The digitized (green) Polhemus points of most of our participants are not within the MEG helmet in Brainstorm. (As shown in the attached picture.) I’m confused and worried that this will lead to difficulties in localizing the signals we have measured.
Is this strictly an acquisition issue? Can anything else cause this?
Is there anything we can do in the processing side to fix this issue or at least maneuver around it?
Will we be able to do any sort of localization if the digitization is an issue and how?
What kind of operations have you applied to get to this result?
When linking the recordings to the database (in fif format I guess?), you should not perform any automatic registration, for two reasons: it looks like you are using an anatomy template therefore the digitized head points cannot match the anatomy, and the head points you picked are too sparse for this algorithm to work well.
Are you sure you are using the same convention for the fiducials (NAS,LPA,RPA) during the digitization and in the MRI (in the MRI viewer in Brainstorm)?
The anatomy templates in Brainstorm use by default the points as defined on this page: http://neuroimage.usc.edu/brainstorm/CoordinateSystems
If you have been digitizing different head points, you have also to change the point selection in the MRI of the subject (or the template anatomy), so that the two match and can be aligned correctly.
I have a similar issue and I am trying to manually align the HPI to the subject's head for a run after importing it. I want to apply this transformation to all runs.
Is there a way to save this to the anatomy, so that if a run is imported later it takes the modified transformation and not the original one 'while being imported'? Or should I do this manually for each run after it is imported?
If all the runs are already coregistered to the same head position (using MaxFilter for instance), you could copy-paste the modified channel file file to the other runs.
Note that this would also copy SSP and ICA projectors, so do this before you do any advanced artifact cleaning on the recordings.
Otherwise, if the run are not registered initially, you can't reuse the same transformation, you need to define a new one for each run.