Artifact created by saccade removal with SSP

Discussions
,
1

Hello Brainstorm forum!

I am new to MEG preprocessing/brainstorm, so bear with me. For some background, I went through the first half of the brainstorm tutorials with my own data. The participants had their eyes closed during the recording, so saccades were the main eye artifact generated. I followed the tutorial's suggestion for removing saccades (detect other events --> rename the 1-7 Hz artifacts saccades --> remove saccades with SSP). The first SSP component was the only component that reflected eye movements, so it was removed. After the removal, an artifact was left at the beginning/end of some of the saccades (highlighted in the image below). Did I do something wrong when removing the saccades? Is there any way I can remove these "steps" in the data, or will I have to mark all trials with these artifacts as "bad"?

Screen Shot 2021-07-31 at 10.05.23 AM
Screen Shot 2021-07-31 at 10.05.23 AM1752×1532 708 KB

Thank you all!
Jennifer

2

Unfortunately, artifact removal is not an exact science...
Saccades create artifacts that are not as reproducible as blinks or heartbeats, removing them is not always possible and often require many trial and errors.

You can try a few more things:

  • Creating shorter events around the saccades (instead of the 3s one on your screen capture)
  • Trying with different frequency bands
  • Trying with ICA instead of SSP. See the EEG-based example from the tutorials:
    https://neuroimage.usc.edu/brainstorm/Tutorials/Epilepsy#Artifact_cleaning_with_ICA
  • Make sure the projectors you apply do not alter much the recordings away from the saccades. If you don't manage to get rid of them by removing one or two SSP/ICA components, it would be safer to mark the segments as bad (if you have enough data left...)