Hi Francois,
Could you help me few more steps forward?
I use 40s long baseline for Z-score normalization. Now when I have done comparison to dSPM, I have calculated noise covariance matrix from longer (300s) period of time. The baseline file is of course the same and the 40s period is part of the 300s period. The baseline is resting state eyes closed.
Question 1) Can you estimate, how big error is made if I use 300s for dSPM and 40s for z-transform normalization and then make comparison of the results? Or should I use 40s also for the noise covariance? In tutorial it was said that the duration of the data used for the calculation of the noise covariance matrix should be as long as possible.
Question 2) Is it possible to make atlas where I have some structures from deep structures like thalamus and amygdala and then some structures from Desian-Kiliany atlas? I tried to export amygdala using "save selected scout to file" icon on the scout tab and then changing to my own atlas and using "Add scouts to atlas" in the atlas menu, but this does not work. It brings the name (Amygdala) to the atlas but I'm not able to display the scouts time series. It gives error: Scout "Amygdala R" is not included in the source model. If you use this region as a volume, create a volume scout instead (menu Atlas > New atlas > Volume scouts)