How to trigger sound based on different EEG frequency band??

[Aizat89]

urgent help!!
from the reading and forum, i found that the threshold part cannot be done directly with available boxes.
I have no ideas since to develop new box, its required me to have some basic in C++

what i would like to do is:
IF brain wave is 7-13Hz, THEN play Sound_1
IF brain wave is 7-13Hz, THEN play Sound_2

How do i design this?

[lbonnet]

Hi and sorry for late reply,
The official support team has a lot of work right now to train a whole new crew of engineers.

To be honest, several solutions can be implemented for a threshold-based BCI.

If you don't want to write C++ code (which I understand perfectly), you can use boxes. But it may sound a bit complex...

To build a BCI scenario that plays a sound if a frequency range is above or below a level threshold, you will need:
- the EEG signal processing: a good start would be computing a band power in the desired frequency range.
see the motor-imagery or the neurofeedback scenarios given with openvibe for example.
- The power-to-stimulation part: this step transforms the band power level into stimulations that will trigers the sound player
To do so, you can use a simple DSP to set the threshold you want (equation : x - Thresh).
Then you can use the new sign change detector box to output a stimulation when the DSP output goes above or below 0.
- Sound player : this box plays a sound upon receiving a defined stimulation.

This is roughly the design I see that would require nothing more than what's inside openvibe.
However, there are some limitations. The threshold for example must be defined manually, and may be hardly dependent on the subject, the time of the experiment, the current installation setup, etc. You would have to configure it manually everytime.

If you have hard time understanding the processing pipelines, please try to search and learn more about signal processing basics.
Otherwise, building BCI can be very hard. Even if openvibe can provide fast and easy way to build BCI, it still requires the user to know what's the purpose of each box and how to use them.

I hope this will help you !

Laurent-

[Aizat89]

Thanks for the very informative response!!
Im really2 sorry for my stupidity and still have not achieve the solution..
is it the correct design flow as ur mention?

[jlegeny]

Hello Aizat89,

the processing chain looks ok by design, however maybe it would be better to use a spatial filter instead of a simple channel selector (for example using a laplacian).

The other questions is what exactly are you trying to detect. Maybe you could use your scenario with filtering in alpha band and testing if closing the eyes triggers the sound. It is also very important to set the right threshold.

Cheers
Jozef

[Aizat89]

thanks for replying..
i'm really appreciate your kindness if you can provide me the roughly idea with an image of the design (since i'm newbie in openvibe and programming, example in image could rather help me for better understanding the concept .) plsss help me..i have no idea whats the problem..the sound still not produced..

[Aizat89]

From the error, its said that it required at least 2 channel..is it due to the unavailability of EEG tool (Mindset)? how should i overcome the problem..?

[jlegeny]

Hello,

the reference channel box's purpose is to subtract one channel from all of the others. By definition this box can only work if there are at least two channels. Neurosky already has an implicit reference (on the ear) and it is used by the hardware directly, you do not need to use a reference channel in your case.

Cheers
Jozef