Although I don't have any experience with electrophysiological recordings from animals, the most common source of EEG noise I've noticed from human work (I do movement-related research) has been due to wire movement. I'd suggest securing the wires as much as possible.

Could you by any chance provide some more information on the movements the animals are doing, the electrode setup, or an image of the noise in question? You might be able to filter out the noise during analysis to save some trials, assuming the noise is not increasing the power in your frequencies of interest.

Kind regards,

Chris

Hello,

If it creates spikes on your signal (artifact noise) maybe first you can try to smooth the signal with a median filter.

Another thing that you can do and I am also reading about, is the decomposition of the signal with EMD into different nonlinear signals to filter out some of the components that might be negatively affecting the raw signal.

In the frequency domain, the spectrum of artifact noise is on a specific range, maybe you can try to filter this band and see if it is still useful for your results. I am curious on what kind of tests are you performing with the animals.

I hope this information is useful.

Motion artifact is as old as EEG recording, occurring at the interface of the electrode with the tissue, be it skin, cortex or deep brain.  Placing the pre-amp close to the electrode helps deal with ambient noise, such as from mains (60 Hz, 50 Hz) but not with motion artifact.  Same thing with stabilizing the wires so that they move less within the ambient electrical field unless, of course, wire movement is causing electrode movement when the animal moves.  Filtering may work.  You might be able to detect motion transients and remove them.  It's possible that you have EMG artifact as well as motion artifact, and filtering may be applied to the EMG artifact.  Generally, however, one must focus on using the clean EEG signal between periods of motion.

Hello everyone,

For the benefit of getting clear clarification i will upload the file which shows the prob i indicates during my study.  Looking forward for your  assistance. Thank you

Hello Christopher and Bruno

Thank you for the reply and its a good idea to work with the filters to resolve the noise. i will work on it.

Regarding the animal movement,  the noise usually occurs when the animals walks in the recording chamber. but no noise when it grooms or sniff. any idea?

check this link for our lab details:

http://lifeserv.bgu.ac.il/wp/azangen/index.php/personnel/addiction-2/

Hello Valentin Neubert ,

Yeah i am using the preamp to record the signals. So do you think the movement of the preamp with electrode produce noise here? but i am making sure that it dont move , however i cannot restrict the  movement of the rat as well. any suggestions?

James C. Miller,

Thank you very much for your suggestions, i will work on the filter.  If you don't mind Could you please specify the name of filter specifically to reduce this kind of noise? . I am a beginner in analysis part.

In the picture you uploaded, the trace speed is too slow to see the characteristics of the noise.  If the noise is at the frequency of the mains, then the motion is causing increased impedance at the electrode-tissue interface and the interface becomes an antenna that picks up the ambient (mains) electromagnetic field.  If the noise is at the frequency of EMG, then filtering in the EMG frequency band may be useful.  If the noise has spike characteristics, then perhaps an electrical  connection is coming slightly loose during motion, making and breaking contact.  Generally, mains and EMG frequencies are above most EEG frequencies of interest (0.5 to 30 Hz); thus, a low pass filter may help.  Historically, the removal of motion artifact has not been accomplished with much success.  Thus, the most practical solution is to simply use EEG data collected while the animal is not moving.  Usually, this requires manual selection of data epochs and a method for averaging across these epochs in the frequency domain.

I to have worked on EEG of wrist rotate signal, leg rotation of human.

Placement of the electrode is important cerebral cortex that to p7, p8 position is where the Mu signals are much accurate. U could sample the signal at a higher rate and use a LPF and BPF. My paper on Hand rotate EEG signal will be helpful for you if you refer it.

I would suggest to make a study of the signals of animals EEG signal and classify them like we have to Humans alpha, Beta, delta, Mu, gama etc.  

U can also use DWT of order 2 to analysis the signal in better way.

Any help you need i can work on ur signals and send it back to you regarding pre processing.

This publication of mine might help you.

https://www.researchgate.net/publication/261301154_Hand_rotate_EEG_signal_feature_extraction_by_second_order_Daubechies_wavelet_transform_DWT

Article Hand rotate EEG signal feature extraction by second order Da...

Hello Stafford,

Thank you for the reply and involvement, Could you please help me for the selection of filters or procedure in Pre processing to avoid the noise during movements.?

Please have a look of this signals

https://www.researchgate.net/file.PostFileLoader.html?id=57b03d2c5b495262000a9964&assetKey=AS%3A394920482361344%401471167788376

Hello sundareswaran 

I have look at the signal that u have sent.. I need the signals in a raw format so that i can apply some filtering and ans back to you.

Since i have not worked on animal EEG.. I need some inputs form your side to work on the filter..

Hello Stafford,

I am working on it and get back to you shortly. Thank you for the reply.