Datenbestand vom 13. Juni 2019

Warenkorb Datenschutzhinweis Dissertationsdruck Dissertationsverlag Institutsreihen     Preisrechner

WICHTIGER HINWEIS
DER VERLAG IST IN DER ZEIT VOM 12.06.2019 BIS 23.06.2019 AUSCHLIESSLICH PER EMAIL ERREICHBAR.

aktualisiert am 13. Juni 2019

ISBN 9783843910590

Euro 84,00 inkl. 7% MwSt


978-3-8439-1059-0, Reihe Informatik

Martin Spüler
Assessing the Benefit of Adaptive Brain-Computer Interfacing

186 Seiten, Dissertation Eberhard-Karls-Universität Tübingen (2013), Softcover, A5

Zusammenfassung / Abstract

Brain-Computer Interfaces (BCIs) have emerged as a new technology that can either be used to enable communication in people suffering from paralysis or as neurofeedback tool for stroke rehabilitation. Recently, adaptive Brain-Computer Interfaces have been proposed as a new method to improve BCI systems. By continuously adapting to the changing brain signals, the problem of non-stationarity can be alleviated, which leads to more robust BCIs and increased performance.

This thesis tries to assess the benefit of adaptive Brain-Computer Interfacing. While the benefit is tackled from a theoretical side by running simulations of the mutual interaction between an adaptive BCI and a learning user, different adaptive and non-adaptive methods have also been developed to improve BCI performance in online experiments.

To properly determine the benefit of adaptive methods, this thesis does not concentrate on a specific BCI system, but several BCI systems and applications are considered. Different adaptive and non-adaptive methods were developed to improve performance in a MEG-based motor imagery BCI. Furthermore, methods for stroke rehabilitation were proposed and evaluated using non-invasive and invasive methods for recording the brain activity. At last, new methods to improve a BCI based on code-modulated visual evoked potentials were proposed and evaluated. These improvements led to the highest bitrates reported so far for a non-invasive BCI. In addition, it could be shown that adaptive methods can be utilized to allow an unsupervised calibration of a BCI.

Concluding from the results that were obtained for this thesis, adaptive methods do not seem to be overall better suited for improving BCI performance than non-adaptive methods. But adaptive Brain-Computer Interfacing is an important prerequisite that allows to apply BCI technology in new areas of application like stroke rehabilitation or that opens new possibilities like unsupervised calibration of a BCI, which might enable communication with complete locked-in patients.