how should I find noise???

Post Reply
Ababy
Posts: 5
Joined: Mon Jan 30, 2023 9:29 pm

how should I find noise???

Post by Ababy »

I think there will be a small amount of noise in measuring EEG, but I don't know how to distinguish noise. Can anyone give me some suggestions or a better way to do this?
User avatar
James
Site Admin
Posts: 1109
Joined: Wed Jan 02, 2013 9:06 pm

Re: how should I find noise???

Post by James »

Noise will show as variance of the RAW EEG channels. You want thin crisp lines with spikes only when you blink.
Ababy
Posts: 5
Joined: Mon Jan 30, 2023 9:29 pm

Re: how should I find noise???

Post by Ababy »

So, you mean that there is noise in the picture below, but what is the criterion for noise?
raw-brain-waves.png
raw-brain-waves.png (165.9 KiB) Viewed 2872 times
User avatar
James
Site Admin
Posts: 1109
Joined: Wed Jan 02, 2013 9:06 pm

Re: how should I find noise???

Post by James »

As a rule of thumb, I'd say around 50uV of variance is ok, so this looks good. You're always going to have spikes outside that because of blinking.. but most of it is within 50mV.
Ababy
Posts: 5
Joined: Mon Jan 30, 2023 9:29 pm

Re: how should I find noise???

Post by Ababy »

You say that a dispersion of around 50uV is acceptable, but in the above figure, the vertical axis scale seems to be more than ±50uV based on 800. To begin with, is it correct that the unit of the vertical axis is uV?
Ababy
Posts: 5
Joined: Mon Jan 30, 2023 9:29 pm

Re: how should I find noise???

Post by Ababy »

I have one more question. In the image below, there are marks such as GF (Good Fit), BF (Bad Fit), etc. By what criteria are these marks made?
raw-brain-waves1.png
raw-brain-waves1.png (145.73 KiB) Viewed 2804 times
User avatar
James
Site Admin
Posts: 1109
Joined: Wed Jan 02, 2013 9:06 pm

Re: how should I find noise???

Post by James »

This image above is terribly noisy. The markers happen when there is large variance on the RAW EEG, such as above. The scale will change automatically depending on the data range.
Ababy
Posts: 5
Joined: Mon Jan 30, 2023 9:29 pm

Re: how should I find noise???

Post by Ababy »

James wrote: Fri Feb 03, 2023 8:26 am As a rule of thumb, I'd say around 50uV of variance is ok, so this looks good. You're always going to have spikes outside that because of blinking.. but most of it is within 50mV.
You say that a dispersion of around 50uV is acceptable, but in the above figure, the vertical axis scale seems to be more than ±50uV based on 800. To begin with, is it correct that the unit of the vertical axis is uV?
User avatar
James
Site Admin
Posts: 1109
Joined: Wed Jan 02, 2013 9:06 pm

Re: how should I find noise???

Post by James »

You want +-50uV from whatever you baseline value it. Different Muse models have different baseline levels. Yours is 800, so you want your RAW EEG to be min 750uV - and max 850uV.
The Graph scale is automatic. If you give it data from -100 to +9999, that'll be the scale it shows. Your data above is from -200 to +2000. It just auto ranges to the min and max data values available.
Post Reply