I am trying to investigate the coherence between two shoulder muscles at 0 - 90Hz. Precious studies used various cuff-off for their band pass filter, e.g. 1–400 Hz(Chen, 2012), 5-300Hz(Boonstra,2008), 3-1000Hz (Kattla, 2010). How do I determine the most appropriate filter for my analysis?

Also, we measured both surface and intramuscular EMG, but some intramuscular EMG signal are not good. Is it acceptable if I downsample the intramuscular EMG of one muscle, and calculate the coherence between it and the surface EMG of another muscle?

Thank you!!