Estimating broadband EMG noise power?

[Steve Miller]

I'm trying to measure muscle tension between two conditions, and it seems like a good way to do it would be to try to measure the broadband EMG noise that is typically removed for EEG. The simplest way to do it would be to use ICA. However, I was wondering if there was some way to estimate the value without ICA because 1. I'm not using very many electrodes so I'm not sure how well ICA will work, and 2. It would be great to be able to measure it in real-time.

[Mike X Cohen]

Hi Steve. If you are getting the data from scalp channels (and not EMG electrodes), then I think ICA is the best approach. The signals are mixed, so you will need some kind of multivariate source separation method. 

You can try filtering the data from, say, 15-40 hz for the ICA. Also keep in mind that ICA can take a while to compute (but should be quick with few channels), but once you have the weights, you can apply them to new data. Thus, perhaps you can have the subject produce the EMG artifact, run ICA, and then run the actual experiment once you've picked a component. 

Mike

On Tue, Aug 20, 2019, 03:03 Steve Miller <slmi...@gmail.com> wrote:

I'm trying to measure muscle tension between two conditions, and it seems like a good way to do it would be to try to measure the broadband EMG noise that is typically removed for EEG. The simplest way to do it would be to use ICA. However, I was wondering if there was some way to estimate the value without ICA because 1. I'm not using very many electrodes so I'm not sure how well ICA will work, and 2. It would be great to be able to measure it in real-time.

--
You received this message because you are subscribed to the Google Groups "AnalyzingNeuralTimeSeriesData" group.
To unsubscribe from this group and stop receiving emails from it, send an email to analyzingneuraltimes...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/analyzingneuraltimeseriesdata/4de81d8c-acd2-4a99-91dd-db343444eef7%40googlegroups.com.