Emotiv Epoc, Highpass filter issue & problem on Cursor task

[tawilliams]

Hi Everyone,

I'm new to BCI research and I am having a difficulty when running a Mu/beta session with the cursor task.

I have the Emotiv Epoc, running on an Vista operation system. I compiled the SVN and am currently using the 3.0.1 BCI Launcher. I have also followed the instructions with respect to the parameters as described here: http://www.bci2000.org/wiki/index.php/C ... ons:Emotiv

However when I launch the Emotiv Signal source, ARSignalProcessing and either the StimulusPresentation or cursor task, I get a reminder to set the high pass filter,the error is as follows:

EmotivADC::Preflight: Set HighPassFilter to 0.1Hz to filter out DCOffset.

I then set the HighPassFilter to this setting and I get the following error:

Warning: SourceFilter::Initialize: High pass corner frequency is outside sampling bandwidth. No filtering will be performed.

Can anyone explain what I am failing to do, or is this a bug?

On another point, a problem exists when I run the Cursor Task, I presume this is because there are no filters set relating to the other issue. But, the ball just keeps entering the screen and then falling to the bottom of the screen, there seems to be no data transmission with respect to the filtering.. Is this a result of the same issue described previously, or am I doing something wrong here as well.

I set the full matrix spactial filters according to that described and then used a large laplacian filter set up, around the sensors of interest., and then changed the classifier accordingly? Any suggestions greatly appreciated.......

Thanks a million for your help.

Tiernan

[mellinger]

Hi Tiernan,

Can anyone explain what I am failing to do, or is this a bug?

you were hit by a bug. It's been fixed in rev. 3378, so when you do an SVN update, and recompile, the HighPassFilter should work for you.

Your problem with the CursorTask could be related to the filtering problem. Proper setup of cursor feedback is pretty complicated, so you should first follow the tutorial on the wiki to get it going, then try your own modifications. The mu rhythm tutorial is provided at
http://www.bci2000.org/wiki/index.php/U ... I_Tutorial

Best regards, and thanks for reporting the bug,
Juergen

[tawilliams]

Yes you are right, the new build does not generate the error mentioned, however now when run BCI I get the following issues in preflight:

2011-07-05T19:31:16 - Warning: StimulusPresentationTask::Preflight: Due to a sample block duration of 62.5ms, the actual value of ISIMinDuration for stimulus 1 will be 62.5ms rather than 80ms.
2011-07-05T19:31:16 - Warning: StimulusPresentationTask::Preflight: Due to a sample block duration of 62.5ms, the actual value of ISIMaxDuration for stimulus 1 will be 62.5ms rather than 80ms.
2011-07-05T19:31:16 - Warning: StimulusPresentationTask::Preflight: Due to a sample block duration of 62.5ms, the actual value of StimulusDuration for stimulus 1 will be 0ms rather than 40ms.
2011-07-05T19:31:16 - Warning: StimulusPresentationTask::Preflight: Due to a sample block duration of 62.5ms, the actual value of ISIMinDuration for stimulus 1 will be 62.5ms rather than 80ms.
2011-07-05T19:31:16 - Warning: StimulusPresentationTask::Preflight: Due to a sample block duration of 62.5ms, the actual value of ISIMaxDuration for stimulus 1 will be 62.5ms rather than 80ms.
2011-07-05T19:31:16 - Warning: StimulusPresentationTask::Preflight: Due to a sample block duration of 62.5ms, the actual value of StimulusDuration for stimulus 2 will be 0ms rather than 40ms.
2011-07-05T19:31:16 - Warning: StimulusPresentationTask::Preflight: Due to a sample block duration of 62.5ms, the actual value of ISIMinDuration for stimulus 2 will be 62.5ms rather than 80ms.
2011-07-05T19:31:16 - Warning: StimulusPresentationTask::Preflight: Due to a sample block duration of 62.5ms, the actual value of ISIMaxDuration for stimulus 2 will be 62.5ms rather than 80ms.
2011-07-05T19:31:16 - Warning: StimulusPresentationTask::Preflight: Due to a sample block duration of 62.5ms, the actual value of StimulusDuration for stimulus 3 will be 0ms rather than 40ms.
2011-07-05T19:31:16 - Warning: StimulusPresentationTask::Preflight: Due to a sample block duration of 62.5ms, the actual value of ISIMinDuration for stimulus 3 will be 62.5ms rather than 80ms.
2011-07-05T19:31:16 - Warning: StimulusPresentationTask::Preflight: Due to a sample block duration of 62.5ms, the actual value of ISIMaxDuration for stimulus 3 will be 62.5ms rather than 80ms.
2011-07-05T19:31:16 - Warning: StimulusPresentationTask::Preflight: Due to a sample block duration of 62.5ms, the actual value of StimulusDuration for stimulus 4 will be 0ms rather than 40ms.
2011-07-05T19:31:16 - Warning: StimulusPresentationTask::Preflight: Due to a sample block duration of 62.5ms, the actual value of ISIMinDuration for stimulus 4 will be 62.5ms rather than 80ms.
2011-07-05T19:31:16 - Warning: StimulusPresentationTask::Preflight: Due to a sample block duration of 62.5ms, the actual value of ISIMaxDuration for stimulus 4 will be 62.5ms rather than 80ms.
2011-07-05T19:31:16 - Warning: StimulusPresentationTask::Preflight: Due to a sample block duration of 62.5ms, the actual value of StimulusDuration for stimulus 5 will be 0ms rather than 40ms.
2011-07-05T19:31:16 - Warning: StimulusPresentationTask::Preflight: Due to a sample block duration of 62.5ms, the actual value of ISIMinDuration for stimulus 5 will be 62.5ms rather than 80ms.
2011-07-05T19:31:16 - Warning: StimulusPresentationTask::Preflight: Due to a sample block duration of 62.5ms, the actual value of ISIMaxDuration for stimulus 5 will be 62.5ms rather than 80ms.
2011-07-05T19:31:16 - Warning: StimulusPresentationTask::Preflight: Due to a sample block duration of 62.5ms, the actual value of StimulusDuration for stimulus 6 will be 0ms rather than 40ms.
2011-07-05T19:31:16 - Warning: StimulusPresentationTask::Preflight: Due to a sample block duration of 62.5ms, the actual value of ISIMinDuration for stimulus 6 will be 62.5ms rather than 80ms.
2011-07-05T19:31:16 - Warning: StimulusPresentationTask::Preflight: Due to a sample block duration of 62.5ms, the actual value of ISIMaxDuration for stimulus 6 will be 62.5ms rather than 80ms.
2011-07-05T19:31:16 - GazeMonitorFilter::Preflight: State "EyetrackerLeftEyeValidity" is inaccessible.
2011-07-05T19:31:16 - GazeMonitorFilter::Preflight: State "EyetrackerRightEyeValidity" is inaccessible.

