[FieldTrip] cortical time-course reconstruction

[Paolo Belardinelli]

Hi Robert.
We have some nice coherence maps with peaks in M1 for a finger tapping 
task. We were wondering how to properly reconstruct the power 
time-course in the M1 area in order to use it as a further reference
signal.

-Is the code necessary for the approach described in (Schoffelen et al., 
NeuroImage 2008) already implemented in ft (we observed in the code that 
it is possible in ft_sourceanalysis.m to give a location for a
reference dipole but we found no documentation)?

-Is the only reliable solution for reconstructing a time-course to 
project the activity in the dominant direction? We have the eigenvalues 
of the first dipole which are normally 3 to 6 times larger
than the ones of the second. We are not sure this is enough to neglect 
the second component.

- One last technical question: in the mailing list it is said that the 
orientation of the dipole is saved in source.avg.ori but it doesn't seem 
like that to us. Maybe we are doing something wrong.

Best regards and thank you for your work,

Paolo and Erick