Removing a lesioned area

[tmedani]

@romboide2004

This is what is not totally clear to me. By "within" the GM, in a volume model, is it meant all the volume enclosed by the GM surface? If so, if a lesioned tissue is (partially) segmented within the volume enclosed by GM, that will also be included in the source space, right?

Probably yes, if you have the final model, we can check it.
But for these specific cases, I recommend that you define a custom source space and then overlay it with the lesioned brain for visual checking.