Large artifact during button press

Post Reply
UlrichP
Posts: 3
Joined: Thu Apr 23, 2015 4:41 pm

Large artifact during button press

Post by UlrichP »

Hi all,

We encountered a strange issue with our Active 2 system. When using the Biosemi response box, there is a huge low frequency artifact/ drift whenever the button is pressed. We tried 2 differend responseboxes, both produce the artifact (please find a figure attached -note the large scaling on the plot!).
Interestingly, the artifact is not visible in the acquisition software during the recording, but only offline after importing the data ino matlab (might be our particular display settings in Actiview). Also, the artifact is visible on all channels, even the ones that hadn't any electrodes connected.

Another hint on the potential issue is that the trigger returned by the button press is the same that is returned by the F keys on the keybord (316).
Could there be some sort of loop/ interaction, that could be avoided by disableing the trigger input from the (recording computer's) keyboard?

Best regards,
Ulrich
Attachments
buttonpress.png
buttonpress.png (99.3 KiB) Viewed 9679 times

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

Re: Large artifact during button press

Post by Coen »

If the artifact is does not appear in ActiView during recording, then you can be 100% sure that it is also not present in the BDF file. The error is surely caused by your Matlab file reader.

The logical first step in such a case is to check the file with proven file readers such as http://www.biosemi.com/download/ToolsFo ... er-808.zip or http://www.teuniz.net/edfbrowser/

Best regards, Coen (BioSemi)

UlrichP
Posts: 3
Joined: Thu Apr 23, 2015 4:41 pm

Re: Large artifact during button press

Post by UlrichP »

Hi Coen,

Thanks for your reply.
We since looked further into this issue, and it seems that the responsebox produces a spike-like artifact at the time of both the button press and release.
The original image I posted was after lp, hp and notch filtering as well as downsampling of the data, all of which presumably transformed the spike into a different pattern. I have now attached an additional image showing the artifact in the raw data.

So I'm pretty sure that this spike was present already during the recording, but went unobserved as we were looking for the other, much bigger artifact pattern.
That being the case, do you have any ideas on where this spiky artifact during press and release might come from?

Best regards,
Ulrich
Attachments
buttonpress 2.png
buttonpress 2.png (184.82 KiB) Viewed 9666 times

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

Re: Large artifact during button press

Post by Coen »

Test the file with a proven filereader

Test the hardware with electrodes in water.

Are these BioSemi response buttons ?

Best regards, Coen (BioSemi)

UlrichP
Posts: 3
Joined: Thu Apr 23, 2015 4:41 pm

Re: Large artifact during button press

Post by UlrichP »

Hi Coen,

Thanks for your reply.
We found out what caused the artifact: right after importing the data, the script we used rereferences the data to average reference.
This introduced the squarewave signal from the response channel (ERG1) to all other channels.
The reason we did't think of this earlier is that our other triggers are conveyed via the A/D box's parallel port, and are not represented as an actual channel in the data. Thus they did not introduce an artifact.

Best regards,
Ulrich

Post Reply