No reference or average reference?

Post Reply
nualab
Posts: 1
Joined: Fri Oct 21, 2005 11:39 am

No reference or average reference?

Post by nualab »

Hi,

We are recording EEG using the a 64-channel Biosemi setup. We were
advised in our initial training session to choose 'no reference' (where display is wrt CMS) during recording and to then re-reference during data analysis - we are using BESA for analysis, and to date the data look good
when re-referenced to the average. This is a face perception experiment
and the ERPs show appropriate P110 and N170. I am confused on two issues:

1. The Biosemi web site FAQ on CMS and DRL suggests that 'no reference'
should only be selected 'as a quick check on electrodes' as it does not provide the full CMMR - can someone explain this? Does it mean we would get better signals after selecting the average reference in ActiView?
To date we have great recordings from short haired male subjects but
pretty poor recordings from females with long hair?

2. Is it wise to select the average reference when recording? e.g., if one
channel is faulty and turned off during recording, we now interploate this
channel in BESA and then use the average reference when averaging to
get the ERP. If a channel was turned off during recording, would this not
effect the definition of the average reference?

thanks in advance for any help
Nuala

Coen
Site Admin
Posts: 1124
Joined: Fri Mar 26, 2004 7:00 pm
Location: Amsterdam, Netherlands
Contact:

Post by Coen »

1) It is important to realize that the choice of reference in ActiView only affect the displayed signals; the signals saved to file are always unreferenced. Display of unreferenced signals can be useful to spot problem electrodes. However, the unreferenced generally still show significant 50Hz interference because of the limited CMRR. In order to achieve full CMRR, the signals have to be displayed with respect to a reference. It is not necessary to try to achieve clean signals in unreferenced form. So long as the signals in referenced form are satisfactorily, all is well.

2) Turning channels off during recording does not affect the channels to file, unless the option "save monopolar displayed channels" is selected in the file dialog.

Note: only the controls in the right sidebar affect the BDF file. Controls in the left sidebar only affect the dsiplay, they do not affect the BDF file. There is one exception: the above mentioned option to save monopolar displayed chanels.

Best regards, Coen (BioSemi)

John Lumsden
Posts: 2
Joined: Fri Jul 09, 2004 1:55 pm
Location: West London MH Trust

reference or not reference

Post by John Lumsden »

Surely this is an interaction with what you include in your saved channels. If you require a channel for your analysis it is important that you include this in your collection parameters. This is not necessarily the case if you use monopolar displayed channels option and do not additionally save the channel that you have used for the reference in the collection display.

My position is that for my pilot study, I have referenced to A2 (right mastoid), and have assumed that the signals recorded are referenced to this channel. As I have saved "monopolar displayed channels" rather than the full set of electrodes, I assume that I have no record of the reference on its own (A2- right mastoid). My subsequent analysis in Brainvision Analyzer does not display the reference (as I haven't saved it). Surely since it is displayed as a reference in the monopolar displayed channels it should also be saved. Or at least this should be abit more explicit in the manual or on the FAQ on the website.

John

Coen
Site Admin
Posts: 1124
Joined: Fri Mar 26, 2004 7:00 pm
Location: Amsterdam, Netherlands
Contact:

Post by Coen »

Dear John,

On March 11th, 2004, I have sent the following personal email to to you:

[quote]To record 32 pin and 3 EX channels, proceed as follows:

1) In the monopolar tabpage, select channel selection "free choice"
2) In the channels list, select channels 1-32 and EX1, EX2 and EX3, there are now 35 channels dsiplayed in the monopolar display
3) In the file dialog, choose the save subset "save as monopolar displayed"
4) Check that the number of saved channels indicator in the right sidebar displayes "35+1" (32 pin, 3 flat, plus trigger channel)

You can save these settings in a (new) CFG file by going to the "about ActiView" tabpage, and saving the configuration[/quote]

I would think that this old email explains the procedure quite explicit.

If some of your recordings have the mastoid electrodes missing from the file (check with BDFreader from http://www.biosemi.com/download/BDFReader.zip), you can take the average between T7, T8, P7 and P8 as a reference. The difference with a real mastoid reference would be insignificant for most experiments.

Best regards, Coen (BioSemi)

Post Reply