Forum

Nobuyoshi · Joined: Jan 15, 2019 Posts: 21

Hi There,

EEG could be recorded clearly when I use Cyton board via the openbci element( FTDISerialPort > F_OpenBCI ...
).

However, with the Ganglion board (BlueGiga Bluetooth 4.0 Smart dongle> ArduinoSerialPort > F_Ganglion..), I could not record it correctly.
EEG amplitudes are extremely small.

Are there any solution?

Best regards,
Nobu

jarek · Joined: Oct 22, 2007 Posts: 1058

I am not sure what can cause that.

When you rescale the signal (multiply by a larger number, like 10 or 1000) do you see a signal which looks like EEG
or is it just noise?

Nobuyoshi · Joined: Jan 15, 2019 Posts: 21

Thank you Jarek,

When I monitored the EEG on OpenBCI GUI using same device (USB dongle, Ganglion board and laptop), EEG were normal. On Bioera, it looked like EEG but extremely tiny.

I will confirm it again and report to you!

Nobu

Nobuyoshi · Joined: Jan 15, 2019 Posts: 21

Hi there,

Thank you for your support.

I have just finished the verification
I belive the signals were EEG.

1. Firstly, I will show you the data when I recorded with OpenBCI GUI using the same device.
> normal

https://drive.google.com/file/d/1_H8RydyLmxUYc38jLOvx87wnRc1mAWcb/view?usp=sharing

2. Next is the data with Bioera.
> tiny

https://drive.google.com/file/d/1tDBdBeUObOAwhBYJAK8niFjS7i4vR1rc/view?usp=sharing

3. Final data are above data at 4-fold magnification
>safe to say EEG

https://drive.google.com/file/d/1KUk_9EySJ7JCBrYheZnm2cGeG0CaV0oZ/view?usp=sharing

Regards
Nobu

jarek · Joined: Oct 22, 2007 Posts: 1058

Thanks, I will try to compare the signal amplitudes with my Ganglion I have here. It could be a matter of proper scaling.

It has been months since I used the OpenBCI GUI, but I don't seem to recognize it on the first picture you show. Are there only BioEra charts there? Or maybe you recorded data with OpenBCI app and imported that into BioEra?

Nobuyoshi · Joined: Jan 15, 2019 Posts: 21

Hi Jarek,

Thank you for your prompt reply!
Yes, I recorded the data with OpenBCI GUI, then imported the ascii data into BioEra via "F_Excell file reader"element.

Nobu

jarek · Joined: Oct 22, 2007 Posts: 1058

It took awhile but I finally got to it. Unfortunately I did not find a problem with my ganglion's amplitude at this moment.

I used this snippet to test:
http://proatech.com/design/ganglion_float.bpd

I attach picture of what I get from BioEra. The RMS amplitude is about 2.5uV (unconnected inputs), which is about the same I see on the OpenBCI_GUI application with bled112 dongle and OpenBCI Hub.

http://proatech.com/photo/ganglion_float.png

Maybe you could try this snippet (which looks very similar to yours, maybe it is even exactly the same) and tell me where you see the difference and what are your RMS values when no input electrode is connected.

Nobuyoshi · Joined: Jan 15, 2019 Posts: 21

Hi Jarek,

Thank you very much for confirmation.

Yes, under unconnected inputs condition, the RMS amplitudes on Bioera(unconnected inputs) are similar to onese on the OpenBCI_GUI application, but Bioera's amplitudes seem to be a bit small.

The differences become more prominent when the EEG are recorded.

Here are my results!

https://drive.google.com/file/d/1gPwbpTqyO-mrhfFPgUNrs7rVwalJwkC1/view?usp=sharing

Nobuyoshi · Joined: Jan 15, 2019 Posts: 21

In addition, we shuold consider the difference between the calculation formula for RMS adopted by BioEra and one incorporated in OpenBCI GUI ?

