Hello everyone, I implemented a few methods to remove muscle artifacts from EEG signals, but I can not understand how I can clearly (not visually, not spectral analyzer, not SNR) determine which is method better? is it possible in % determine how much EEG signal was damaged after artifacts were removed?
I attached an example of my data, before and after artifact cleaning
you could ask yourself why you want to remove muscle artifacts. Is it to have cleaner ERPs, for example? Then you could compute your “final measure of interest” (ERP, a classification), and compute quantitative metrics on that final measure (SNR of ERP, classification score, …), and repeat that once per method of your implemented methods. Finally, you would compare these quantitative metrics.
