Strange synchronous ECG and EEG artifact

ytersyters Maryland, USA
edited November 2019 in General Discussion

Hi, BCI newbie here. Watching videos on youtube I expected to see a fairly homogenous signal from the EEG. For example, this video shows a fairly steady frequency domain such that the wearer can clearly demonstrate alpha waves:

However, when both my wife and I wear the Mark IV headset, we see consistent synchronous pulses that come from all electrodes. The frequency varies slightly, and can change significantly seemingly based on activity. Consequently, the frequency domain goes all over the place.

To see if this was somehow related to my cardiac rhythm, I also attached two skin tags to my biceps and grounded on my elbow to generate ECG data. I got exactly the same, synchronized signal from the ECG. However, the signal is not clearly paired to my heart rate. At various times and levels of activity I measure 2, 4, 6, 8 or 11 heart beats per pulse.

The other weird thing is that not only is the big pulse synchronized across electrodes, but the general shape of the overall signal is synchronized, with some variance. This makes me think somehow a signal is bleeding across all channels. But when I remove my headset, leaving the ECG skin tags on, all head electrodes are wired, and the ECG continues to receive the same signal.

You can see the ECG and EEG data here. Channel #1 is the ECG data, and the rest are EEG. Any idea what could be causing these artifacts, and perhaps how to get rid of them to get a clear signal? Did I maybe setup my headset incorrectly and get wires crossed? Thanks for any assistance you can provide!

Also, for reference, this is how the electrodes are distributed across my head (ignore position #1).

Comments

  • ytersyters Maryland, USA

    I've resolved this issue. It was due to the headshot widget in the OpenBCI GUI. The widget took up too much CPU leading to packet latency and the pulse pattern. https://github.com/OpenBCI/OpenBCI_GUI/issues/349

    Once I turned off the widget the pulse artifact disappeared and I got clean brainwave signals.

  • retiututretiutut Louisiana, USA
    edited November 2019

    @yters It's true that the current GUI headPlot is very greedy with resources. We are working on solutions for GUI 5.0 to improve frame rates, specifically in TimeSeries, Accelerometer, and HeadPlot Widgets.

Sign In or Register to comment.