help with creating scenario

Concerning processing components: filters, file load/save, visualizations, communication ...
Post Reply
bayo
Posts: 6
Joined: Sun Oct 25, 2009 1:18 am

help with creating scenario

Post by bayo »

Hi there,

Does anyone know what the "Level measure" does? I'm trying to create a scenario where i will be comparing a the band power of two frequency band e,g alpha and beta. I have an idea how to calculate the power in each band but I don't know if there is any "comparator" in openvibe.

Also, I intend to measure the power in a particular frequency band and compare it to some kind of baseline. for example reading in the band power for 10s initially, storing that value (as the baseline), and comparing future values of power band to the baseline.

Please help

Bayo

yrenard
Site Admin
Posts: 645
Joined: Fri Sep 01, 2006 3:39 pm
Contact:

Re: help with creating scenario

Post by yrenard »

Hello bayo,

this is not a kernel related question but anyway...

The level measure has been developed very quickly for a specific need and probably lacks a lot of functionality / documentation ! Facts are that it computes the relative contribution of each cell of the input matrix to the sum of all the cells. For example, if you compute the band power in 16-24 Hz for C3 and C4, at the end of the processing you will have a matrix with two values representing the beta power of C3 (let's call it BetaPowC3) and the beta power of C4 (let's call it BetaPowC4) respectively. Then the level measure will show you two gauges, the first one being BetaPowC3 / (BetaPowC3+BetaPowC4) and the second one being BetaPowC4 / (BetaPowC3+BetaPowC4).

For your second question about the comparison of two different levels, our partner GipsaLab is currently working on a box that is able to compute such comparison... It is not included in the standard distribution of OpenViBE yet because it's sill considered as unstable.

For the baseline problem, a box is currently missing that would just "substract" two matrices. This is easy to do but it does not exist for now. If such box existed, it would be possible to stim-based epoch the baseline signal, compute the band power and substract the result to the live measure. If you are interested, this would be a nice contribution. In any case, our partner GipsaLab is working on the "neurofeedback" pipeline implementation in OpenViBE so they will probably create such a box in the next few months.

In hope this helps,
Best regards,
Yann

Post Reply