Sample rate of 1kHz for EMG cursor control, How?

Comments

• Billh

  February 2018

  Do you know from prior experience that the frequency at >250 Hz works better for your cursor control with your application than amplitude at for example 150 Hz? 

  
  

  Individual facial motor units would not generally be expected to fire at over 100 Hz. Higher apparent frequencies are due to overlapping of many motor units. So at that point why not use the summed amplitudes at, say, 50 to 200 Hz instead, since it ought to be measuring much the same thing as frequency above 200 Hz would give you?

  
  
• skiria8

  February 2018

  I'm still in the design phase. I have experience in EEG but I've just started with EMG. All the papers I found, use 1kHz widn bandpass 20-500Hz, that's why I assumed those were the reqs. 

  
  

  Do you know any good reference (paper/blog/site)?

  I was thinking of detecting if a muscle gets fired by bandpass->rectification->enevelope detection->thresholding. This seems like a standard procedure for EMG. However, if you know a better method for facial EMG of 1 muscle, I would appreciate. 
• Billh

  February 2018

  Surface EMG right? That sounds okay. If it works well at the 1000 Hz sampling rate, it should also work with a data rate of 250 Hz. The threshold you want to detect might change, but you can adjust responses accordingly.

  
  

  
  

  
  

  
  

  
  

  
  

  
  

  
  
• skiria8

  February 2018

  Thanks so much. I'll try with 250Hz sampling rate using a bandpass of 10-125Hz. Hopefully that works. 
• pushtheworld

  March 2018

  @skiria8
  
  You should purchase a WiFi shield! This is exactly why we made it
  
  https://shop.openbci.com/collections/frontpage/products/wifi-shield?variant=44534009550
  
  There is also a 10% discount going on now