My interest in Muse has led me to find articles where they compute connectivity features for AI with Muse device. They compute in this article "Mental State Recognition via Wearable EEG" arXiv:1602.00985 pairwise correlation between each power within each frequency band and each electrode. Is this feasible with up to date Brainstorm? Is it possible to compute other connectivity features other than correlation? If not any suggestions on how to perform this on my own with any toolbox(MNE, FieldTrip, ...)?
If not any suggestions on how to perform this on my own with any toolbox(MNE, FieldTrip, ...)?
You can also do this with MNE-Python and FieldTrip.
You'd need to learn how to write code (Python or Matlab) instead of learning how to click on buttons., but you could get to more or less the same results.
Brainstorm reads natively the Muse .csv files, I'm not sure if this is the case with the other programs.
I am aware of Brainstorm tutorials the issue here is that with Muse only 4 channels I am afraid for connectivity features you have to compute the inverse solution but that is not possible for Muse from previous posts. Right?
That would not be reasonable indeed to compute full brain maps from only 4 electrodes. Also, the pairwise correlations at the electrode level you mention above are prone to crosstalk and signal leakage between channels, that tend to increase cross-correlation estimates. This needs to be controlled carefully.
partial lag coherence could be a choice(overcoming crosstalk)? if that is the case how to apply that to the muse device data? I don't know if you can compute PLC without full brain maps. Is it possible in Brainstorm? Or shall I use other software instead? Or is it even possible at all?