Hello, do you have any ideas for removing walking artifacts from a few channels of EEG signals? Currently, I have found one way to use accelerometers to summarize as noise with EEG signal and after that, via the EMD method, let’s say “components” with these types of fluctuation and remove these components from the original EEG data, is it sounds logic and maybe exist any methods for few channels without any additional sensors?
Hi,
I would suggest you to take a look at Artifact Subspace Reconstruction methods as it is used in BCI for movement removal. There is some work going on to implement ASR into MNE see the discussion here .
I used also sometime this python package GitHub - DiGyt/asrpy: Artifact Subspace Reconstruction for Python.
Best,
Samuel
1 Like
Hi Samuel, ASR is amazing method, only in the basic of this method located ICA which requires many channels, in my case I have only 8 channels or I am wrong and this method can works with few channels ?
Hi! Sorry for the late answer, for 8 channels I think it will be difficult indeed to apply ASR. Your setup with accelerometer is interesting but you will not capture some muscles artifacts.