> OpenBCI GUI

　[William'comments]
　https://openbci.com/index.php/forum/#/discussion/comment/8021

>BioeEra

　[TimeTransform]
　SquareRoot(Mean(Sum of Squares))

Raw signals recoreded by each program are different.

　https://drive.google.com/file/d/1gPznLOx2ltjizK8eC9GBG64kBPXc2Vxa/view?usp=sharing

Thanks,
Nobu

jarek · Joined: Oct 22, 2007 Posts: 1058

Please try version update 4.066.

I removed the /4 rescale which was there I guess to match the RMS amplitudes more closely. Now, without this rescale, this is not the case any more, but the RAW values seem to match, which is probably the right way to compare.

Nobuyoshi · Joined: Jan 15, 2019 Posts: 21

Hi Jarek,

I am deeply grateful for your support.

I checked it again. The calculation formula used by BioEra is applied to the recorded data using the OpneBCI GUI. The comparison results are shown below.

RMS values (average)
[BioEra]
input1 input2 input3 input4
8.95 9.45 9.39 11.39

[OpneBCI GUI]
input1 input2 input3 input4
11.74 12.64 12.66 15.78

Each value has a similar value.
However, it is still found that the values recorded by OpneBCI GUI are about 1.3 times larger.

Nobu

jarek · Joined: Oct 22, 2007 Posts: 1058

The RMS calculation takes a time period. Which means that in order to compare apples to apples we need to use the same time periods. I my snippet I set it to 1 second, but it can be anything. I don't know what it is set to in the OpenBCI GUI, which could be the reason of the difference. Also it is a question of RMS calculation, we would need to make sure we calculate it exactly the same.

So I think it is better to compare only RAW values.

If it is still 1.3 times different, I can't explain it at this moment. I need to build or buy a microvolt generator with a known amplitude.

Nobuyoshi · Joined: Jan 15, 2019 Posts: 21

Thank you for your reply.

Yes, setting the time period is an important parameter for calculating RMS.
Each values of RMS presented by me is one calculated by setting the time period to 1 second.

However, Bioera's data and OpenBCI's data were not simultaneously measured (though each was measured immediately in the same enviroment), so we can not simply compare.

I agree your suggestion.
Unfortunately, I don't have a microvolt generator, but I will try to examine another method and report to you.

Nobu

Nobuyoshi · Joined: Jan 15, 2019 Posts: 21

Hi Jarek,

We can send '[' to generate a fake square wave signal.

http://docs.openbci.com/OpenBCI%20Software/06-OpenBCI_Ganglion_SDK

Using this command, I compared the raw signals recorded on each program.

https://drive.google.com/file/d/1woUdsZL_sgs1HqA3fCWycCzsU0oKM3AJ/view?usp=sharing

It tells us that the amplitude recorded by openbci is about 1.44 times (14.96/10.39) larger than the signal recorded by Bioera.

https://drive.google.com/file/d/1xZFiGHULM0Ivb7YF8hnFfH_Ns2mhMvvL/view?usp=sharing

Thanks,
Nobu

jarek · Joined: Oct 22, 2007 Posts: 1058

Thanks Nobu. Good find.

I created BioEra version 4.067 with the 1.44 amplitude multiplier. I have not seen such multiplier anywhere in the OpenBCI documentation or source code. But your method of comparing makes sense, so let's hope it is correct.

Nobuyoshi · Joined: Jan 15, 2019 Posts: 21

Thank you Jarek!

Using v4.07, I tried to compare them again.
The values recorded by both programs are exactly the same value.

https://drive.google.com/file/d/1bILHD_Evd3KaW01RRRiYYELEsoIf0dDt/view?usp=sharing

I deeply appreciate you have comfirmed and made an adjustment with promptly.

Nobu

hiimpaul · Joined: Oct 25, 2025 Posts: 3

Hello, I tried the ganglion_float design with my Ganglion and got around 11 uVrms. From my limited understanding of what is going on in this thread, that is a normal reading correct??

https://imgur.com/a/DNZfLL0

jarek · Joined: Oct 22, 2007 Posts: 1058

It is probably correct, depending what exactly you are measuring.