SSP projectors in Neuromag system

Hi all,

I am preprocessed the data acquired with a Neuromag system and I have a couple of questions regarding the SSP cleaning.

  1. Is it normal that when computing the SSP projectors for both heartbeats and eyeblinks separately for the MAG and the GRAD, the percentage that the gradiometers components explain is always much smaller than the magnetometers?

  2. I am aware that the order matters, and SSP to heartbeats should be applied before the eyeblinks, but between magnetometers and gradiometers the order matters?

And should I do first SSP ECG for MAG and GRAD and then eyeblinks or first both ECG and EOG for gradiometers and later for magnetometers?
(I mean, which order would be better: ECG MAG / ECG GRAD / EOG MAG / EOG GRAD or ECG MAG / EOG MAG / ECG GRAD / EOG GRAD).

  1. I have been playing around with it a bit, and for some subjects the eyeblink activity seems to be detected fine, but there is one subject that the SSP MAG projector that explains the maximum amount of signal gives me this topography:

It seems that it explains some dipole signal in the left temporal, but to me it doesn't seem like eyeblink activity. Am I correct? Should I consider this as an artifact as well? I checked the sensors and none of them seem to be bad..

Sorry for so many questions, I am still trying to get the SSP procedure :slight_smile:
Many thanks in advanced,

NatĂ lia.

1 Like

Hi Natalia,

  1. Yes it is normal.
  2. The order GRAD/MAG doesn’t matter, the sensors are processed independently.
  3. No you should not remove any component that do not match the type of artifact you are targeting. You need to go back to the list of events that were marked as bad and keep only the ones that clearly eye movements.

Additional information:
http://neuroimage.usc.edu/brainstorm/Tutorials/ArtifactsSsp#Troubleshooting
http://neuroimage.usc.edu/brainstorm/Tutorials/SSPCookbook
http://neuroimage.usc.edu/brainstorm/Tutorials/TutMindNeuromag#Heartbeats_and_blinks

Cheers,
Francois

1 Like

Thanks Francois!

It appears that my subject only blinked 9 times in 5 minutes and the detect EOG algorithm only detected 5 of them, so that is why the SSP projector was so weird.

I will manually select all the EOG events, although I am afraid that it may be too small number…

Anyway, thanks!