I am working with about 180 seconds of resting state data for 32 participants (pediatric OCD and controls).
I was wondering if anyone is able to tell me the source of these big low frequency jumps in my data? I see them in almost all my participants. It seems to be coming from sensors in the right frontal and right temporal region. Do I need to worry about cleaning them?
If I do have to correct for this, what would be the best procedure in brainstorm? I noticed that if i apply a 1Hz high pass filter most of it seems to disappear, but not sure if that is the best way to remove this.
Otherwise, if you are not interested in what is happening below 1Hz (no slow activity expected in the brain), you could simply apply a high-pass filter at 1Hz. But it would affect a lot more the recordings…
Thank you for your response. Unfortunately, we did not use EOG sensors so it is hard to tell if the origin is related to eye movement. Do typical eye artifacts have such high amplitudes? Also is ICA the best way to remove eye artifacts when no EOG sensors were used?
For SSP, mark the peaks manually. For ICA use all the recordings.
For the general procedure, follow what is presented in the introduction tutorials (if you haven’t read them at least until #19, I recommend to go back to them).
If with either technique you manage to get one or two components that capture most of the artifact, and with topographies that look like blink or saccade topographies, you’re good.
You have to try and see what you get.
