[Eeglablist] Physical filter Circuits

Ross Fulham ross.fulham at newcastle.edu.au
Mon Dec 8 16:21:29 PST 2014

I can't imagine why you would wish to build your own EEG amp when there are a range of good quality commercial systems available. However, if you are seeking general advice on EEG amplifier design try the Open EEG project:


Note: if you are building your own EEG amplifier, then it is critical that you meet safety requirements for bio-medical instrumentation for human use. There are limits on leakage current, isolation, etc that must be meet.

Dealing with 50Hz noise should in the first instance be about eliminating it from the raw signal, rather than filtering it out. For this you need either a high quality differential amplifier, or an active electrode system, with very high input impedances. The presence of visible 50 Hz noise in the raw signal usually implies poor electrode impedances, grounding, etc, in which case simply filtering out the 50Hz may not be enough to obtain good EEG signals.

Filtering of 50 Hz online can be an issue, depending upon application, as it introduces phase shifts and/or latency delays into the signal.  However, if you are only interested in signal up to ~20Hz, then a basic notch filter is unlikely  to be an issue.

Ross Fulham.

From: eeglablist-bounces at sccn.ucsd.edu <eeglablist-bounces at sccn.ucsd.edu> on behalf of Martijn De Neeling <martijn.dene at gmail.com>
Sent: 06 December 2014 06:11
To: eeglablist at sccn.ucsd.edu
Subject: [Eeglablist] Physical filter Circuits

Hi All,

I am fairly new to BCI and I am trying to make an EEG signal amplifier. I do not know if this is the correct mailing list for my questions, but I will ask them anyways. What is the optimal amplifier setup for filtering out the 50 Hz frequency while keeping good EEG signals? Note: I am interested in measuring the motoric signals ranging from 7.5 to ~20 Hz.
At the moment I use this setup:

Pulse->electrodes->Industrial amplifier->Notch filer 50 Hz->Low-pass filter 32 Hz->Amplifier with variable gain of 60-440-> another 50 Hz Notch filter(optional)->AD-converter->computer.

I have read that it is also important to have high pass filters for noise below 7.5 Hz. Does anyone recommend this? Also, are physical filters good for pre-filtering the data or is it better to use digital filters?

I hope someone can give me some clear answers on these questions.
Thanks in advance,

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://sccn.ucsd.edu/pipermail/eeglablist/attachments/20141209/b75f56d2/attachment.html>

More information about the eeglablist mailing list