OpenViBE forum

Hello.

Ι have been dealing with Motor Imagery for the last year and a half. I use CSP + LDA and have tested it on about 20 people. Unfortunately there are people with whom the algorithm cannot be trained and it always predicts one class (usually the right). There are people it works very well, others not well, and others it doesn't work at all.
So I started trying to find alternative ways or modifications that could be made to make my system more powerful. Unfortunately I couldn't find any other scenarios with OpenVibe except from Graz Protocol with CSP and Laplacian.
Are there any other new scenarios?
Is it possible to use Deep Learning with OpenVibe? I would greatly appreciate any guidance, tutorials, etc.
Any solution or ideas are welcome.

Thanks in advance!

Hi, I'm Fabio PhD student, my doctoral thesis is based on the use of visual imagination as a control paradigm for BCI systems. I have implemented a block in OpenVibe to use CNN networks for classification of EEG signals in the time domain. If you are interested we could get in touch and make some collaboration. I am also interested in extending OpenVibe functionalities with more blocks, especially classification blocks based on deep learning.

Hi, I'm Fabio PhD student, my doctoral thesis is based on the use of visual imagery as a control paradigm for BCI systems. I have implemented a block in OpenVibe to use CNN networks for classification of EEG signals in the time domain. If you are interested we could get in touch and make some collaboration. I am also interested in extending OpenVibe functionalities with more blocks, especially classification blocks based on deep learning.

Dear @FabioBCI.

Sounds very interesting.
Please check your private messages for collaboration.