Emotiv EPOC recorded values (inverted or not?)

About Brain Computer Interfaces, EEG, neurofeedback, related event announcements, jobs, etc.
Post Reply
lbatistic
Posts: 5
Joined: Tue Aug 09, 2016 1:21 pm

Emotiv EPOC recorded values (inverted or not?)

Post by lbatistic »

Hello people,
I would like to know how does EMOTIV Epoc store EEG data, positive as positive or positive as negative? I've read on multiple articles that sometimes EEG data is plotted in a way where positive voltages are below x axis (on the side where usually negative voltages are plotted) and negative voltages are plotted above x axis (where usually positive voltages are plotted). You can see an example in the photo below:

Image


Is there a reason that some people plot EEG signal values inverted? How can I see if certain signal is inverted or not, for eg. I use eeg signal recorded on EMOTIV Epoc in eeglab and I just plot it, how can I tell if it has been inverted or not (are positive values actually positive or entire signal has been multiplied by -1)? Also, does eeglab invert signals?
Cheers,
Luka

jtlindgren
Posts: 775
Joined: Tue Dec 04, 2012 3:53 pm
Location: INRIA Rennes, FRANCE

Re: Emotiv EPOC recorded values (inverted or not?)

Post by jtlindgren »

Hi Luka,

you could try to average the trials (EEGLABs study functions should allow you to do an ERP average plot) and then notice which direction the average bump bumps. Is this the same direction as the P300 bumps seem to bump in the literature? The result also depends on the choice of the reference electrode, and the fact that the amp is a differential amplifier (i.e. potential of a signal channel is actually a potential difference between the corresponding electrode and some reference electrode). Although the sign may be significant for ERP and brain activitiy analysis, for typical BCI purposes the sign doesn't matter: the usual machine learning techniques such as P300 classification used in openvibe do not really care if a global sign has been flipped or not, as long as this flip is consistent and doesn't change during the experiment.

Afaik openvibe shouldn't flip the sign (unless there is a bug), and if Emotiv does, that should be a question posed to the company making the device and the sdk.


Cheers,
Jussi

seanypoo
Posts: 7
Joined: Wed Feb 22, 2017 5:55 pm

Re: Emotiv EPOC recorded values (inverted or not?)

Post by seanypoo »

Hi

I am reading an old Emotiv research paper http://neurofeedback.visaduma.info/EmotivResearch.pdf The paper use openvibe p300 speller and Epoc to detect P300.

The P300 component is showing negative results in this research paper. I screenshot one figure here below

Image

P300 should be positive voltage. why the results shown negative voltage?
jtlindgren wrote: Afaik openvibe shouldn't flip the sign (unless there is a bug), and if Emotiv does, that should be a question posed to the company making the device and the sdk.
I didn't see any post on Emotiv company talks about sign.

I presume openvibe test the Epoc as well, do you also get negative voltage when measure EEG (or P300)?

Thanks
Sean

jtlindgren
Posts: 775
Joined: Tue Dec 04, 2012 3:53 pm
Location: INRIA Rennes, FRANCE

Re: Emotiv EPOC recorded values (inverted or not?)

Post by jtlindgren »

Hello,

the Emotiv EPOC driver gets data from the device by the call

Code: Select all

					m_ui32EDK_LastErrorCode = IEE_DataGet(m_tDataHandle, g_ChannelList[i], l_pBuffer, l_ui32nSamplesTaken);
What data that call returns in the 'l_pBuffer' is totally internal to Emotiv's proprietary SDK, we cannot inspect that. On OV side (checked OV 2.0), the Acquisition Server does not flip the signal sign, and neither does Designer unless you use some box that has such an effect. By default the data is passed as it is received from the above call.

Regarding the paper, its best to try to directly contact the authors who wrote it.


Best,
Jussi

Post Reply