The way to conduct wPLI for unconstrained source maps

Dear Brainstorm contributors,

I have a question regarding phase-based connectivity for unconstrained source maps.

I have calculated my EEG data to unconstrained source maps and have attempted to conduct wPLI analysis. However, whenever I try to do so, I receive the following message: "The PLV measures are not yet supported on unconstrained sources." The tutorial indicates that "if using unconstrained source maps, with 3 time series at each brain location, this requires an extra step to derive connectivity estimates." However, I cannot find an explanation for this "extra step."

Could you please advise me on how to address this problem?

Best regards,

@Sylvain @Raymundo.Cassani @Marc.Lalancette @leahy @danielemarinazzo?

The other contributors, in particular those who coded this part, will be able to give a much better answer. What I think this message refers to is that you have three series for each dipole, and you should choose what to do (use all three of them, or a combination of them, which could on the other hand defeat the original purpose of having unconstrained sources). For example for visualization you reduce the dimensionality by taking the norm of the vectorial sum of the three orientations at each vertex.


It means converting unconstrained source map to constrained prior to doing connectivity. The process is under Sources > "Unconstrained to flat map".

We've been working on fixing PCA, which is probably what you'd want to use for this. This fix is close to being distributed, probably a matter of weeks, though you can try the existing version as well. Another option is to use a constrained source model from the start if possible.

Dear Daniele and Marc,

Thank you for your kind reply. I understand that I need to convert my unconstrained data to constrained data. I tried using the "Sources > Unconstrained to flat map" function, but I received the following message: "Cannot use this operation on source files attached to continuous RAW files." Does this mean that the function cannot be applied to continuous EEG data?

I am analyzing resting state EEG data, which is all continuous. If I cannot use "Unconstrained to flat map" on my data, is the only solution to create a source model again with a constrained map instead of an unconstrained map?

Best regards,

I think you just need to import the data so it is no longer a "raw" type.

Thank you very much for your kind response. I would love to try it out.

Best regards