Hello, I am not sure if things like this belong to this forum but I guess there's nothing wrong about it:
I am newbie to EEG signal processing and I am wroking on my bachelor thesis. I am analyzing what effects does tACS delivered with gamma freq have on subjects. I have before and after recordings. Now I am doing segment and channel rejection and I would like to ask an expert. Are those 2 (frontal) channels with significantly higher gamma power probably bad (noise) or can it also be that the gamma activity is really that high?
From my research and discussion with my supervisor I know, that gamma activity is significantly harder to record because it is weaker than the lower freq bands and it can easily get confused with noise. So if it is almost 10 dB stronger at some freq ranges + it is consistently distributed from 50 - 90 Hz I assume it is just noise and I would throw the channels away. But I hesiate a bit in this case.
Plus are there some more fool-proof methods reliable enough for determining if something is/isn't noise for this kind of case?
Thank you very much for advice.
1 Like
EEG can often be decomposed in background activity, decaying with frequency and having some periodic components, as you can see, for example, at around 10 Hz. If there would be an enhancement in gamma activity I would not expect it to be so equally distributed in power in the whole band from 50 to 90 Hz, but still have a bit of decay over frequency. In your case, it looks more like hitting some kind of noise floor. There is no "proven" foolproof way of determining if something is noisy or not, but it might be interesting to look at the specific time series of the channels that stick out. Good luck!
2 Likes
Hello,
I'm really not and expert in this topic, but I'm very interested in correlation within ASD/ADHD/BPD and Schizophtrenia in terms of measurable fMRI or EEG data (these are most available data in Europe, IDK where are you from), as there are proven DNA inherited correlations (not necesseraly cause and effect relation). I have limited budget as well, but I'm detetermined in terms of analizying anonalies within this given values, It can be difficult, know my limitations, but I really want to go deeper into that. IMO free hardware limitations data is the major obstacle. Are you interested?
Greetings,
Mike Zommer
