6533b82ffe1ef96bd1295948

RESEARCH PRODUCT

A method for extracting subspace of deterministic sources from EEG data

Heikki LyytinenTapani RistaniemiAndriy IvannikovTommi Kärkkäinen

subject

Quantitative Biology::Neurons and Cognitionmedicine.diagnostic_testBasis (linear algebra)business.industryComputer scienceNoise reductionSpeech recognitionPattern recognitionElectroencephalographyLinear subspaceNoiseSignal-to-noise ratioEeg datamedicineArtificial intelligencebusinessSubspace topology

description

In this paper, an algorithm for separating linear subspaces of time-locked brain responses and other noise sources in multichannel electroencephalography data is proposed. The search criterion used by method discriminates time-locked brain components and noise components on the basis of the assumed deterministic behavior that the time-locked brain sources obey. The comprehensive derivation of the method is given together with the description and the analysis of the results of the method's application to simulated and real EEG data sets. The possibilities of improving the results are also discussed.

yearjournalcountryeditionlanguage
2008-03-012008 3rd International Symposium on Communications, Control and Signal Processing
https://doi.org/10.1109/isccsp.2008.4537438