Publica
Hier finden Sie wissenschaftliche Publikationen aus den FraunhoferInstituten. A maxmin approach to optimize spatial filters for EEG singletrial classification
 Cabestany, J.: 10th International WorkConference on Artificial Neural Networks 2009. Vol.1: Bioinspired systems: Computational and ambient intelligence : Salamanca, Spain, June 1012, 2009 ; proceedings Berlin: Springer, 2009 (Lecture Notes in Computer Science 5517) ISBN: 3642024777 ISBN: 9783642024771 ISSN: 03029743 pp.674682 
 International Work Conference on Artificial Neural Networks (IWANN) <10, 2009, Salamanca> 

 English 
 Conference Paper 
 Fraunhofer FIRST () 
Abstract
Electroencephalographic singletrial analysis requires methods that are robust with respect to noise, artifacts and nonstationarity among other problems. This work contributes by developing a maxmin approach to robustify the common spatial patterns (CSP) algorithm. By optimizing the worstcase objective function within a prefixed set of the covariance matrices, we can transform the respective complex mathematical program into a simple generalized eigenvalue problem and thus obtain robust spatial filters very efficiently. We test our maxmin CSP method with real world braincomputer interface (BCI) data sets in which we expect substantial fluctuations caused by daytoday or paradigmtoparadigm variability or different forms of stimuli. The results clearly show that the proposed method significantly improves the classical CSP approach in multiple BCI scenarios.