Project CEPBCI – Code-modulierte evozierte Potentiale zur Steuerung eines Brain-Computer Interface

Basic data

Acronym:
CEPBCI
Title:
Code-modulierte evozierte Potentiale zur Steuerung eines Brain-Computer Interface
Duration:
01/10/2015 to 31/10/2018
Abstract / short description:
A Brain-Computer Interface (BCI) allows to communicate or to control a computer by means of brain activity only. The aim of this technology is to enable completely paralyzed persons to communicate and interact with their environment. In previous works, it was shown that visual evoked potentials that are modulated by a pseudorandom code can be utilized for BCI control. Based on these code-modulated visual evoked potentials (c-VEPs), the applicant has developed a BCI system, which currently provides the highest communication speeds in the area of non-invasive BCI systems. However, the application of this system for completely paralyzed patients is not possible yet. The goal of this project is the extension of the current c-VEP BCI system to increase the accuracy and improve the usability. Further, the applicant wants to make the c-VEP useable by patients who are completely paralyzed. Since the latter is of special importance, different approaches shall be used. First, the visual stimulation should be adapted to be useable without eye movement and the method using code-modulated Stimuli to evoked potentials shall be transferred to auditory stimulation to make the BCI system useable by people who are blind or have impaired vision. Therefore, also the methods used for signal processing need to be improved. Further, unsupervised machine learning methods should be used for calibration of the BCI. This is a completely novel approach and should enable the use of BCIs also for patients with complete locked-in syndrome, for whom all existing BCI systems does not work. At the end of the project, the user-friendly BCI system should be tested with completely paralyzed in cooperation with external partners.
Keywords:
brain
Gehirn
machine learning
Maschinelles Lernen
Mensch-Maschine Interaktion

Staff

Managers

Wilhelm Schickard Institute of Computer Science (WSI)
Department of Informatics, Faculty of Science

Local organizational units

Wilhelm Schickard Institute of Computer Science (WSI)
Department of Informatics
Faculty of Science

Funders

Bonn, Nordrhein-Westfalen, Germany
Help

will be deleted permanently. This cannot be undone.