Leff ratio in dipole scanning and whitened recordings

[ftadel]

@jcmosher suggested we change the scaling factor applied in two functions:

1. Dipole scanning (not implemented)

https://github.com/brainstorm-tools/brainstorm3/blob/master/toolbox/process/functions/process_dipole_scanning.m#L124 Factor = sqrt(DataMatP.Leff / sResultP.Leff); Replaced with: Factor = sqrt(sResultP.Leff);

Problem: When testing with the introduction tutorials, the goodness of fit of the dipoles drops to very small values, example for standard condition at 90ms (Leff=193):

Current code (top) vs. John's suggestions (bottom):

2. Whitened recordings (already implemented)

https://github.com/brainstorm-tools/brainstorm3/blob/master/toolbox/tree/tree_callbacks.m#L3547 Factor = sqrt(DataMat.Leff / ResultsMat.Leff); Replaced with: Factor = sqrt(sResultP.Leff);

This looks more reasonable (but do we want to see Z=20 here?). Below, figures obtained by right-clicking on the MN source link for the standard condition (Leff=193) > Model evaluation > Save whitened data (top=before, bottom=after modification):

@jcmosher @dimitriospantazis please validate whether this is what you want to obtain.

[jcmosher]

https://neuroimage.usc.edu/brainstorm/Tutorials/TutDipScan

[ftadel]

@jcmosher ??

[ftadel]

@jcmosher @dimitriospantazis @sbaillet If you have time now, can you please have a look at this?

[ftadel]

@jcmosher @dimitriospantazis @sbaillet Do you just want to ignore this question until someone spots a possible bug?

[ftadel]

This issue is still pending....

[ftadel]

@jcmosher @dimitriospantazis @sbaillet