Hello
Question 1) @Sylvain: Should I include all the subcortical structures or should I include only those that might get activated most probably? I only had 64 active EEG electrodes and I have no idea how inaccurate localization is in these deep structures as compared to more superficial ones.
Question 2) @Sylvain: Do you think that including some subcortical structure(s) is better option than using whole brain as source model if there is at least small hint of subcortical activations?
With 64 EEG electrodes, you can't expect a high spatial resolution in your source analysis. Trying to model individually of the various deep structures may improve your results but not much, while they can introduce a lot confusion in the interpretation of the results. These "mixed head model" results are difficult to visualize and process.
I'd recommend you start by getting results you can explain with a more generic model, and then, only if you need to, refine the location of the sources with this "mixed head model" approach. Use a "volume head model" with a regular grid of sources, which won't impose any constraint of location or orientation on the dipoles in your model:
https://neuroimage.usc.edu/brainstorm/Tutorials/TutVolSource
If you really want to use these deep structures, use them all, not only the ones were you want to observe activity, otherwise you introduce a strong bias in your analysis
The fact that you get all your activity at the center of the head might be an issue related with the dSPM computation and/or your noise covariance matrix. Try to compare the results obtained with the option "current density maps", normalized with a Z-score, as in the introduction tutorials:
https://neuroimage.usc.edu/brainstorm/Tutorials/SourceEstimation#Z-score
Question 3) I have a baseline measurement in separate file. Can I use this to calculate Z-score transformation for MNE? Is this in fact identically same as calculating dSPM? In Z-score transformation, the only ínput option shows the time span in current file and I cannot select input from another file. So what I would like to do, is to compare dSPM and MNE with Z-score normalization in deep sources by using the same baseline definition.
If the two files (baseline and the recordings of interest) are in two separate files: compute separately the sources for the baseline, the sources for the data of interest, and use the Process2 tab to standardize one with the other (FilesA=baseline source file, FilesB=source file, process "Standardize > Baseline normalization").
Question 4) How can I compare activation maps between subjects? If I select group analysis, how should I proceed with the mixed model. For example if I want to use Desian-Kiliany atlas as scouts and some structures from aseg atlas for comparison purposes?
- Option 1: Use scouts (process "Extract > Scouts time series) for each subject and compare the time series you obtain for the various regions of interest.
- Option 2: Estimate the sources for a common source grid, projected to the individual subject spaces with the MNI transformation computed with SPM:
https://neuroimage.usc.edu/brainstorm/Tutorials/CoregisterSubjects#Volume_source_models
Question 5) I have one collaborator who is doing some FEM modelling and solving inverse problems. We would like to compare BEM and FEM results. Is there some place where I can see all the parameters that are used during source reconstruction in Brainstorm? For example, he would like to know what is the time window used for calculations etc?
For the estimation of the minimum norm source maps, you define all the time windows yourself.
Time and actual EEG recordings are not used in the actual computation of the forward model, no matter which one (single sphere, BEM, FEM...). Only the anatomical information and the position of the electrodes is needed.
Please list the exact list of details you would like to know more about and I will try to get this information for you.
François