Spikes in EMG signals

[Larry To]

Hi all,

I perform Long-Term EMG study. I use BioStamp as my main data collection platform. For some reason, lots of my data were corrupted when I use the BioStamp electrodes. There are always a large spike before each recorded muscle activation. ( I attached a picture of the EMG data.) My filtering process is High pass (20Hz), Rectified and Low Pass (15Hz). Signal is collected in 1000 Hz

My question is what causes them and how I can remove them.

My muscle placement is Bicep, Tricep (lateral), ECR and FCU
Screen Shot 2017-07-14 at 2.09.30 PM.png

[Huseyin Celik]

Re: Spikes in EMG signals

Hi Larry, we had a similar issue with our force plate. It was solved when we changed the USB cable that connects the amplifier and the PC with a shielded USB cable. If you still want to use the data you have already collected, we had some success to clean the data using median filtering. Check this MATLAB document if you are interested:

https://www.mathworks.com/help/signal/ug/remove-spikes-from-a-signal.html

Huseyin

[Roux Terence]

Re: Spikes in EMG signals

Hi Larry,

Maybe, it comes from the skin of your subject. Do you shave him before putting the EMGs ? Sometimes, hairs can give you some noises.
As Huseyin Celik said, use a filter on matlab also.

Terence

[Ankita Pramanick]

Re: Spikes in EMG signals

Hi Larry
We used to have similar issues. We are using 8 channel Noraxon wireless EMG. We always shave and clean the area with sterilium before putting on the EMG electrodes
You can try this and see

Ankita