Dear SPM-M/EEG experts,
I am currently trying the "Imaging approach" on the MEG-data. I succeeded
all parts of it including Source Space Modelling (mesh), Data Coregistration
(datareg) and Forward Computation (forward). But I got a problem in the end
of the Inverse Solution for induced (and evoked) power - m-file:
spm_eeg_inv_induced.m. I've a mistake on the line number 244:
% cross-energy in source space
Gind = MAP*Eind*MAP' + C*trace(K*V);
Gind = diag(Gind);
The matrix Gind is not calculated.
In my settings I have 148 channels and 3004 points of the brain for inverse
solution. Therefore, C*trace(K*V) has dimension 3004 x 3004 and it is
calulatable, but first members of the first part have dimensions: MAP =
3004 x 148 and Eind = 148 x 1. As you can see the equation is incalculable.
My question is - did I mistaken somewhere or there is a 'bag' in SPM. If I
am wrong could you please tell me presumably what mistake it might be and
how should the matrices look like.
Thank you a lot in advance.
Mikhail Zvyagintsev
|