Q. real-time EEG to categorize human emotions using Enobio3G

General information about brain-computer interfaces
Locked
aasedeeq
Posts: 4
Joined: 05 Jun 2013, 09:00

Q. real-time EEG to categorize human emotions using Enobio3G

Post by aasedeeq » 24 Jul 2013, 10:45

Dear Friends,
Hope you all doing well.
I'm quite new to BCI and BCI2000, so apologies in advance if my queries sounds naive.
I'm trying to create a real-time EEG system to categorize human emotions( i.e. Fear, anger...etc) using Enobio3G device and a audio-video stimuli.
My questions are:
1- is BCI2000 is the most suitable software or there is other.
2- In BCI2000, I got the signal source for Enobio3G (Enobio3G.exe) with 8 channels, can I increase it to 20? I wonder if the ARSignalProcessing.exe is the most suitable Signal Processing module? if not, what are my options for the Signal Processing module apart from ARSignalProcessing.exe? and how can I get the source code for any module to modify its filters or create a new one?
3- The input and the output of the signal processor module sounds ambiguous, regarding how to create an Application and connect the Signal Processing Module to it.

Any suggestion will help.

Thank you very much for your patience.
Looking forward to hearing from you.
Regards

boulay
Posts: 382
Joined: 25 Dec 2011, 21:14

Re: Q. real-time EEG to categorize human emotions using Enob

Post by boulay » 26 Jul 2013, 05:55

Hi,

Do you require online classification of the EEG to provide real-time feedback of the emotional state to the user or the experimenter? For example, do you want to change the stimulus to something more calming if the BCI system detects too much anger? If so, BCI2000 is probably the best way for you to achieve that but you could also look into OpenVibe, Fieldtrip, or BCILab+LabStreamingLayer.

Based on a simple Google search it seems as though emotion classification is not so simple. It is likely you will have to first collect a lot of training data (i.e., without meaningful feedback) then use that training data to build a subject-specific classifier. It also seems like the typical "power in a frequency band in a single channel" type feature is not sufficient so you will have to learn about different types of feature extraction and feature selection techniques.

If you do not need to change the task based on the online EEG analysis and instead you can analyze the data after the recording is finished then you might be better off using a simpler system. For example, you could simply record using whatever recording software Enobio gives you and include in the recorded data a digital signal that indicates when your stimuli are played.

Good luck.
-Chad

aasedeeq
Posts: 4
Joined: 05 Jun 2013, 09:00

Re: Q. real-time EEG to categorize human emotions using Enob

Post by aasedeeq » 05 Aug 2013, 08:42

Thank you Chad.
As you said, its better to do offline analysis and then deal with the real-time one.
Kind regards

Locked

Who is online

Users browsing this forum: No registered users and 3 guests