Trigger display

Post Reply
Erin McMullen
Posts: 3
Joined: Thu Jan 12, 2006 4:12 pm

Trigger display

Post by Erin McMullen »

Hello! I am a brand new user of BioSemi, so please forgive me if I'm asking about something obvious. I'm using a 128-channel setup for an ERP experiment, in conjuction with Presentation. I have a somewhat complicated design involving approximately 200 different stimulus trigger signals. The triggers display properly onscreen during the experiment, which suggests to me that ActiView can read them properly, but then when I try to import the bdf file to an analysis program (BESA), it won't read them intelligibly -- most of them appear as the same, abnormally-large number (65280, if I remember correctly). How should the triggers behave on export? Is there anything special I need to do to get them to display as expected? I have also contacted BESA support since I don't know which end the problem is on.

Any help you can offer would be greatly appreciated! Thanks!

Mosimann
Posts: 10
Joined: Fri Jan 13, 2006 2:13 pm

Post by Mosimann »

Hi, BESA seems to have a problem to handle incoming triggers at the exact start of a recording. So, you can do this to prevent seeing instead of a trigger code 3 a huge number such as 65280: you start the recording and save a few samples and then hit Presentation to start stimulating. This will work, and in BESA you will see the correct codes.

However, a far better solution to the problem is this:
Go to Notepad and start a new text file, and write the simple text as below shown:

65280 Positive

Then save the file as TriggerLogic.txt under the following directory:
C:/Programme Files/BESA5/system/Biosemi .

If you now open your earlier file, that read the wrong code (ie 65280) and not the 3, then press CRTL L and clear database. Once done, and you reopen your file, you should see nicely the trigger codes as recorded with ActiView!
Good luck, and keep well, Marc

Erin McMullen
Posts: 3
Joined: Thu Jan 12, 2006 4:12 pm

Post by Erin McMullen »

Thank you, Mosimann -- that solution worked perfectly! I will share it with my labmates.

Karsten Hoechstetter
Posts: 2
Joined: Fri Nov 12, 2004 12:06 pm
Location: MEGIS Software GmbH

Post by Karsten Hoechstetter »

Dear Erin,

I know you have meanwhile received help from the BESA Support, but here's an explanation for the Forum users: Your problem had nothing to do with triggers appearing at the start of the recording.

Rather, in BioSemi data triggers can be deined in different ways in the trigger channel: In some data, triggers are defined as voltage drop in the trigger channel, in other data as an increase in voltage. Also the baseline level can differ between files. BESA can handle all these situations:

As a default, BESA expects the baseline level to be zero (all 16 trigger bits are off). In your data this is not the case (rather, the left 8 bits are set all the time and bits are set to 1 to define an event). This is what you tell BESA by creating the File TriggerLogic.txt in folder BESA5\System\BioSemi containing the text "65280 Positive". The baseline value 65280 represents the signal of the left 8 bits (65535 - 255 = 65280).

Best,
Karsten

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

Post by Coen »

Unconnected trigger lines are high due to the internal pull-up resistors in the USB receiver. The "baseline correction" described above is a good solution. An alternative is to connect all unused trigger lines to ground (pin 37 on the trigger port).

The 16 trigger bits should not be regarded as an analog voltage level, but rather as a digital code. The 16 bits (lines) can be used fully independent from each other (16 separate trigger signals). Another popular choice is to use 2 groups of 8 bits (two sets of 256 different trigger codes).

The ActiView software is based on the convention that positive logic is used for the triggers (event = high, no event = low). Using negative logic for the triggers may lead to missed triggers when the data is downsampled in software (because an OR function is applied on groups of bits).

Best regards, Coen (BioSemi)

Post Reply