I have a concern about the "Source estimation" tutorial. In the "Source map normalization" paragraph it is recommended to use non-normalized current density maps in scenarios such as single trial analysis or computing shared kernel. However, there is strong evidence in literature (as also stated at the beginning of that paragraph) on the inaccuracy of this method compared for instance with sLORETA. Is it therefore being suggested to opt for lower accuracy in order to gain a more efficient computational approach, maybe?
Thank you in advance for your help