INSTITUTO NACIONAL DE ASTROFÍSICA, ÓPTICA Y eLECTRÓNICA

Brain computer interfaces (BCI) using multivariate analysis and recurrent neural networks.

This project has to do with the use of computational intelligence and signal processing for characterizing EEG signals, in a way that they can be useful for identification of task in brain computer interfaces (BCI). This project is developed in conjunction with some researchers in the coordination of electrical engineering in INAOE.

 

A BCI is a communication channel, which does NOT depend on the conventional “outputs” of the brain, like peripheral nervous or muscles (Kubanek et al. 2009). Currently, BCI applications are found in medical applications, video games, wearable systems etc.

 

Our objective is to build a system, capable of recognizing the orders given by a user with his/her thoughts. To do so, we are designing models to characterize the information embedded in a brain signal represented in EEG over time. Such characterization must contain information produced over time. Also, we are designing classifiers based on artificial neural networks that can be adapted to different users.

Currently, there are many research and practical applications related to this challenge; however their performance requires to be improved

Last modification: March 19, 2015