Hi Moahmed,

actually there are many methods for denoising and removing artifacts. Which one is best suited will depend on your signals and on what you further plan to do with them. For an overview on EEG preprocessing see (for example, there is much literature): https://www.ncbi.nlm.nih.gov/pubmed/27751622

Maybe this gives you some more detailed idea.

Greetings, Sebastian

Hello,

One of the most common techniques is the Hilbert-Huang Transform. It is used to separate nonlinear components and isolate the noise by this method:

http://inter-eng.upm.ro/2012/files/proceedings/papers/paper71.pdf

I hope that this information was useful

There are several kinds of artifacts and noise in EEG data, and so the best techniques depend on the type of noise (e.g., spatial or temporal, narrowband or broadband).

I have a lot of EEG data analysis teaching resources linked on my website that you might find helpful. mikexcohen.com

Mike

In a review, i found fastICA may be the most used technique for noise removal. It depends on the type of noise also. Band pass filters have been used to de-noise a signal if the frequency of noise is know like 50 Hz AC line noise.

What broadband sources of noise could there possibly be in rat EEG in the gamma range (40-80 Hz) if narrow band 60 cycle nose is filtered out? This seems too slow for muscle fibers, and too fast for EOG, EKG, or movement artifact. There are no movements in the rat that are that fast.

Hi John Salamone

is this actually extra-cranial recording or implanted? EMG is a likely suspect (in humans fast-twitch EMG can be that low in frequency (not sure about the rat). I'd be happy to help point towards some good cleaning pipelines that may help. depending on what you're trying to look for ICA/PCA may do well or you may be better off with some aggressive filtering. Happy to chat!