CAN ICA HELP IDENTIFY BRAIN TUMOR RELATED EEG SIGNALS?
M. Habl, Ch. Bauer , Ch. Ziegaus, E. W. Lang, F. Schulmeyer
elmar.lang@biologie.uni-regensburg.de
Scalp EEG has been used as a clinical tool for the di-
agnosis and treatment of brain diseases. A probabilis-
tic ICA algorithm modi ed by a kernel-based source
density estimation is studied to separate EEG signals
from tumor patients into spatially independent source
signals. A statistical method to automatically identify
artifactual and tumor related ICA components is also
presented.