I'm going to examine the relation between the the surface electromyography (sEMG) signals and the arm movements. I have a large number of sEMG signals acquired simultaneously with a large number of human arm joint angles during different movements. Those signals were recorded from forearm and upper-arm muscles. I need to select from those sEMG a subset that contains the most informative channels representing those movements, i.e., I need to exclude the noisy signals from the analysis. So, rather than the visual inspection of the sEMG with the angle signals, is there a method to accomplish this goal? I know there are dimensionality reduction techniques, like the PCA, but I have to avoid this approach.

More Mohammad Alothman's questions See All
Similar questions and discussions