> University of Cologne
> Biocenter
> Zuelpicher Str. 47b
> 50674 Cologne
> [log in to unmask]
>
