Hello,
We're getting a signal in BCI2000 from the biosemi system using 64 channels but the signal information is all over the place.
We've tried to make it align with the channels by changing the AlignChannels, SourceChOffset, SourceChGain but none of these seem to help.
What are some reasonable settings for AlignChannels, SourceChOffset, SourceChGain in other systems?
As you can't calibrate the biosemi system and gains are internally fixed.
Is there anyone besides the acquisition module contributor (Sam) who have used Biosemi with BCI2000 for research? I've contacted Sam but he hasn't been doing any work with the system for a few years now.
Thanks,
Emily
Aligning brain signals
Hi Juergen,
In the source signal window;
On the vertical axis each channel has a horizontal baseline and the source signal for each channel basically runs along this line. This is how it looks in the mu tutorial with a dummy signal and in other EEG capture programs we run such as actiview.
What we are seeing in the source signal window with BCI2000 with live signal from Biosemi though are no baselines and the source signals are not in line with their channel labels. They are unevenly spaced some bunched together some spaced very far apart. So we cannot tell which signal corresponds to which channel and changing the parameters I mentioned in the first email seems to have no effect.
Thanks,
Emily
In the source signal window;
On the vertical axis each channel has a horizontal baseline and the source signal for each channel basically runs along this line. This is how it looks in the mu tutorial with a dummy signal and in other EEG capture programs we run such as actiview.
What we are seeing in the source signal window with BCI2000 with live signal from Biosemi though are no baselines and the source signals are not in line with their channel labels. They are unevenly spaced some bunched together some spaced very far apart. So we cannot tell which signal corresponds to which channel and changing the parameters I mentioned in the first email seems to have no effect.
Thanks,
Emily
Biosemi
Emily,
It sounds like the signals have a DC offset, which can be perfectly normal if the amplifiers have a low high pass filter. I do not know about Biosemi, but from what you describe, this is the most likely reason. Why don't you right-click on the Source display and select a high-pass of, say, 1 Hz. The signals should then all be aligned around the baseline. Just understand that this procedure is only applied to the display, and not to the data stored or processed.
Gerv
It sounds like the signals have a DC offset, which can be perfectly normal if the amplifiers have a low high pass filter. I do not know about Biosemi, but from what you describe, this is the most likely reason. Why don't you right-click on the Source display and select a high-pass of, say, 1 Hz. The signals should then all be aligned around the baseline. Just understand that this procedure is only applied to the display, and not to the data stored or processed.
Gerv
Emily,
the signal display assumes that the SourceChGain and SourceChOffset values are set such that, after their application to the signal, signal values are calibrated in muV, and their offset is zero.
If your signal offsets are very large, you might try setting all SourceChOffset entries to 32768, in case the Biosemi system represents data as unsigned integers.
The SourceChGain entry corresponds to the Biosemi system's voltage resolution, which is usually given in the manual, typically in units such as "muV per least significant bit (LSB)".
For a 16-bit EEG system with an input range of 1mV, a typical value for SourceChOffset would be 0.03 muV per AD unit.
HTH,
Juergen
the signal display assumes that the SourceChGain and SourceChOffset values are set such that, after their application to the signal, signal values are calibrated in muV, and their offset is zero.
If your signal offsets are very large, you might try setting all SourceChOffset entries to 32768, in case the Biosemi system represents data as unsigned integers.
The SourceChGain entry corresponds to the Biosemi system's voltage resolution, which is usually given in the manual, typically in units such as "muV per least significant bit (LSB)".
For a 16-bit EEG system with an input range of 1mV, a typical value for SourceChOffset would be 0.03 muV per AD unit.
HTH,
Juergen
Emily,
currently, filter settings for display windows are not stored permanently. This is intentional, to avoid users overlooking bad signal quality when filtering is activated permanently. You should always try to get a reasonable signal without filtering, and then use filtering to suppress artifacts that you cannot avoid.
Actually, your amplifier should send zero-mean data. If it doesn't, try 32768 or 16384 as SourceChOffset entries.
Regards,
Juergen
currently, filter settings for display windows are not stored permanently. This is intentional, to avoid users overlooking bad signal quality when filtering is activated permanently. You should always try to get a reasonable signal without filtering, and then use filtering to suppress artifacts that you cannot avoid.
Actually, your amplifier should send zero-mean data. If it doesn't, try 32768 or 16384 as SourceChOffset entries.
Regards,
Juergen
Who is online
Users browsing this forum: No registered users and 4 guests