PASCAL - Pattern Analysis, Statistical Modelling and Computational Learning

Artifact-Insensitivity of CSP in Motor Imagery BCI
Irene Winkler and Michael Tangermann
Int J Bioelectromagnetism Volume 13, Number 2, pp. 72-73, 2011.


While Brain-Computer Interfaces (BCIs) should generate control commands based on neural activity only, the electroencephalogram contains artifacts such as eye-or muscle activity, and healthy subjects might use those (sub-)consciously for BCI-control. We analyze the influence of an automatic, subject independent artifact reduction step on the performance of a motor imagery setup that uses Common Spatial Patterns. The offline test conducted on data from 80 subjects revealed no performance drop of the Berlin Brain-Computer-Interface after rigorous artifact reduction based on Independent Component Analysis (ICA).

EPrint Type:Article
Project Keyword:Project Keyword UNSPECIFIED
Subjects:Brain Computer Interfaces
ID Code:9483
Deposited By:Martijn Schreuder
Deposited On:16 March 2012