How to remove rhythmic spiking in EEG signal?
I have been having a strange spike occur in my data every couple seconds that I don't understand. Here is a picture of what it looks like in the GUI with only the ear clips attached:
How can I stop these spikes from happening?
What I have tried:
-searching the forum
-changing operating systems and computers: Linux and Windows
-changing the number of connected sensors
-updating the firmware to the latest version
I also played around with many parameters in the driver properties on Windows, though the parameters did not seem consistent with the tutorial for the latency timer at https://docs.openbci.com/docs/10Troubleshooting/FTDI_Fix_Windows
Finally, I went in to the registry to alter the latency timer there and set the value to 1.
I have attached a csv file of a run of the spikes with only the reference ear clips attached to the board.
Comments
DMan, hi.
All EEG requires THREE connections to the scalp / head:
If the channel being measured on the scalp is floating / not connected -- then much ambient EMF noise in the room will be picked up. The EEG amplifier is measuring a DIFFERENTIAL voltage between channel and reference.
Regards, William
https://docs.openbci.com/docs/01GettingStarted/01-Boards/GanglionGS
https://docs.openbci.com/docs/01GettingStarted/02-Biosensing-Setups/EEGSetup
Thanks for responding so quickly. I wish I had more time to work with this.
I hooked up one electrode and noticed that the noise went away for a while. I was able to see eye blinks on the GUI. Then the spikes came back. So if I understand correctly that is because a connection is disappearing somewhere?
I played around with it for a while, and it seems like if I hold the ref and D_G wires at a certain angle then the spikes go away and I can detect eye blinks again. I am not sure if this is helpful because it seems like there isn't a consistent angle.
I am sorry if these are beginner questions.
How can I improve my set up so that I don't have to hold wires down in order to get meaningful data?
What electrodes are you using? How are they fastened to the scalp? Did you complete all the tutorial steps in the links previously shown?
https://docs.openbci.com/docs/01GettingStarted/01-Boards/GanglionGS
https://docs.openbci.com/docs/01GettingStarted/02-Biosensing-Setups/EEGSetup
I am using the spikey electrodes from the headband kit that were sent to me as well as the ear electrodes.
I did complete that tutorial but the results were the same.
I fasten them to the scalp by lifting the hair and then tightening the headband. I noticed better results when I placed them on the bare skin of my forehead.
So you've never been able to get the results shown in the tutorial? When did you purchase the Ganglion and headband kit? Do you have any other types of electrodes there, such as gold cups and paste?
I purchased the headband kit and ganglion two years ago.
I think I also have solid gel emg sticky electrodes.
I don't have gold cups and paste.
Did the EMG tutorial work? If you did not try it, please give it a try.
https://docs.openbci.com/docs/01GettingStarted/02-Biosensing-Setups/EMGSetup
When you purchased two years ago, did any of the tutorial steps work for you?
Also have you seen this tutorial on the headband?
https://docs.openbci.com/docs/04AddOns/01-Headwear/HeadBand
I have gone through the tutorial on the headband.
I also have gone through https://docs.openbci.com/docs/10Troubleshooting/minimizingNoise.
I just tried the emg for the first time. All my other runs were on eeg. On the emg it was easy to see when the subject was flexing or relaxed. That was clearly working. I was still getting the spikes though. I thought that might be heartbeat, but it didn't seem to match the pulse in timing or bpm.
Here are some screenshots with two channels set up:
While flexing:
While relaxed:
I notice that you do not have your notch filter in the GUI, turned on. Your images show huge amount of 60 Hz mains noise.
Another verification test you can try is ECG. All of the EEG, ECG, EMG usage requires the notch filter.
https://docs.openbci.com/docs/01GettingStarted/02-Biosensing-Setups/ECGSetup
I'm having the same issue with the OpenBCI Headband & Ganglion board. I'm trying to measure EEG when blinking and doing a few other facial gestures.
I have the notch filter on 60Hz and BP on 7-13hz. I also make sure the electrodes ade very tight and I use the Ten20 electrode paste.
Did anyone ever figure out how to fix this problem?
@sng, hi.
If you are seeing noise in your OpenBCI board, see these suggestions:
https://docs.openbci.com/docs/10Troubleshooting/minimizingNoise
Your noise spikes are most likely from some equipment in your area. So trying an experiment in another room / building may result in less noise. Also see previous comments in this thread.
William
I tried notch filters and all the troubleshooting tips but I only got it to work for ecg and emg. I couldn't ever get rid of the spikes for eeg.
Instead I just went and bought a cyton board and that does great for eeg.
If you are seeing spikes with Ganglion EEG and the dry headband electrodes, then that is due to the less than optimal contact and skin impedance of the dry headband electrodes. Cyton has a higher amplifier input impedance, so is more forgiving of poor skin impedance.
But the solution to your Ganglion EEG spikes, may be to simply use better electrodes: cups and paste, cap with gel, or any system which uses saline soaked pads, such as the MindAffect band.
William
I've heard that IDUN makes some spiky gel electrodes that are great for going in hair. I'd love to buy some but I can't find any.
https://shop.openbci.com/collections/frontpage/products/idun-dryode-kit?variant=31081774874696
If you have cup electrodes and paste, those should give equally good results as far as noise reduction.