Is this another bug or misconfiguration?

[mellinger]

The warning messages tell you that the durations set for the presentation of stimuli cannot be kept exactly because the BCI2000 system's timing accuracy is given by the duration of a sample block.
You may do one of the following:
* Ignore the warning messages if you are fine with the reported values (which I suppose you are not because actual stimulus duration is 0 for all stimuli, so no stimulus will be presented).
* Change the entries in Application->Stimuli to multiples of 62.5ms.
* Change sample block size via the Source->SampleBlockSize parameter. There, sample block size is specified as the number of samples per block, not in ms. Note that the Emotiv hardware might impose restrictions as to what values of SampleBlockSize make sense; when you get an uneven curve for "block" in the BCI2000 timing window, you need to increase sample block size in steps of one sample until the "block" curve is smooth.

The two error messages at the end tell you that the GazeMonitorFilter depends on system state variables which are only present when eyetracker logging is enabled. To fix this, remove the file CMakeCache.txt from BCI2000/build, run the appropriate CMake batch file again, and say "no" when it asks whether you want to include the GazeMonitorFilter. Then, do a rebuild of your Application module.

Regards,
Juergen

[tawilliams]

Hi Juergen

Thanks for your advice, but the filter issue still exists in the new build and the timing issue with the StimulusPresentation module is now an additional problem. BCI2000 is not running in an effective manner to ascertain meaningful or coherent data in the latest build, for the Emotiv Epoc device.

As part of my research, I am attempting to examine sensors C5, and C6 in a reverse placement of the Emotiv Epoc EGG device to attempt to obtain data from the sensory motor cortex during a mu/beta session using BCI2000. I obviously need the use of the StimulusPresentation Module to preprocess and classify meaningful information from the the raw signal. So, I am currently using the BCI2000 Offline analysis tool to formulate a spectral analysis using a CAR filter. The problem exists for me in two areas with the new build.

1. StimulusPresentation: No stimuli are being presented due to the new problem, the timing issue.
2. The Filtering Issue: No filtering means that regardless of what I see in the Offline Analysis Tool, using a CAR, I can not then use the spatial filter CAR in the CursorTask, as the filtering is not being preformed in BCI2000. This means that I may get a good classification matrix, but information obtained brom the offline analysis is useless in a Mu Beta Cursor task because all that is being used is a raw signal devoid of spatial filters such as the CAR. This is the case if the filters are not being used in BCI2000

Is my interpretation of this sitation correct? ...

Thanks again for your help Juergen,

Kind Regards,

Tiernan

[mellinger]

ad 1:
You should be able to fix stimulus presentation using the second or third suggestion from my previous post.

ad 2:
The ARSignalProcessing module contains a spatial filter, which may be configured to use CAR. Just go to Filtering->SpatialFilterType, and choose "CAR".

Hope this helps,
Juergen