Beta 1


Title Signal Detection in EEG Brainwaves -a classification based approach
Author Hede, Simon Christian
Supervisor Hansen, Lars Kai (Cognitive Systems, Department of Informatics and Mathematical Modeling, Technical University of Denmark, DTU, DK-2800 Kgs. Lyngby, Denmark)
Institution Technical University of Denmark, DTU, DK-2800 Kgs. Lyngby, Denmark
Thesis level Master's thesis
Year 2010
Abstract Supervised classification within EEG signals has previously been applied only within Brain Computer Interfaces. In this thesis, classification rates are examined for EEG data originating from the cognitive science experiment called the Standard Object Recognition Paradigm, where the activation of human memory is activated by visual stimulation. Applying methods previously used within Brain Computer Interfaces such as the Support Vector Machine and feature selection algorithms, the classification rates for the data are improved, showing that the data contains enough signal to achieve feasible classification rates. Aspects within Support Vector Machine classification and general EEG classification are discussed. By applying a kernel based visualisation method with the Support Vector Machine, it is shown that this method is capable of extracting information from EEG signals, by finding that the signal content used for the classification, is the one expected to be present from previous studies of the Standard Object Recognition Paradigm.
Imprint Technical University of Denmark (DTU) : Kgs. Lyngby, Denmark
Series IMM-M.Sc.-2010-67
Fulltext
Original PDF ep10_67_net.pdf (1.49 MB)
Admin Creation date: 2010-09-03    Update date: 2010-10-28    Source: dtu    ID: 266445    Original MXD